ENMAX Energy Corporation v TransAlta Generation Partnership

In the Court of Appeal of Alberta

Citation: ENMAX Energy Corporation v TransAlta Generation Partnership, 2022 ABCA

206

Date: 20220609

Docket: 1901-0237AC

1901-0235AC

Registry: Calgary

1901-0237AC

Between:

ENMAX Energy Corporation

Appellant

(Applicant)

- and -

TransAlta Generation Partnership

Respondent

(Respondent)

1901-0235AC

And Between:

The Balancing Pool

Appellant

(Applicant)

- and -

TransAlta Generation Partnership

Respondent

(Respondent)

_______________________________________________________

The Court:

The Honourable Justice Marina Paperny

The Honourable Justice Patricia Rowbotham

The Honourable Justice Jo'Anne Strekaf

_______________________________________________________

Memorandum of Judgment

Appeal from the Decision by

The Honourable Justice P.R. Jeffrey

Dated the 26th day of June, 2019

Filed on the 6th day of November, 2019

(2019 ABQB 486, Docket: 1601 16992)

_______________________________________________________

Memorandum of Judgment

_______________________________________________________

The Court:

Introduction

[1]

The duty to comply with the rules of natural justice and procedural fairness is fundamental

to a fair hearing. The nature and extent of that duty are fluid, depending on the context of the

dispute, the issues to be decided and, importantly, the nature of the forum, whether that be a court,

an administrative tribunal or, as in this case, an arbitral panel.

[2]

The Arbitration Act, RSA 2000, c A-43 (the Act), which governs domestic arbitrations in

Alberta, expressly requires an arbitral tribunal to treat the parties equally and fairly (section 19(1))

and give them an opportunity to present their case and meet that of the other side (section 19(2)).

The Act also provides that a court may, on the application of a party, set aside an arbitration award

if there is a fundamental flaw in the process or the make-up of the tribunal, including if the

applicant was treated “manifestly unfairly and unequally” or “not given an opportunity to present

a case or to respond to another party’s case”: section 45(1)(f).

[3]

The appellants say that, in the course of a lengthy and complex commercial arbitration,

they were treated “manifestly unfairly and unequally” by the arbitration panel and not given an

opportunity to present their case and that the arbitration award should therefore be set aside. This

appeal requires us to consider what constitutes manifest unfairness such that judicial intervention

in setting aside the award is warranted.

[4]

The fact that the issue arises in the arbitral context is important. The Act sets the guidelines

and rules for parties who either choose, or are required, to use arbitration to resolve their disputes,

recognizing that arbitration is intended to be timelier and more private, allow for specialized

expertise, and offer finality. The Act intentionally limits access to the court system. An appeal to

the court from an arbitration award is permitted only if the parties’ arbitration agreement so

provides (section 44(1)) or if the court grants permission to appeal on a question of law (section

44(2)). In this case, the arrangement that governs the parties’ relationship expressly precludes an

appeal and permission to appeal, although sought, was not granted: see ENMAX Energy

Corporation v TransAlta Generation Partnership, 2018 ABQB 142.

[5]

Section 45, pursuant to which the appellants made their application, allows the court to set

aside an award in specific and limited circumstances. The chambers judge found those

circumstances did not exist and declined to set the award aside. The matter before the arbitration

panel was complex and has significant financial implications for the parties, but the specific issue

raised and decided by the chambers judge was relatively discrete. The issue before this court on

appeal is likewise narrow: did the chambers judge err in concluding that the arbitration panel’s

Page: 2

rulings on document disclosure with respect to one matter did not prevent the appellants from

making their case in this multi-party commercial arbitration nor result in manifest unfairness. For

the reasons that follow, we have determined that there was no manifest unfairness, and the

chambers judge did not err in declining to set aside the arbitration award.

Brief background

[6]

The dispute arises out of an arbitration between the respondent, TransAlta Generation

Partnership (TransAlta), and the appellants, ENMAX Energy Corporation (ENMAX) and the

Balancing Pool. TransAlta owns and operates the Keephills Generating Units. TransAlta, as

owner, entered into a Power Purchase Arrangement (PPA), created pursuant to the Power

Purchase Arrangements Determination Regulation, Alta Reg 175/2000, to supply electrical power

from Keephills Generating Unit #1 (the Unit) to ENMAX, as buyer. The Balancing Pool is

established by the Electric Utilities Act, SA 2003, c E-5.1, and has financial obligations under the

PPA. It is entitled to participate in any arbitration between an owner and a buyer.

[7]

On March 5, 2013, the Unit “tripped” and went off-line for 216 days.

[8]

The issue underlying the dispute is whether the losses from the Unit being off-line should

fall on the owner, TransAlta, or on ENMAX and the Balancing Pool. TransAlta claimed the

shutdown was an event of force majeure, as defined in the PPA. If the event is a result of force

majeure the Balancing Pool is required to pay TransAlta’s fixed and variable costs of generation,

and TransAlta is relieved of its obligation to pay ENMAX for the time the Unit was off-line.

However, if the shutdown was a foreseeable part of normal operations, then TransAlta is required

to pay compensation in the form of an Availability Incentive Payment. The stakes are high. During

the period the Unit was off-line, ENMAX incurred between $128 million to $168 million in costs;

the Balancing Pool incurred $40 million.

[9]

TransAlta claimed force majeure. ENMAX and the Balancing Pool disagreed.

[10] The dispute was referred to arbitration under the Act, as required by Article 19.4 of the

PPA. The arbitration panel, in a lengthy award decision, concluded that the shutdown of the Unit

was an event of force majeure as defined in the PPA. The arbitration panel found TransAlta had

acted in accordance with Good Operating Practice, or GOP, pursuant to its obligations under the

PPA, and that the circumstances were beyond TransAlta’s reasonable control.

[11] ENMAX and the Balancing Pool sought to challenge the award. They each applied for

permission to appeal the award pursuant to section 44(2) of the Act, which contemplates appeals

on questions of law from an arbitration panel to the Court of Queen’s Bench with permission of

the court in circumstances where the arbitration agreement does not provide for an appeal.

Page: 3

However, the PPA expressly precludes an appeal on any ground, and the chambers judge

concluded no appeal of the arbitration award was available: see Article 19.4(f) of the PPA and

ENMAX Energy Corporation v TransAlta Generation Partnership, 2018 ABQB 142. That earlier

decision of the chambers judge not to grant permission to appeal pursuant to section 44(2) was not

appealed to this Court, and is not before us now.

[12] The appellants also applied to set aside the award pursuant to section 45 of the Act, which

allows the court to “set aside” awards on specified grounds:

45(1) On a party’s application, the court may set aside an award on any of the

following grounds:

(a) a party entered into the arbitration agreement while under a legal incapacity;

(b) the arbitration agreement is invalid or has ceased to exist;

(c) the award deals with a matter in dispute that the arbitration agreement does

not cover or contains a decision on a matter in dispute that is beyond the

scope of the agreement;

(d) the composition of the arbitral tribunal was not in accordance with the

arbitration agreement or, if the agreement did not deal with the matter, was

not in accordance with this Act;

(e) the subject-matter of the arbitration is not capable of being the subject of

arbitration under Alberta law;

(f) the applicant was treated manifestly unfairly and unequally, was not given

an opportunity to present a case or to respond to another party’s case, or was

not given proper notice of the arbitration or of the appointment of an

arbitrator;

(g) the procedures followed in the arbitration did not comply with this Act or

the arbitration agreement;

(h) an arbitrator has committed a corrupt or fraudulent act or there is a

reasonable apprehension of bias;

(i) the award was obtained by fraud.

Page: 4

[13] It is evident from this list that the Act contemplates the setting aside of awards on the basis

of fundamental flaws in the establishment of the arbitration or its processes. The Act as a whole

limits judicial intervention and contemplates that arbitration awards will bind the parties except in

limited circumstances: see, e.g., sections 37 and 44. The provisions of section 45 permit the court

to set aside an award in circumstances that go to the heart of the arbitration itself. Section 45 is not

a substitute for an appeal or a judicial review and ought not be used to circumvent the statutory

limitations on court intervention.

[14] The appellants’ application to set aside the award was based on section 45(1)(f); they

argued that procedural decisions made by the arbitration panel in relation to certain document

disclosure resulted in “manifest unfairness” and deprived them of a fair “opportunity to present a

case or to respond to another party’s case”. The application to set aside was dismissed by the

chambers judge: ENMAX Energy Corporation v TransAlta Generation Partnership, 2019 ABQB

486. This is the decision under appeal. For convenience, we will refer to the arbitration panel’s

decision as “the award” and the decision of the chambers judge as “the reasons”.

[15] The specific records that the appellants say should have been disclosed related to

TransAlta’s reliance on a diagnostic test used to determine the location of a problem with the Unit,

known as the Tennessee Valley Authority, or TVA, test. As part of their submissions in the

arbitration, the appellants took the position that TransAlta unreasonably relied on the TVA test as

showing that the problem with the Unit originated with the top T5 stator bar located inside the

generator, and that TransAlta’s decision to take the generator off-line to do a complete stator

rewind was unreasonable. In addition to arguing that the TVA test was not widely recognized as a

reliable primary diagnostic test, ENMAX sought records relating to TransAlta’s experience with

the TVA test at its other facilities (the fleet-wide records). Records with respect to the Keephills

generator were disclosed, but the appellants did not receive records relating to the conduct of TVA

tests at other TransAlta generating facilities that had not been taken off-line (and were not the

subject of the dispute). In its procedural orders, the arbitration panel declined to require TransAlta

to provide the fleet-wide records. The appellants submit the undisclosed records were relevant and

material to their case and that the reasons for the award make clear that the arbitration panel relied

on the absence of the undisclosed records to support its conclusion that the event was the result of

force majeure.

[16] The chambers judge identified issues with some of the procedures followed by the

arbitration panel, but declined to set aside the award, stating at para 147 of the reasons:

I dismiss the applications to set aside the Award. When the entirety of the process

is considered within the particular circumstances of the dispute, I do not find any

manifest unfairness or inequality, nor do I find the Applicants were prevented from

Page: 5

meeting the case of TransAlta. This is my conclusion not just on the individual

concerns raised about the Panel’s process, but when assessing their cumulative

effect.

Issues on appeal

[17] Permission to appeal was granted on one issue: whether the chambers judge erred in

concluding the appellants were not treated manifestly unfairly and unequally, or prevented from

presenting their case or responding to TransAlta’s case: ENMAX Energy Corporation v TransAlta

Generation Partnership, 2020 ABCA 68.

[18] In addressing that issue, it is necessary to consider the correct approach to an application

to set aside pursuant to section 45, and whether the chambers judge erred in his assessment of

whether the process adopted by the arbitration panel, in the whole context of the arbitration,

resulted in manifest unfairness such that judicial intervention in the award is warranted. There are

two aspects to that question: Did the arbitration panel deny the appellants document disclosure as

they allege? Did the absence of the undisclosed records deny the appellants the opportunity to

present their case or result in manifest unfairness within the meaning of section 45(1)(f) of the

Act?

Standard of review

[19] The application before the chambers judge was a section 45 application to set aside the

award on specific grounds that go to the fundamental nature of the arbitration. The chambers judge

was required to identify the correct test under that provision and apply that test to the award to

determine if the process followed was “manifestly unfair” or deprived the appellants of the

opportunity to present their case or respond to the case of TransAlta, contrary to section 45(1)(f).

Permission to appeal to this Court was granted pursuant to section 48 of the Act, which provides

for appeals from decisions of the Court of Queen’s Bench under sections 44, 45 or 47 of the Act.

[20] The appellants say this Court owes no deference to the assessment and decision of the

chambers judge, referring to Alberta (Minister of Infrastructure) v Nilsson, 2002 ABCA 283 at

para 38. In Nilsson, the Queen’s Bench judge was acting as an appellate court on a section 44

appeal of an arbitrator’s decision; this Court held that “a second appellate court does not owe a

duty of deference to a first appellate court”. In this case, the chambers judge was considering an

application to set aside the award under section 45; he was not hearing an appeal from the

arbitration panel. Nor was he presiding over a judicial review and, accordingly, the Agraira

standard that this Court, on appeal, will “step into the shoes” of the chambers judge and perform a

Page: 6

de novo review of the arbitration panel’s decision does not apply: see Agraira v Canada (Public

Safety and Emergency Preparedness), 2013 SCC 36 at para 46.

[21] The appeal from the chambers judge to this Court is subject to the usual appellate standards

of review, as set out in Housen v Nikolaisen, 2002 SCC 33. The interpretation of section 45(1)(f),

which incorporates elements of natural justice and procedural fairness, is a question of law

reviewable for correctness. The chambers judge’s factual findings are reviewed for palpable and

overriding error. His application of the test under section 45(1)(f) to those facts is a matter of mixed

fact and law that, in the absence of extricable legal error, is accorded deference on appeal. His

determination of whether the process followed by the arbitration panel, in the circumstances of the

arbitration as a whole, rises to the level of manifest unfairness or is otherwise contrary to section

45(1)(f) is therefore reviewed for palpable and overriding error.

[22] If there is a flaw in the arbitral process sufficient to contravene section 45(1) of the Act, the

decision of whether to set the arbitration award aside engages the discretion of the chambers judge

and is entitled to deference on appeal: see Popack v Lipszyc, 2016 ONCA 135 at para 25.

Standard of review on an application to set aside an arbitration award under section 45

[23] The appellants submit that, given the circumscribed nature of access to judicial intervention

and the significance of fundamental fairness to the arbitration process, the chambers judge should

not have shown any deference to the arbitration panel’s procedural decisions. They say the court

must play a supervisory role to ensure parties are afforded natural justice. A claim that an arbitrator

has breached natural justice raises a question of law, or more specifically, a question of fairness,

and no deference is due.

[24] In section 45, the legislature has provided for a discretionary review by the court of certain

aspects of arbitrations, enumerating nine grounds pursuant to which the court may set aside an

arbitration award. The context of this review is critical to understanding the legislative intent

behind this provision. Arbitration, as opposed to litigation in court, is intended to be more

expeditious, private, and economical. Arbitrators can be chosen for their expertise. The procedures

can be set by the parties and disagreements dealt with by the arbitration panel. The Act was

designed to limit access to the courts in the absence of agreement by the parties to an appeal

(appeals are precluded by the PPA that governs the relationship of the parties here), or some

fundamental flaw in the establishment or conduct of the arbitration as set out in section 45. As

noted by J. Brian Casey, Arbitration Law of Canada: Practice and Procedure, 3rd ed (New York,

JurisNet, LLC, 2017) at p 508:

Page: 7

The setting aside of an award is something different from an appeal. Rather than

attacking the tribunal’s judgment as occurs in an appeal, an application to set aside

an award is a challenge to the validity of the arbitral process, the result of which, if

successful, takes away the fundamental underpinning of the award.

[25] As noted, section 45 grants a discretion to the court in the setting aside of arbitration

awards: Kawchuk v Kovacs, 2016 ABCA 210 at para 35; Popack at para 25. That discretion must

be exercised with due regard to the legislative scheme, including section 20(1) of the Act, which

grants to arbitrators a wide discretion on procedural matters. Section 45(1)(f) is not intended to

override that provision. Thus, in considering an application to set aside, it is appropriate for the

court to give deference to the procedural choices of the arbitration panel, subject to the requirement

in sections 19 and 45(1)(f) that those choices not result in manifest unfairness to the parties or

deprive a party of the opportunity to present its case.

[26] The questions before the chambers judge are: what constitutes manifestly unfair or unequal

treatment or the deprivation of an opportunity to present a case for purposes of section 45(1)(f);

whether the procedures adopted by the arbitration panel fail to comply with those stipulations; and,

if so, whether the award should be set aside.

Interpretation of section 45(1)(f)

[27] The appellants submit that the chambers judge misconstrued section 45(1)(f) and that the

term “unfairness” in section 45(1)(f) must be given a robust interpretation because of the

circumscribed access to the courts under the Act. As section 45(1)(f) is the only safeguard and

opportunity for judicial scrutiny, they say, the provision should be interpreted broadly with a view

to remedying any perceived unfairness, regardless of whether it affected the ultimate award, or

any lack of opportunity to present a case.

[28] Determining what constitutes manifest unfairness and unequal treatment or lack of

opportunity to present a case under section 45(1)(f) is a matter of statutory interpretation. It

requires reading the section in its entire context, and in its grammatical and ordinary sense,

harmoniously with the statutory scheme and the object of the statute, and the intention of the

legislature: Rizzo & Rizzo Shoes Ltd (Re), [1998] 1 SCR 27, 154 DLR (4th) 193 at para 21.

[29] The overarching purpose of arbitration legislation is to promote adherence to agreements,

efficiency, and fairness and to lend credibility to an important dispute resolution process: EPCOR

Power LP v Petrobank Energy and Resources Ltd, 2010 ABCA 378 at para 16. Arbitrations are

intended to provide a faster and less expensive process to resolve disputes: Mitchell v Mitchell,

2019 ABQB 420 at para 17.

Page: 8

[30] The arbitration panel is vested with discretion to determine the procedure to be followed

and the way in which evidence is admitted. Further, the arbitration panel is not bound by the rules

of evidence or any other law applicable to judicial proceedings and has power to determine the

admissibility, relevance, and weight of any evidence: Act, sections 20 and 21.

[31] Section 19 of the Act provides that the parties must be treated equally and fairly, and each

party must be given an opportunity to present its case and respond to the other party’s case. Under

section 6, the court may intervene in matters governed by the Act on four specified grounds only,

one of which is to prevent manifestly unfair or unequal treatment of a party to an arbitration

agreement. The three other grounds on which the court may intervene are to assist the arbitration

process, to ensure that an arbitration is carried on in accordance with the arbitration agreement,

and to enforce awards. This Court has held that with respect to section 6, “[t]he starting point is

non-intervention, subject to a stated list of fairly narrow and discrete exemptions to be interpreted

in accordance with the expressed legislative purpose”: EPCOR Power at para 18.

[32] A nearly identical provision is found in Ontario’s arbitration legislation, Arbitration Act,

1991, SO 1991, c 17. The Supreme Court in TELUS Communications Inc v Wellman, 2019 SCC

19 at para 56 described that section as follows:

Stated succinctly, s. 6 signals that courts are generally to take a “hands off”

approach to matters governed by the Arbitration Act. This is “in keeping with the

modern approach that sees arbitration as an autonomous, self-contained, self-

sufficient process pursuant to which the parties agree to have their disputes resolved

by an arbitrator, not by the courts” (Inforica Inc. v. CGI Information Systems and

Management Consultants Inc., 2009 ONCA 642, 97 O.R. (3d) 161, at para. 14).

[33] Once an arbitration award has been issued, there are also limited, circumscribed remedies

available to the parties to challenge the award. The arbitration award binds the parties unless it is

set aside or varied under section 44 or section 45: Act, section 37.

[34] Section 44 is the appeal provision. It states:

44(1) If the arbitration agreement so provides, a party may appeal an award to the

court on a question of law, on a question of fact or on a question of mixed law and

fact.

(2) If the arbitration agreement does not provide that the parties may appeal an

award to the court on a question of law, a party may, with the permission of the

court, appeal an award to the court on a question of law.

Page: 9

(2.1) The court shall grant the permission referred to in subsection (2) only if it is

satisfied that

(a) the importance to the parties of the matters at stake in the arbitration

justifies an appeal, and

(b) the determination of the question of law at issue will significantly affect

the rights of the parties.

(3) Notwithstanding subsections (1) and (2), a party may not appeal an award to the

court on a question of law that the parties expressly referred to the arbitral tribunal

for decision.

[35] The appeal provision limits judicial intervention and promotes finality by restricting the

circumstances under which the court may review an arbitration award. The availability and scope

of an appeal depends on what the parties agreed to under the arbitration agreement. If the parties

did not agree that they could appeal an award, the appeal cannot proceed unless the court grants

permission after determining that the appeal involves a question of law and is warranted in the

circumstances pursuant to the requirements in subsection (2.1). Court intervention is further

restricted under subsection (3), by precluding appeals involving a question of law which was

expressly referred to the arbitration panel.

[36] The second remedy available to a party after the issuance of an arbitration award is to apply

to the court to set it aside under section 45. As with appeals, the Act limits when an award may be

set aside to specific enumerated grounds:

45(1) On a party’s application, the court may set aside an award on any of the

following grounds:

(a) a party entered into the arbitration agreement while under a legal

incapacity;

(b) the arbitration agreement is invalid or has ceased to exist;

(c) the award deals with a matter in dispute that the arbitration agreement

does not cover or contains a decision on a matter in dispute that is beyond

the scope of the agreement;

(d) the composition of the arbitral tribunal was not in accordance with the

arbitration agreement or, if the agreement did not deal with the matter, was

not in accordance with this Act;

Page: 10

(e) the subject matter of the arbitration is not capable of being the subject of

arbitration under Alberta law;

(f) the applicant was treated manifestly unfairly and unequally, was not

given an opportunity to present a case or to respond to another party’s case,

or was not given proper notice of the arbitration or of the appointment of an

arbitrator;

(g) the procedures followed in the arbitration did not comply with this Act

or the arbitration agreement;

(h) an arbitrator has committed a corrupt or fraudulent act or there is a

reasonable apprehension of bias;

(i) the award was obtained by fraud.

[37] From this list, it is clear that the Act contemplates the setting aside of awards on the basis

of fundamental flaws in the establishment or conduct of the arbitration or its process. Section 45

is not a substitute for an appeal or a judicial review and ought not be used to circumvent the

statutory limitations on court intervention. Additionally, the language “may set aside an award”

(emphasis added), grants the court discretion on whether to set aside the award, even if one of the

listed grounds is established. This interpretation is supported by subsection 45(8), which allows

for alternative remedies: “Instead of setting aside an award, the court may remit it to the arbitral

tribunal and give directions about the conduct of the arbitration.”

[38] If a party wishes to appeal a decision of the Court of Queen’s Bench made under section

44 or section 45, permission of a justice of the Court of Appeal is required: Act, section 48. This

again reinforces the court’s non-interventionist role in arbitrations and encourages finality.

[39] Regarding the specific phrase “treated manifestly unfairly and unequally” in section

45(1)(f), we agree with the chambers judge that “manifestly” speaks to how apparent the unfairness

is and whether it is obvious or clear: see Oxford English Dictionary, (Oxford University Press,

2021) sub verbo “manifest, adj.”.

[40] On a review of section 45(1)(f) as a whole, in the context of the entire Act and the purposes

of the legislation, we make several observations. The Act imposes an obligation on arbitral panels

to accord parties natural justice. While it is accepted as a foundational principle that the arbitration

panel will treat parties fairly and give them an opportunity to present their case, it does not follow

that the same procedural rigors present in civil litigation are required. The inclusion of the language

“manifestly” unfair, the nature of the grounds enumerated under section 45, the statutory

Page: 11

limitations on judicial intervention, the primacy placed on the arbitration agreement and party

autonomy, the legislated restrictions on appeals, and the arbitration panel’s discretion to determine

the procedure to be followed, all suggest that procedures which are more efficacious and

economical and perhaps less robust than those present in civil litigation are acceptable.

[41] A review of the jurisprudence offers additional guidance in determining what amounts to

manifestly unfair and unequal treatment or a failure to provide an opportunity to present a case

within the meaning of section 45(1)(f) of the Act.

Unfairness and natural justice in arbitration

[42] The appellants object to any reliance on international arbitration cases, arguing that factors

such as international comity and the global marketplace, which are relevant to international

arbitrations, do not apply in the domestic context. They say the wording of section 45(1)(f) is not

identical to, and is far more expansive than, the wording found in legislation governing

international arbitrations with respect to when a court may set aside an award due to unfairness.

Article 18 of the UNICTRAL Model Law on International Commercial Arbitration, which is

adopted by Alberta’s International Commercial Arbitration Act, RSA 2000, c I-5, requires that

parties “be treated with equality and each party shall be given a full opportunity of presenting his

case”. Article 34(2)(a)(ii) provides for the setting aside of an arbitral award in specified

circumstances, including where a party was “unable to present his case”.

[43] While we acknowledge that there are points of distinction between domestic and

international arbitrations, the overarching purposes of arbitration in both contexts are similar. As

such, international arbitration decisions provide helpful guidance in assessing what treatment rises

to the level of manifest unfairness to justify setting aside an award under section 45(1)(f) of the

Act.

[44] Beginning with Alberta decisions, Mitchell was a judicial review of a matrimonial property

arbitration award. With respect to section 45(1)(f) of the Act, the Court of Queen’s Bench stated

at para 18:

The strong wording under section 45(1)(f), which allows a Court to set aside an

award only if the party was treated manifestly unfairly and unequally, is consistent

with the purposes of dispute resolution through arbitration rather than the courts.

The specific inclusion of the word “manifestly” demonstrates a legislative intent

that not every potential unfairness in an arbitration proceeding or decision will

permit the Court to set aside an arbitration award.

Page: 12

[45] In Contract Policy Committee v FortisAlberta Inc, 2012 ABQB 653, the court interpreted

the language “manifestly unfair or unequal treatment” under section 6 of the Act to mean that “the

Legislature is willing to accept that a result may be unfair, but not so unfair as to justify

intervention”: at para 36.

[46] In Ary Homes Inc v Triple-A Classic Homes Ltd, 2017 ABQB 510 at para 45, the court

noted that procedural fairness depends on the context including the subject-matter of the dispute,

the circumstances of each case, the nature of the inquiry, and the rules under which the parties

have agreed to arbitrate the dispute. Natural justice in arbitrations requires the arbitrator to act in

good faith, fairly listen to both sides, and give the parties a fair opportunity to make

representations, including to correct or contradict any statements that are prejudicial to their view

(citing 0927613 BC Ltd v 0941187 BC Ltd, 2015 BCCA 457 at paras 59-60). In Flock v Flock,

2007 ABQB 307 at para 49, leave to appeal ref’d 2007 ABCA 287, the court concluded that the

protections under section 45(f) include “the right to a fair hearing, free from bias and free from

procedures which confer relative advantages or disadvantages on one of the parties”.

[47] The court in Ary Homes also noted that it was not unusual in arbitrations for evidence to

be produced in a tight time frame. If a party receives evidence that it did not anticipate just before

or during the arbitration, it may respond as best it can or ask for an adjournment, which is often a

strategic or tactical decision (at paras 75-76), and the applicant “must live with its procedural or

strategic choices or omissions”: at para 80.

[48] Several decisions from other jurisdictions are also instructive.

[49] In Rhéaume c Société d'investissements l'Excellence inc, 2010 QCCA 2269, leave to appeal

to SCC ref’d [2011] SCCA No 57, at issue was whether an arbitration panel member had breached

the obligation of deliberative secrecy imposed on arbitrators under Quebec’s Code of Civil

Procedure (as it existed at the time), and if so, whether the arbitration award ought to be set aside.

The court determined that a breach had occurred when the arbitration panel member disclosed

publicly to the parties and third parties that the panel’s decision would be unanimous. However,

the court refused to set aside the arbitration award as the procedural breach did “not affect the

integrity of the process as a whole”: at para 63. In coming to this result, the court found that the

absence of judicial surveillance except on limited grounds and the clear legislative intent favouring

a “hands-off” approach is entirely inconsistent with the automatic annulment of arbitral awards for

procedural defects, however minor: para 57. The court went on to state, at para 61:

A court called upon to adjudicate such a proceeding must balance the nature of the

breach in the context of the arbitral process that was engaged, determine whether

Page: 13

the breach is of such a nature to undermine the integrity of the process, and assess

the extent to which the breach had any bearing on the award itself.

[50] Interlink Business Management Inc v Bennett Environmental Inc, 2007 BCSC 1538 was

decided under an earlier version of British Columbia’s domestic arbitration statute, the

Commercial Arbitration Act, RSBC 1996 c 55, where an “arbitral error” was defined to include a

failure to observe the rules of natural justice. The applicant argued that by denying its application

for oral examinations for discovery the arbitrator prevented it from proving its case, since the very

lack of evidence was the basis for the arbitrator dismissing a portion of its claim. The court

ultimately granted leave to appeal on several questions of law, but found no denial of natural

justice, stating “[t]he process was fair, even if it was not the most favourable”: at para 13. The

court held that determining appropriate procedures was within the arbitrator’s jurisdiction and the

applicant had the opportunity to obtain documents and cross-examine witnesses at the hearing.

[51] Mungo v Saverino, [1995] OJ No 3021 (SCJ) (QL), 1995 CarswellOnt 3298 (WL) involved

a shareholder dispute resolved by way of a domestic arbitration under an earlier version of

Ontario’s legislation, the Arbitration Act, SO 1991, c 17. Section 46 provided that the court may

set aside an arbitration award on several grounds, including if a party “was not treated equally and

fairly, was not given an opportunity to present a case or to respond to another party’s case…”. The

unsuccessful parties applied to vary or resubmit to the arbitrator certain aspects of the award on

the basis that the arbitrator had exceeded his jurisdiction and denied the applicants natural justice

in issuing an award without hearing direct evidence or giving the applicants an opportunity to lead

evidence. In dismissing the applicants’ claim that they had been denied natural justice, the court

recognized that the choice of arbitration involves a trade-off between perfection versus economy

and finality. Thus, the court noted, at para 80:

It is therefore important for the court to resist its natural tendency, faced with a clear

and attractive argument on jurisdiction and natural justice, to plunge into the details

of the arbitration and second-guess the arbitrator not only on the result but also on

the punctilio of the process. If an arbitration is basically fair, courts should resist

the temptation to plunge into detailed complaints about flaws in the arbitration

process.

[52] The principle that not every procedural breach during an arbitration will result in judicial

intervention has been reiterated by other courts. Courts should intervene “only in cases of the most

egregious procedural breaches”: Ryan v Cacioppo, [2008] OJ No 4738 (QL), 2008 61245

(ONSC) at para 16. In Fuego Digital Media Inc v DAC Group (Holdings) Limited, 2018 ONSC

2897, the court put it this way: “[w]hen the allegation is one of denial of natural justice, the

applicant must demonstrate that the arbitrator’s conduct was sufficiently egregious to offend basic

Page: 14

notions of morality and justice”: at para 23, citing Consolidated Contractors Group SAL (Offshore)

v Ambatovy Minerals SA, 2017 ONCA 939 [Consolidated Contractors], leave to appeal to SCC

ref’d, [2018] SCCA No 46.

[53] Both Ryan and Fuego Digital Media involved arbitrations under Ontario’s domestic

arbitration legislation, whereas Consolidated Contractors considered a challenge to an

international commercial arbitration award. As noted earlier, while we recognize the wording in

the domestic and international arbitration legislation is not identical, we see no principled reason

to disregard judicial comment and insight from cases involving international arbitrations. Several

international arbitration cases that include consideration of these issues are discussed below.

[54] Corporacion Transnacional de Inversiones SA de CV v STET International SpA (1999), 45

OR (3d) 183 (SC), [1999] OJ No 3573, aff’d (2000), 49 OR (3d) 414 (CA), leave to appeal to SCC

ref’d, [2000] SCCA No 581, involved an international commercial arbitration under the

UNCITRAL Model Law. The applicants sought to set aside the award, in part on the ground that

they were denied fundamental procedural rights because the arbitration panel failed to order the

disclosure of certain agreements contrary to Article 18 and Article 34(2)(a)(ii), which, as noted

earlier, require that a party be treated with equality and be given an opportunity to present its case.

[55] The court provided the following guidance on interpreting unfairness in the international

arbitration context, which we find applicable to domestic arbitrations under the Act:

a. “One commentator has proposed that conduct justifying intervention is ‘such a

mishandling of the arbitration as is likely to amount to some substantial miscarriage of

justice’”: at para 31, citing Mustill & Boyd, The Law and Practice of Commercial

Arbitration in England, 2nd ed (London: Butterworths, 1989) at 550, footnote 4.

b. “[I]n order to justify setting aside an award for a violation of Article 18, the conduct of

the Tribunal must be sufficiently serious to offend our most basic notions of morality and

justice”: at para 33.

c. “[J]udicial intervention for alleged violations of the due process requirements of the

Model Law will be warranted only when the Tribunal’s conduct is so serious that it

cannot be condoned under the law of the enforcing state”: at para 34.

d. “The concept of ‘fairness’ obviously entails fairness to both sides…The English Court of

Appeal… pointed out [in reference to an arbitrator’s refusal to grant an adjournment] that

any decision-maker has to balance the prejudice that might result from delay with the

duty to hear all material evidence”: at para 43.

Page: 15

e. “The purpose of Article 18 is to protect a party from egregious and injudicious conduct

by a Tribunal. It is not intended to protect a party from its own failures or strategic

choices. It was [the applicant] who deprived itself of the rights it now asserts were denied

to it by the Tribunal. This cannot and does not amount to a lack of opportunity to present

one’s case or to be treated with equality”: at para 73.

[56] In Consolidated v Ambatov, 2016 ONSC 7171, the court held that Article 34(2)(a)(ii) of

the Model Law must be interpreted to include procedural and substantive justice but the threshold

for setting aside an arbitration award remains high. The tribunal’s conduct must be sufficiently

serious to offend our basic notions of morality and justice: at para 56, citing Corporacion

Transnacional. A party may be unable to present its case when: (a) the award is based on a theory

of liability that either or both of the parties were not given an opportunity to address, or based on

a theory of the case not argued for by either of the parties; (b) a party was not given an opportunity

to respond to arguments made by an opposing party; or (c) the tribunal ignored or failed to take

the evidence or submissions of the parties into account: at para 57. Citing the Ontario Court of

Appeal in Popack (which had in turn cited the Court of Appeal of Quebec in Rhéaume), the court

noted at para 154 that “[i]t is necessary to balance the nature of the breach in the context of the

arbitral process, determine whether the breaches are of such a nature as to undermine the integrity

of the process, and assess the extent to which the breach had any bearing on the award itself”, and

concluded “[i]n the context of this arbitration as a whole, the alleged procedural shortcomings

were simply not that serious” and there seemed to be little likelihood that they “would have

affected the outcome”: at para 162.

[57] That decision was upheld by the Ontario Court of Appeal (in Consolidated Contractors),

which endorsed the lower court’s observation that it is difficult to isolate “a single alleged

procedural defect in the treatment of one out of many issues in a complex and lengthy arbitration”:

at para 76.

[58] Finally, a decision from the Singapore Court of Appeal, Soh Beng Tee & Co Pte Ltd v

Fairmount Development Pte Ltd, [2007] SGCA 28, offers helpful commentary on when an

arbitration award ought to be set aside due to a breach of natural justice:

a. “The delicate balance between ensuring the integrity of the arbitral process and ensuring

that the rules of natural justice are complied with in the arbitral process is preserved by

strictly adhering to only the narrow scope and basis for challenging an arbitral award that

has been expressly acknowledged under the Act and the [International Arbitration Act

(Cap 143A, 2002 Rev Ed)]”: at para 65.

Page: 16

b. “In short, there must be a real basis for alleging that the arbitrator has conducted the

arbitral process either irrationally or capriciously”: at para 65.

c. “It must always be borne in mind that it is not the function of the court to assiduously

comb an arbitral award microscopically in attempting to determine if there was any

blame or fault in the arbitral process; rather, an award should be read generously such

that only meaningful breaches of the rules of natural justice that have actually caused

prejudice are ultimately remedied”: at para 65.

d. “Fairness includes the right to be heard and mandates equality of treatment. Arid,

hollow, technical or procedural objections that do not prejudice any party should never

be countenanced. It is only where the alleged breach of natural justice has surpassed the

boundaries of legitimate expectation and propriety, culminating in actual prejudice to a

party, that a remedy can or should be made available”: at para 98.

Breaches said to arise from document disclosure

[59] In this appeal, the appellants say insufficient document disclosure denied them an

opportunity to present their case or respond to that of TransAlta, and that reliance by the arbitration

panel on the absence of the undisclosed records either supports that conclusion or is another basis

for finding manifest unfairness that warrants judicial intervention in the award.

[60] It has been noted that the need for relevant evidence is the same, regardless of whether the

parties are participating in an arbitration or court process: Jardine Lloyd Thompson Canada Inc v

SJO Catlin, 2006 ABCA 18 at para 42, leave to appeal to SCC ref’d, [2006] SCCA No 87.

Document disclosure can be critical to ensuring meaningful participation and a fair hearing: see

Nortel Networks Inc v Calgary (City), 2008 ABCA 370 at paras 19-20, leave to appeal to SCC

ref’d, [2008] SCCA No 533, where a lack of disclosure was found to have “effectively precluded

Nortel … from tendering evidence through its own expert witness … or demonstrating the same

through cross-examination”.

[61] However, not every evidentiary ruling about the admission of evidence “rises to the level

of procedural fairness”: Ary Homes at paras 46-47. In Corporacion Transnacional, the court noted

the applicants could have pursued several procedural avenues, including seeking production of the

redacted portions of the documents from the panel or cross-examining the respondents’ witnesses

and attempting to establish a basis for seeking an order from the panel directing production of the

redacted portions.

Page: 17

[62] In Syndicat des employés professionnels de l'Université du Québec à Trois-Rivières c

Université du Québec à Trois-Rivières, [1993] 1 SCR 471, 101 DLR (4th) 494, at issue was a

labour arbitrator’s refusal to admit certain evidence. The Supreme Court concluded that the

rejection of relevant evidence does not automatically give rise to a breach of natural justice: at 491.

The court ultimately found there was a breach of natural justice because the evidence the arbitrator

refused to admit was “prima facie crucial” to the case the respondent sought to make. Moreover,

the arbitrator had disposed of “an extremely important point in the case before him…without

having heard any evidence whatever from the respondent on the point, and even having expressly

refused to hear the evidence which the respondent sought to present on the point”: at 492 (emphasis

in original).

[63] The appellants rely on Syndicat to argue that, in determining whether the parties were

treated unfairly, the reviewing court cannot consider the result of the arbitration. In support of this

proposition, they rely on the following excerpt from Syndicat at 493:

Secondly, and more fundamentally, the rules of natural justice have enshrined

certain guarantees regarding procedure, and it is the denial of those procedural

guarantees which justifies the courts in intervening. The application of these rules

should thus not depend on speculation as to what the decision on the merits would

have been had the rights of the parties not been denied. I concur in this regard with

the view of Le Dain J., who stated in Cardinal v. Director of Kent Institution, 1985

23 (SCC), [1985] 2 S.C.R. 643, at p. 661:

. . . the denial of a right to a fair hearing must always render a

decision invalid, whether or not it may appear to a reviewing court

that the hearing would likely have resulted in a different decision.

The right to a fair hearing must be regarded as an independent,

unqualified right which finds its essential justification in the sense

of procedural justice which any person affected by an administrative

decision is entitled to have.

[64] This statement must be considered in context. The Supreme Court found a breach of natural

justice because the rejected evidence was crucial to the respondent’s case, and then held it was

inconsequential whether the evidence would have affected the result. We read Syndicat to say that

a reviewing court should not speculate on what the arbitration award would have been if rejected

evidence had been admitted; it does not follow that a court cannot look to the award and the

arbitration panel’s reasons in determining the materiality of the evidence to a party’s case and the

ultimate question of whether unfairness resulted from its absence. This is particularly so when,

like this appeal, the allegation is that the arbitration panel relied on the absence of undisclosed

Page: 18

records to support its finding in favour of the claimants. A review of the reasons and whether the

lack of evidence played a part in the conclusions reached by the arbitration panel is necessary to

answer the issues put forward by the parties on appeal.

[65] We also take note of Rhéaume, a more recent appellate decision which considered the

extent to which a breach undermined the fairness of the arbitration and the effect it had on the

award, an approach that was cited with approval by the Ontario Court of Appeal in Popack and

with which we agree.

[66] Several principles may be extracted from this review of the jurisprudence. First, parties to

an arbitration are entitled to a fair hearing, not a perfect hearing. Second, the overall fairness of

the proceedings must be considered, not individual rulings. Third, parties must take advantage of

and exercise diligence in pursuing issues and cannot later complain of some perceived unfairness

resulting from their failure to do so. Fourth, not every refusal to admit relevant evidence is a breach

of natural justice. Fifth, the threshold for setting aside an arbitration award on grounds of

unfairness has been described in various ways, but the cases make it clear that something of a

significant nature is required, or that the excluded evidence was crucial to the case being presented.

Conclusion on the interpretation of section 45(1)(f)

[67] In summary, the grounds to set aside an arbitration award should be construed in

accordance with the purposes of the Act; that is, narrowly and only to ensure the proceedings are

not fundamentally or fatally flawed. The Act was intended to circumscribe appellate review and

there is no exception for procedural rulings that do not lead to a fundamentally flawed process.

Setting aside an arbitration award because a party was “treated manifestly unfairly and unequally”

or not given an opportunity to present its case or respond to the other party’s case requires a

determination as to whether the errors alleged challenge the fundamental validity of the process.

It must go to the heart of the process and effectively undermine its fairness or have the effect of

preventing the party from putting forward its case. Where the exclusion of evidence is said to be

at the root of an unfairness, that evidence must be crucial to the party’s case. If the alleged

unfairness does not impact the proceedings in a meaningful way, it cannot rise to the level

necessary to set aside the arbitration award under section 45(1)(f).

[68] The appellants have not demonstrated that the chambers judge erred in his interpretation

of the provision.

Did the chambers judge err in finding that the criteria to set aside an arbitration award

pursuant to section 45(1)(f) were not met?

Page: 19

[69] To assess whether the procedural issues alleged by the appellants resulted in manifest

unfairness or denied the appellants an opportunity to make their case, a detailed understanding of

this arbitration is necessary. It is important to appreciate the plethora of issues before the arbitration

panel, the nature of the undisclosed records and their relevance and significance to those issues,

and the findings and determinations made by the arbitration panel. Accordingly, the background

of the dispute and the arbitration award will be discussed in some detail below.

The Power Purchase Arrangement

[70] TransAlta owned and operated a coal fired thermal electrical generating facility known as

the Keephills power plant. ENMAX is the buyer of that power.

[71] The Unit is subject to the Keephills Power Purchase Arrangement, or PPA, between

TransAlta (as owner) and ENMAX (as buyer). The PPA is made under the Power Purchase

Arrangements Determination Regulation and its terms are mandated by Regulation. The

arbitration panel noted, and the parties agreed, that the rules of statutory interpretation apply in

interpreting the language of the PPA. Under the terms of the PPA, TransAlta is to provide an

agreed upon level of electricity to ENMAX. If actual availability falls below that target, TransAlta

must pay ENMAX an “Availability Incentive Payment”. ENMAX receives the generation capacity

and pays TransAlta a regulated wholesale price for the electricity, set by the terms of the PPA, and

sells the electricity.

[72] Both parties are relieved of their obligations under the Power Purchase Arrangement in

case of force majeure, which is defined as “any event or cause which is beyond the reasonable

control of the affected party … including a HILP Event, a mechanical breakdown but only insofar

as such breakdown results from a HILP Event…”

[73] A HILP Event, or “High Impact, Low Probability Event”, is defined as:

… a major failure of some or all of the components of the Plant (or a reasonable

prediction by the Owner that a major failure of some or all of the components of

the Plant will occur before the next scheduled Planned Outage) and which results

(or could be reasonably expected to result) in the Plant being unable to operate or

being forced to operate at a lower level (or is reasonably predicted by the Owner to

be unable to operate or forced to operate at a lower level) and (a) it is reasonably

predicted by the Owner that the Plant will be unable to operate or forced to operate

at a lower level for a period in excess of six (6) weeks; and (b) the Balancing Pool

has confirmed that the above conditions have been met …

Page: 20

[74] In the case of a force majeure event, TransAlta is relieved of its obligation to make penalty

payments to ENMAX under the Availability Incentive Payment mechanism. Also, during periods

of force majeure, the obligation to make Capacity Payments to TransAlta shifts from ENMAX to

the Balancing Pool: PPA, Article 14.4.

[75] The key issues before the arbitration panel were whether the failure was an event of force

majeure and a HILP event, both as defined in the PPA. The arbitration panel noted that “any event

beyond the reasonable control of an affected party can be a matter of Force Majeure, including a

HILP, but if the Force Majeure is from a mechanical breakdown, it must be a HILP.”: award at

para 41. Accordingly, the questions to be determined included whether there was a mechanical

breakdown, whether it was a HILP event, and whether it was beyond the reasonable control of

TransAlta.

[76] TransAlta had an obligation to operate the Unit in accordance with “Good Operating

Practice”, or GOP, defined as follows:

Any of the range of practices, methods and acts engaged in or approved by a

significant proportion of the industry in North America involved in the supply of

electricity from and the operation of generating units similar to the Units, from time

to time, or any other practices, methods and acts which, in the exercise of

reasonable judgment in light of the facts known or reasonably ascertainable, could

have been expected to accomplish the desired result at a reasonable cost consistent

with applicable Laws, reliability, safety and expedition.

[77] The arbitration panel noted there are two parts to this definition: i) practices in common

use in the industry, and ii) other practices reflecting “reasonable judgment in light of the facts

known or reasonably ascertainable”. If the failure resulted from poor practice, then TransAlta

could not rely on the force majeure exception: award at para 11. Such failures would not be

“beyond the reasonable control” of TransAlta.

The “tripping” event and the positions of the parties

[78] On March 5, 2013, the Unit stator ground fault protection system tripped the Unit off-line.

It remained off-line until October 6, 2013.

[79] TransAlta retained various consultants to diagnose the problem. Some of the testing

suggested the Unit could be brought back online, but as the chambers judge summarized at para

15 of his reasons:

Page: 21

A Tennessee Valley Authority (“TVA”) probe test registered elevated localized

partial discharge activity on the generator’s top stator bar (“Bar T5”). Kinectrics

performed the TVA test with an Iris PPM97 corona probe (the “PPM97 Meter”).

On March 20, 2013, TransAlta was advised of the TVA test readings. Hitachi, the

generator’s manufacturer, was also advised and recommended that Bar T5 be

replaced.

[80] TransAlta decided to replace Bar T5 and issued a notice of force majeure on March 25,

2013. On March 27, 2013, TransAlta requested that the Balancing Pool confirm a HILP event had

occurred, as required by the PPA to establish a force majeure. The Balancing Pool did not issue

the requested confirmation.

[81] When the relay tripped, TransAlta was presented with two possible courses of conduct:

(a)

It could reset the relay and restart the generator. Any needed repairs could be done

about 14 months later at the next scheduled maintenance shutdown in May 2014.

(b)

It could take the generator off-line and replace or repair the component that had

caused the trip.

Before deciding which course to take, TransAlta needed to diagnose the problem and identify what

caused the trip. Depending on the cause, simply resetting the relay and restarting the generator

might cause further damage. On the other hand, taking the generator off-line would incur

significant financial costs.

[82] If the Unit had been restarted and returned to service at the time of the initial testing, the

outage would not have qualified as a HILP event. However, TransAlta’s decision to replace Bar

T5 was originally estimated to take six weeks and five days, which would potentially qualify it as

a HILP event. In the end, the replacement of Bar T5 disclosed other problems, leading to the

decision to conduct a full rewind of the stator. That actually took seven months, largely due to the

time required to obtain a stator rewind kit from Japan.

[83] To diagnose the problem, TransAlta conducted a number of tests and consulted with

various persons, including Hitachi, the manufacturer of the generator, and Kinectrics, a third party

expert. In its statement of position, TransAlta stated that one of the tests used was the Tennessee

Valley Authority, or TVA, test:

89 The TVA probe test is a recognized test within the North American utility

industry and TransAlta has used the test as part of its standard generator diagnostics

testing program across its fleet for many years. If there is a problem with the stator

Page: 22

bar insulation, conducting a TVA probe test is the best way to pinpoint its exact

location.

[84] The relay had not tripped previously in 30 years of operation, and diagnostic testing and

routine maintenance prior to the tripping event had not disclosed any problems, so TransAlta took

the position that the tripping of the relay in 2013 was unpredictable and unexpected. TransAlta

had “uprated” the generator’s capacity during the planned maintenance outage in 2012 but had not

rewound the stators at that time, because it was not recommended or indicated.

[85] Possible external causes of the trip or a failure of the relay itself were eliminated, indicating

that the problem was inside the generator. A number of the tests conducted did not identify any

abnormality, and the generator was dismantled for further testing. A TVA test, performed by

Kinectrics using an Iris PPM97 corona probe, suggested that the problem originated in the top T5

stator bar located inside the generator: reasons at para 15. This test showed that the T5 stator bar

was an “outlier” and was generating a “spiked” reading far in excess of the readings coming from

the other stator bars, and higher than the readings taken during the 2012 routine maintenance

outage.

[86] The Iris PPM97 corona probe was owned by Kinectrics. It subsequently became apparent

that this probe had a faulty diode, causing it to show results two or three times the actual reading.

However, this problem was not the sole explanation for the high readings. Even adjusted for that

magnification of the readings, the “spiked” readings it was reporting from the T5 stator bar were

still anomalous and significantly in excess of the readings coming from the other stator bars: award

at para 173.

[87] As noted, Hitachi and Kinectrics initially recommended replacement of the T5 stator bar.

However, when that replacement was attempted, problems were detected in the T31 stator bar and

elsewhere in the generator. TransAlta decided to take the generator off-line and do a complete

rewind of the stator. This decision was supported by Hitachi and the consultants. Hitachi warned

that there was a significant risk of a double-line-to-ground failure if the generator was returned to

service without a full stator rewind: award at para 189. This could lead to significant damage to

the generator, including damage to the stator core. TransAlta concluded that rewinding the stator

was a HILP event that would take more than six weeks to accomplish, and therefore was a force

majeure.

[88] In this situation, the HILP event was not simply the tripping of the relay, but the underlying

problem in the generator that caused the tripping. Since the generator had not actually failed,

ENMAX took the position that to qualify as a HILP event TransAlta had to demonstrate that it

Page: 23

was “a reasonable prediction by the Owner” that a major failure of some or all of the components

of the generator would occur before the next scheduled outage.

[89] In its Statement of Position, ENMAX argued that TransAlta was not entitled to declare

force majeure because its investigation of the relay trip was flawed and inadequate and failed to

meet Good Operating Practice. It alleged that TransAlta did not do sufficient testing to eliminate

the prospect of simply restarting the generator. Hitachi’s recommendation to rewind the stators

was unreasonable. There was accordingly no “reasonable prediction by the Owner” of a major

failure before the next scheduled outage in May 2014.

[90] ENMAX took the position that the TVA test was not widely recognized within the North

American utility industry as a reliable primary diagnostic test, but only as a secondary diagnostic

test to pinpoint the location of potential problems. Hitachi was not familiar with the TVA test and

had recommended other testing be done before the generator was dismantled. Given that other

tests were normal, ENMAX argued it was unreasonable for TransAlta to rely on the TVA test in

deciding to rewind the stator. If TransAlta’s prediction of imminent failure was not reasonable,

there was no HILP event, and therefore no event of force majeure.

[91] In its Statement of Position, the Balancing Pool argued that whatever the cause of the

shutdown of the generator, it was the result of poor operational decisions by TransAlta and was

entirely under TransAlta’s control. The aging of equipment is not unpredictable. As a result, the

shutdown could not qualify as a HILP event. TransAlta should have rewound the stator in 2012

when it uprated the generator’s capacity. At the least, it should have had a spare stator or rewind

kit on hand, which would have significantly reduced the length of the shutdown.

[92] The Balancing Pool stated that, instead of following the advice to return the generator to

service, TransAlta unreasonably relied on the “erratic and inconsistent” and uncorroborated TVA

test. After-the-fact testing failed to show any defect that required a complete shutdown. There was

nothing to show that the generator could not be restarted after the trip or that a full or partial rewind

of the stator was required. The Balancing Pool argued that, at best, TransAlta should have only

replaced bars T5 and T31, and then returned the generator to service until the next planned service

outage, or until the rewind kit arrived from Hitachi.

[93] In its Reply, TransAlta responded that the tripping event was unexpected. Testing done

during the scheduled 2012 outage had indicated that the insulation was in good condition. Hitachi

had never recommended a stator rewind. TransAlta kept spare stator bars, but it was not standard

industry practice to maintain a full spare stator rewind kit. TransAlta did not rely just on the TVA

test, but concluded that there was a predictable failure before the next scheduled maintenance

outage based on (i) the trip itself, (ii) anomalies in the data, (iii) “worrisome physical observations”

Page: 24

(a burning smell and high carbon monoxide reading), (iv) other concerning partial discharge test

results, (v) the recommendation of Hitachi and Kinectrics, (vi) elimination of external causes for

the trip, together with (vii) the TVA test. The TVA test is widely recognized and reliable.

Subsequent testing showed that the T5 bar was unfit for further service. The Balancing Pool and

ENMAX were taking inconsistent positions by arguing that a stator rewind was required in 2012

but not in 2013 after the tripping event.

[94] In addition to their argument that the TVA test was not used in the industry to diagnose

problems of this type, the appellants wanted to establish that TransAlta knew or ought to have

known that the Iris PPM97 corona probe was unreliable and could give “false positives”.

TransAlta’s reliance on the TVA test as a justification for replacing the Bar T5 was itself

unreasonable, but TransAlta should also have known that this particular probe was unreliable.

Instead of taking the Unit off-line, TransAlta should have searched for and eliminated the

possibility of a different diagnosis or solution.

[95] The arbitration panel noted that there was “a series of occurrences or events that

collectively form the Event as defined by the parties”, with the key issue being “whether at each

or any decision point during the outage” it was reasonable for TransAlta to predict that a major

failure would occur prior to the next scheduled planned outage: award at para 38. An underlying

issue was whether TransAlta’s decision to rewind the stator was reasonable or whether the relay

should have been reset and the generator restarted.

The arbitration process and procedural orders

[96] Both appellants disputed the claim of force majeure and a demand for arbitration was

issued under the PPA. The arbitration panel made four procedural orders dealing with scheduling,

document disclosure, production of additional records, requests of further production and removal

of redactions. There were also email directions respecting the conduct of the proceedings. The

procedural orders were made between May 1, 2015 and March 3, 2016, and the hearing took place

in May 2016. Witness statements and expert reports were filed by the parties and tendered as their

evidence. The witnesses were also examined and cross-examined at the hearing. Closing written

submissions were received in June 2016 and the follow up written questions from the arbitration

panel were received and responded to by the parties by way of written reply dated Friday, July 22,

2016.

[97] Issues related to record production came before the arbitration panel in a series of

procedural applications.

Procedural Order No. 1

Page: 25

[98] Procedural Order No. 1 set out the general procedures to be followed in the arbitration,

including record production. Initially, all the parties supported a type of record production similar

to the one used in civil litigation: all records that were relevant and material to the outcome would

be disclosed: reasons at para 21.

[99] The arbitration panel directed a method of “reliance and request” disclosure. Initially, a

party would only disclose those records on which it intended to rely, but the other party could

request production of “additional documents that are relevant to an issue and material to its

outcome believed to be in the possession of or under the control of an adverse party”: reasons at

paras 22-23. Record production was essentially to be done in two stages, with the ultimate

obligation being to disclose all relevant and material records: reasons at para 28. Procedural Order

No. 1 confirmed that it was not comprehensive or final, and that the parties could apply for further

directions or changes: reasons at para 27. The appellants subsequently made a number of

applications for production.

Procedural Order No. 2

[100] Requests for further disclosure resulted in Procedural Order No. 2. The arbitration panel

outlined its general approach to production (Procedural Order No. 2 at para 5):

Without the benefit of engineering evidence, it is difficult for the tribunal to easily

determine proportionality and the scope of the discovery process on this issue. As

a general principle, however, the tribunal approaches production on the basis that

the closer the request is to the stator or the generator protection systems, the more

it is clear that the documents should be produced on the basis that they are relevant

to an issue and material to its outcome. While recognizing that the pleadings

disclose a number of other potential causes for the Trip, the further away one gets

from what was identified at the time as the problem with the generator, the more

we would need to see some evidence of relevance before those documents are

ordered produced. This may well mean that our decision in this Order may be

revisited as further evidence is disclosed in the production process. It will also be

noted that in a number of the items we believe it appropriate that experts from the

parties meet and confer with a view to establishing the reasonable limits of some

of the productions. If no agreement is reached a further motion may be brought to

the tribunal supported with appropriate evidence. [emphasis added]

[101] The arbitration panel made it clear that it was prepared to revisit the issue if further

evidence was received and acknowledged that the issue was still outstanding: reasons at para 38.

Page: 26

Dealing specifically with claims of privilege, and redactions of produced records based on

irrelevance, the arbitration panel directed (Procedural Order No. 2, para 17):

Having considered this Order, if a claim for privilege is still maintained, or there

continues to be a redaction based on irrelevance, the requesting party has three

options: It may 1) accept the characterization by counsel, 2) ask the tribunal to

review the challenged document and decide the issue, or 3) the parties can agree to

an independent lawyer being appointed, either the parties or the tribunal, who will

review and decide the issue. It should be mentioned that the tribunal is not in favour

of option 2.

[102] Of importance to these appeals was ENMAX’s Request #71:

Copies of all records related to or regarding the use of TVA probe testing for

diagnostic purposes across TAGP’s fleet, including: (a) any formal or documented

protocols or procedures; (b) any test reports or results of testing; (c) any reports,

memorandums, recommendations or decisions that contemplate or consider the

outcomes of TVA testing results; and (d) any reports, memorandums,

recommendations or decisions that consider or contemplate the merit or efficacy of

TVA probe testing, whether produced or prepared by TAGP, Kinectrics or any

other 3rd party. [emphasis added]

[103] ENMAX argued that this information was relevant to whether the procedure used when

testing the stator accorded with Good Operating Practice and whether, given TransAlta’s

experience with the TVA test, reliance on that test was reasonable given contra-indications

provided by other diagnostic tests. TransAlta replied that it had provided all these records with

respect to the Keephills generator and that records relating to other facilities were not relevant and

material.

[104] In addition, ENMAX wanted production of records held by Kinectrics and other

contractors under TransAlta’s control.

[105] Counsel for ENMAX advised the arbitration panel that the parties were attempting to

resolve outstanding items, including ENMAX’s Request #71, and “if we can’t, we will bring them

back to you for resolution”. The arbitration panel never specifically dealt with ENMAX’s

Request #71. The arbitration panel noted that “ENMAX and TransAlta were still attempting to

resolve” it and other requests, and therefore “No Order necessary”: reasons at para 38.

Page: 27

Procedural Order No. 3

[106] Another request for production was made a few months later, resulting in Procedural Order

No. 3. Still said to be outstanding were records related to “whether TransAlta’s extensive reliance

on TVA testing during the Event accords with good operating practice”. ENMAX’s position was

that TransAlta’s disclosure revealed “the Corona testing instrument PPM 97 contained

manufacturing defects that resulted in it returning inconsistent and unreliable readings”, and it

sought information respecting the calibration of the probe. No specific reference was made to

ENMAX’s outstanding Request #71.

[107] The Balancing Pool requested further record production, based on issues that had been

revealed by what TransAlta had produced to date. It stated that the reliability of the TVA test was

a key issue, and it also requested records regarding the inaccurate readings given by the probe,

together with detailed records about the testing methodology used.

[108] TransAlta repeated its position that record production should be limited to what happened

with the Unit and that “fleet wide” records were not relevant. TransAlta relied on the test of “the

closer the request is to the stator or the generator protection systems” set in Procedural Order No.

2 and argued that the appellants were on a “fishing expedition”.

[109] Procedural Order No. 3 reiterated the obligations of each party to disclose records and

confirmed that the Reliance and Request method of disclosure was not intended to limit the

ultimate scope of disclosure: all relevant and material records under the control of each party were

to be disclosed.

[110] The arbitration panel directed TransAlta’s counsel to expressly confirm they had made

TransAlta aware of its disclosure obligations, that diligent searches had been conducted, and that:

“. . . all documents or categories of documents relied on, agreed to be produced, or ordered

produced have been produced”: reasons at para 52.

Procedural Order No. 4

[111] The Balancing Pool made a further application for the production of records before the

commencement of the hearing, including “relevant and material records in the possession of

TransAlta and Kinectrics” relating to the TVA probe and its use at Keephills. A particular focus

was on a redacted October 1, 2013, email from Andrew Brown, an employee of Kinectrics, that

disclosed Brown had discovered the Iris PPM97 corona probe was not properly calibrated and

gave inaccurate readings.

Page: 28

[112] Following Procedural Order No. 3, TransAlta asked Brown if he had any further records,

but he replied that he was no longer employed by Kinectrics. TransAlta then asked Kinectrics if it

had any records about the use of the probe. Kinectrics found three documents, but they had already

been produced. The Balancing Pool asked for information about the scope of the request made to

Kinectrics and why no other documents could be located. TransAlta replied that Kinectrics was

consulting its counsel, prompting the Balancing Pool to ask the arbitration panel to issue directions

(i) that TransAlta produce an unredacted copy of the Brown email, and (ii) that Kinectrics itself

produce relevant and material records in specific categories.

[113] The records or categories of records being sought related to TVA test records at the time

of the 2013 trip and records about the Iris PPM97 corona probe. The Balancing Pool argued that

while the arbitration panel may not have jurisdiction over third parties like Kinectrics it could seek

the assistance of the court and requested a declaration that the Kinectrics records were relevant

and material. If Kinectrics refused to voluntarily produce them, then the Balancing Pool would “be

at liberty to apply to the Court of Queen's Bench of Alberta seeking that court's assistance to

enforce the Tribunal’s direction”.

[114] The Brown email redacted information about other TransAlta units where the Iris PPM97

corona probe had been used. The appellants had had a copy of this redacted email for some time

but did not immediately ask for an unredacted copy.

[115] This application resulted in Procedural Order No. 4, which confirmed the arbitration

panel’s premise that each party had an obligation to produce records that were relevant to an issue

and material to its outcome believed to be in the possession of or under the control of an adverse

party. The arbitration panel directed that TransAlta provide an unredacted copy of the Brown

email, as it met this test.

[116] The arbitration panel noted that the appellants had made repeated requests for production

of all records relating to the TVA test and the calibration of the Iris PPM97 corona probe.

TransAlta had agreed to produce all of this category of records in its possession, so the arbitration

panel did not find it necessary to make any further order. The arbitration panel rejected the request

for a declaration that records held by Kinectrics were relevant and material:

11 As TransAlta has apparently agreed that these records should be produced, the

tribunal does not consider it necessary or appropriate to make any further order

respecting the necessity for production from a third party, or to consider the extent

to which records in the possession of Kinectrics are relevant and material to the

issues beyond that which has already been produced. We are also not prepared to

make any order with respect to scope or relevancy that might directly or indirectly

Page: 29

affect the rights of a third party, not resident in Alberta, who presumably was not

given notice of this motion and who has not made any submissions to us.

12 This decision is without prejudice to the Balancing Pool to pursue this issue

elsewhere as it may see fit, keeping in mind the schedule in place for the hearing of

this case.

As a result of Procedural Order No. 4, TransAlta produced an unredacted copy of the Brown email,

which reported that the “Iris PPM97 meter reads high, dependent upon the range selected…”.

[117] TransAlta subsequently provided surreply evidence from Brown on its use of the Iris

PPM97 corona probe: reasons at para 65. In its written argument following the hearing, the

Balancing Pool observed:

208 Having vigorously resisted production of Kinectrics records on the grounds

that it cannot give what it does not have, TransAlta then produced a surreply

statement from a witness whose evidence is based on alleged test results and other

facts that he cannot have had access to since he left Kinectrics in January, 2014.

Paragraph 6 of Mr. Brown's Surreply asks the Tribunal to accept as fact four

detailed characteristics of the TVA tests without producing the underlying test

results, or indeed any records in support. Paragraph 7 purports to list facts that

differentiate Keephills Unit 1 from the other three Units based on information in

TransAlta’s possession. If the statements made in TransAlta’s February 22 Brief

[respecting Procedural Order No. 4] are accurate, Mr. Brown’s surreply is either

made entirely from memory, or it is improperly based on his review of TransAlta

records that have not been produced in this dispute. If it is the former, it is entitled

to no weight. If it is the latter, it is improper and the test results and documentation

that Mr. Brown relied upon should have been produced.

[118] The appellants made no application for further production between the receipt of the

surreply evidence on April 18, 2016 and the commencement of the hearing on May 9, 2016 or

during the hearing itself. In the end, the arbitration panel accepted the Balancing Pool’s argument

and gave little weight to Brown’s evidence: award at para 162.

The arbitration award

[119] The arbitration panel issued its award on November 14, 2016. It found that TransAlta had

acted in accordance with Good Operating Practice, that there was a HILP event and that it was

beyond the reasonable control of TransAlta and constituted a force majeure.

Page: 30

[120] As noted earlier, any event beyond the reasonable control of an affected party can be a

matter of force majeure, including a HILP event, but if the event is from a mechanical breakdown,

it must be a HILP event to constitute a force majeure. The key issues to be resolved by the

arbitration panel were, therefore, whether the shutdown of the Unit was a HILP event and

constituted force majeure. The panel described those issues as follows:

38 In this case, there is a series of occurrences or events that collectively form the

Event as defined by the parties. The key issue in the definition of a HILP is whether

at each or any decision point during the outage, there was “a major failure of some

or all of the components of the Plant”, or “… a reasonable prediction by the Owner

that a major failure of some … of the components of the Plant will occur before the

next scheduled Planned Outage and which … could be reasonably expected to

result in the Plant being unable to operate or be forced to operate at a lower level

… and it is reasonably predicted by the Owner that the Plant will be unable to

operate or be forced to operate at a lower level for a period in excess of six (6)

weeks.”

39 ENMAX puts the second part of the definition this way in its post-hearing Reply

submission: “based on the information that TransAlta knew, or reasonably ought to

have known, at the time, was it reasonable for TransAlta to predict that a major

failure would occur prior to the next Planned Outage (May 2014).”

[emphasis added]

[121] In approaching these issues, the arbitration panel considered whether there was a

mechanical breakdown, whether the event was beyond the reasonable control of TransAlta, and

whether the activities carried out by TransAlta, including testing, constituted Good Operating

Practice. These issues were considered at each decision point, and the award was structured

accordingly.

[122] The focus of the arbitration was, therefore, on what happened, why it happened, could it

have been prevented, what did TransAlta know, what choices did TransAlta have and what choices

did it make, and what information did TransAlta choose to rely on or fail to rely on, all considered

at each decision stage. The structure of the award highlights, in chronological order, each of the

sub-issues and related facts that went in to determining whether the outage was a matter of force

majeure. When required, the arbitration panel found facts and made determinations regarding

whether each step was reasonable and within Good Operating Practice. Finally, the arbitration

panel summarized its findings and discussed them in relation to the PPA definitions of force

majeure and HILP event.

Page: 31

[123] The arbitration panel explored the trip and its initial investigation. It found that the DGSH

relay and related wiring was functioning properly at the time of the trip, and the testing done to

confirm that was reasonable: award at paras 66-67. It then considered the theory posited for the

trip and found that it was triggered by a change in the third harmonic voltage ratio caused by an

issue with the generator stator winding: award at para 77.

[124] The arbitration panel described its key findings and determinations on each issue as

follows, at award para 254:

a. Operating the Unit without dehumidifiers prior to the 2012 uprate did not cause any

detectible problem with the stator insulation in 2013, or cause or contribute to the Trip.

b. The investigation of the DGSH relay was appropriate and confirmed it was working

properly.

c. The decision to dry out the stator fell within GOP and was reasonable.

d. The decision to pull the rotor fell within GOP and was reasonable.

e. There was no good reason to exclude the TVA testing.

f. At the end of the investigation stage all of the evidence pointed to an aged insulation

system with Bar T5 being the suspected cause of the Trip.

g. At the time the decision was made to replace Bar T5, TransAlta knew:

i. A Trip had occurred that was in all likelihood caused by an issue in the generator;

ii. Very high TVA readings were found on one bar that increased when the wedges

were removed evidencing delamination or a breakdown in the insulation, caused by

heat and mechanical forces on the bar when the generator was operating;

iii. Confirmation of the TVA test was obtained from a partial discharge test done at

the neutral end, phase A, showing two times the reading at the line end, and on the

neutral of the other two phases on the Unit. This indicated there was an issue at the

neutral end, phase A;

iv. It had the opinions from both Hitachi and Kinectrics that Bar T5 needed to be

replaced, or it was likely the relay would trip again.

Page: 32

h. TransAlta reviewed and considered the recommendations before coming to a conclusion

to replace Bar T5.

i. Based on the facts available in March, TransAlta did not have sufficient information to

question the reliability of the test results or the opinions of Hitachi and Kinectrics.

j. In the circumstances it was reasonable for TransAlta to accept the results and the

recommendations of Hitachi and Kinetrics and replace Bar T5.

k. The replacement of Bar T5 was expected to return the Unit to operation by April 21,

2013.

l. Based on the information available at the time, the decision to replace Bar T5 was

reasonable and fell within the definition of GOP.

m. The decision to replace T31 was reasonable.

n. The decision not to order a full rewind kit until after it was determined that the

replacement of Bar T5 had not solved the problem, but had made matters worse, but was

still reasonable.

o. Given the test data and the opinions of Hitachi, Kinectrics and Mr. Klempner, it was

reasonable for TransAlta to have decided to wait for the full stator rewind before returning

the Unit to service.

p. The Panel has no evidence that the 15 weeks it took for delivery of the full rewind kit

was unreasonable.

q. The failure to have a full rewind kit on hand was not unreasonable or contrary to GOP.

r. There is no evidence of any significant delay in returning the Unit to service.

[125] The award painstakingly went through all the challenges to the decision-making process

involved in the failure. Every decision at every stage was challenged by the appellants. One of

those decisions was TransAlta’s reliance on TVA testing, a sub-issue to the reasonableness of the

decision to replace Bar T5. It is this sub-issue that underlies the appellants’ argument regarding

unfairness. Before the arbitration panel, the appellants took the position that TransAlta ought not

to have relied on those tests because (i) the use of the TVA test in this context was not common in

the industry and (ii) from its previous fleet-wide experience, TransAlta should have known that

the testing was unreliable. A related issue involved a faulty meter used in the March TVA tests,

Page: 33

something which arose in an October 1, 2013, email from Brown and was the subject of his

surreply evidence.

[126] The arbitration panel addressed the use of the TVA tests in its award. It acknowledged

evidence that the TVA test is not used very much in North America but found that “it is an accepted

test to determine the extent to which partial discharge and thus delamination is occurring in

individual bars”: award at para 100. The arbitration panel accepted that the TVA test may not be

used to determine fitness for service, but concluded, based on the evidence of the experts, that it

is not inappropriate to use the test as part of an investigation to try to determine the cause of the

trip and to gather information on individual bars: award at paras 106-107. The arbitration panel

concluded that, “[t]here was no good reason to exclude the TVA testing. Saying it is not part of

the usual suite of tests used to determine a return to service is too narrow a view of what was

occurring”, and that the inclusion of TVA testing at this point of the investigation of the cause of

the trip fell within the definition of Good Operating Practice: award at paras 109-110.

[127] Abnormally high TVA test readings had been detected at other TransAlta facilities, but

those facilities had not been taken out of service. The appellants argued this showed that the

decision to take the Unit out of service based on the TVA test was unreasonable. TransAlta’s

witnesses were vague on the details but recalled that the other readings were taken during routine

maintenance, not in response to a “trip”, so presumably were not of the same level of concern, and

other testing minimized their importance: award at paras 158-163.

[128] With respect to the calibration problem with the meter, the arbitration panel found that

TransAlta had no reason to question the reliability of the readings at the time the decision was

made to replace Bar T5: award at paras 158, 169-172, 176, 179. Most importantly, the arbitration

panel found that the mis-calibration of the probe did not have any effect on the reasonableness of

TransAlta’s decision. Even if corrected for the miscalibration, the readings from Bar T5 were still

an “outlier” and significantly different from readings on the other stator bars. The arbitration panel

said, at para 173 of the award, “regardless of any inaccuracy in the meter used, … Bar T5 was

showing significantly different readings.”

[129] The arbitration panel noted that, at the time the decision was made to replace Bar T5,

TransAlta knew that very high TVA readings were found on one bar, had confirmation of the

concerns raised by the TVA test through a partial discharge test, and had recommendations from

both Hitachi and Kinectrics that Bar T5 needed to be replaced or it was likely the relay would trip

again.

[130] The arbitration panel’s ultimate conclusion with respect to the replacement of Bar T5 was

that, based on the facts known to TransAlta in March and the recommendations of its consultants,

Page: 34

the decision to replace Bar T5 was reasonable and within Good Operating Practice. The decision

fit the definition of GOP as “the exercise of reasonable judgment in light of the facts known or

reasonably ascertainable”: award at paras 176-179.

[131] It is important to place the argument regarding TVA testing within the entire context of the

arbitration. In context, it was a minor issue. One of the 18 findings and determinations by the

arbitration panel was that “there was no good reason to exclude the TVA testing”. The TVA testing

issue related primarily to the decision to replace Bar T5. By the time the decision was made to

conduct a complete rewind, the TVA test had been repeated using different probes, and significant

other testing had been conducted: award at paras 188-189.

[132] The arbitration panel also reviewed the decision not to return to service until the rewind

was complete, the timing of the decision to order a full rewind kit, the decision not to have a spare

rewind kit on hand, and the overall length of time it took to complete each step. On each point, the

arbitration panel outlined the evidence, set out the differing perspectives and made fact findings

and determinations, some of which were set out previously. At each point, the arbitration panel

found that the decisions taken were reasonable and not contrary to Good Operating Practice.

[133] The arbitration panel concluded there was a HILP event, noting that the Unit itself had

experienced a major failure. The trip was not caused by any external factor, and the Unit was

neither running nor generating electricity due to something internal to the Unit. Test results

indicated significant partial discharges for Bar T5 which, coupled with increased PD readings,

indicated deterioration of the insulation on Bar T5, which would lead at least to a further trip or

worse. The decision to replace Bar T5 was reasonable, and the estimate of time it would take to do

so made it reasonable for TransAlta to predict the Unit would remain off-line for more than 6

weeks, thus qualifying as a HILP event: award at paras 257-262.

[134] Extensive testing carried out after Bar T5 was replaced disclosed a larger problem. Due to

the deterioration of the insulation system, a further trip, or actual ground fault, or worse, would

occur prior to the next planned outage: award at paras 264-265. The HILP event had gone from a

trip caused by a single bar to a predicted failure of the insulation system generally: award at para

267.

[135] On the issue of force majeure, the arbitration panel found that the fact the Unit was tripped

was beyond the reasonable control of TransAlta, giving several reasons for that conclusion (award

at para 269), as was the length of time it took to investigate the problem and carry out repairs

(award at para 270). Accordingly, the trip was a HILP event and constituted a force majeure: award

at para 271.

Page: 35

Do the procedural issues alleged by the appellants rise to the level of manifest unfairness?

[136] There are three aspects to the assessment of whether the procedures adopted by the

arbitration panel as they pertain to the undisclosed records resulted in manifest unfairness to the

appellants or deprived them of an opportunity to present their case:

1. Did the procedures adopted repeatedly deny the appellants access to the undisclosed

records, as the appellants allege?

2. To what extent were the undisclosed records crucial to the issues before the arbitration

panel and to the appellants’ ability to put forward their case or respond to that of TransAlta?

3. Did the arbitration panel’s comments regarding the TVA testing elevate the importance

of the undisclosed records?

1. Did the procedures adopted deny the appellants’ access to the undisclosed records?

The reliance and request standard for document production

[137] The appellants take issue with the methodology adopted by the arbitration panel with

respect to document production. The parties had proposed a modified form of the document

production process used in civil litigation, whereby parties produce the documents in their

possession that are relevant and material. However, the arbitration panel determined that a

“reliance and request” approach was more appropriate. In that approach, each party produces the

documents it proposes to rely upon, and the other parties can then request other relevant and

material records. One of the potential flaws identified by the appellants before the chambers judge

was that in this case all the relevant documents were in the hands of TransAlta and not the

appellants.

[138] The chambers judge correctly noted that the reliance and request method is commonly used

in arbitrations¹and is not inherently an unfair procedure. He also noted that it was entirely open to

the arbitration panel to adopt this method of production and that section 20 of the Act grants the

arbitration panel discretion in determining the procedures to be followed. We agree with the

chambers judge’s holding that the arbitration panel’s adoption of this process was not unfair, much

less manifestly so, nor did it deprive the appellants of the opportunity to present their case.

[139] We cannot agree, however, with the comments of the chambers judge that the reliance and

request process was an “unfortunate choice” (reasons at para 109), seemingly suggesting that as a

¹see J Brian Casey, Arbitration Law of Canada: Practice and Procedure, 3^rded (New York: Juris, 2017) at 260

Page: 36

method of production it is of a lower standard. Those comments are, in our view, a

mischaracterization of the procedure. The underlying purpose and fundamental obligation of the

“reliance and request method” and the method of document production generally used in civil

litigation are the same – each party is to produce relevant and material records in its possession.

The only difference is timing. The obligation to produce arises at a different time, usually upon

the delivery of a series of documents that give rise to a request for further documents. Regardless

of which method is employed, the system hinges on the integrity and diligence of the litigants in

ensuring full disclosure is made, as well as the obligation of counsel to advise clients of their

disclosure obligations: see Demb v Valhalla Group Ltd, 2017 ABCA 340 at para 4; Catalyst

Partners Inc v Meridian Packaging Ltd, 2007 ABCA 201 at para 5; Kent v Martin, 2018 ABCA

202 at para 32. We take a similar view of the chambers judge’s criticism of the arbitration panel’s

decision in Procedural Order No. 3 (see reasons at para 51), which criticism also seems to flow

from his concerns with the reliance and request process.

[140] In a related argument, the appellants submit it was unreasonable and unfair for the

arbitration panel to insist that the appellants establish some degree of relevance for its document

requests, a requirement that seems to be part of the reliance and request process. A review of the

procedural orders shows that the arbitration panel’s initially narrow description of relevant

documents as being “closer to the stator” expanded as more information and evidence came to

light. As the chambers judge noted, the arbitration panel acknowledged at the outset that the scope

of discovery would change over time as it came to understand the nature of the documents being

requested, saying in Procedural Order No. 2:

Without the benefit of engineering evidence it is difficult for the tribunal to easily

determine proportionality and the scope of the discovery process on this issue.

...

... our decision in this Order may be revisited as further evidence is disclosed in the

production process.

[141] As the chambers judge found, at paras 125-126 of the reasons:

As the process unfolded the Panel became increasingly informed. While it did not

accede fully to the requests for Information, it did accede in part and it never

precluded the Applicants from succeeding more fully at a later date, either prior to

or during the hearing.

Page: 37

... it was not unfair for the Panel to deal with production issues in a measured

fashion, on a graduated basis, taking care to keep the scope of the entire matter

manageable and focused, not distracted by possible collateral facts.

[142] When the process was complete, the parties were apparently satisfied with the level of

knowledge they had obtained. The Balancing Pool acknowledged as much in its final written

argument, saying that it “was satisfied with the state of the evidence following review of the

unredacted email [of Brown] and did not pursue document requests of Kinectrics in other forums”.

[143] As noted earlier, the Act contemplates that the arbitration panel will determine the

procedures to be followed, and those procedural decisions are accorded deference: section 20(1).

This is not a license to proceed in an unfair manner; the Act also requires that arbitrators treat the

parties equally and fairly and give them an opportunity to present their case and respond to that of

the other parties: section 19. The standard, however, is not one of perfection. “The duty of fairness

is flexible and variable, and depends on an appreciation of the context of the particular statute”:

Baker v Canada (Minister of Citizenship and Immigration), [1999] 2 SCR 817, 174 DLR (4^th) 193

at para 22. In the arbitral context, the inclusion of the word “manifestly” in the governing statute

“demonstrates a legislative intent that not every potential unfairness in an arbitration proceeding

or decision will permit the Court to set aside an arbitration award”: Mitchell at paras 17-18.

[144] The document production process employed by the arbitration panel here resulted in

several distinct procedural rulings and repeated requests for additional documents, but it was not

unfair. It certainly did not result in a level of manifest unfairness that goes to the validity of the

arbitration, as required by section 45.

The procedural orders

[145] The appellants submit that, in the series of procedural orders issued by the arbitration panel,

they were repeatedly denied production of the disputed records, and that this in turn denied the

appellants the opportunity to present their case or respond to that of TransAlta.

[146] A review of the procedural orders demonstrates, as found by the chambers judge, that it is

inaccurate to say that the arbitration panel’s directions repeatedly denied the appellants the

undisclosed records, or that the arbitration panel ever closed the door to the ability of the appellants

to obtain those documents.

[147] Procedural Order No. 1 set out general procedures for the arbitration, including the issues

to be resolved and the document production standard. It contemplated that each party would

provide all documents relied upon by that party with its statement of position, following which

parties would be entitled to request the production of further records that are “relevant to an issue

Page: 38

and material to its outcome”. It also stated that the order “is not comprehensive as to the procedure

to be followed for the Arbitration. The Parties are at liberty to apply to the Panel at any time for

further directions…”.

[148] The appellants requested that TransAlta produce additional information, including the

results of TVA tests, largely in an effort to challenge their reliability and TransAlta’s decision to

rely on them in the decision to replace Bar T5. The chambers judge found that fleet-wide TVA test

records were not requested at that time, as the parties continued to negotiate a number of document

requests: reasons at para 32.

[149] As one of many lines of inquiry, ENMAX sought information related to TVA probe testing

elsewhere by TransAlta (in its Request #71). TransAlta noted that it had produced records

regarding its procedure and practice of using the TVA probe test, referred to its reliance on “third

party industry experts and other publicly available documents” for guidance on TVA testing, and

redacted other information said not to be relevant. In Procedural Order No. 2, the arbitration panel

set out its general approach to relevance (at that point in the proceedings) as being “the closer the

request is to the stator or the generator protection systems, the more clear it is that the documents

should be produced” as being relevant to an issue and material to its outcome. The arbitration panel

also noted that ENMAX and TransAlta continued to attempt to resolve some requests, including

Request #71. No order for production was made, but the arbitration panel left the door open if

further negotiations on document production proved unsuccessful.

[150] The appellants continued to push for additional document disclosure on several issues.

They sought broader production to deal with issues that went beyond the cause of the trip,

including information that related to when TransAlta might have become aware of the suspect

reliability of the TVA tests. TransAlta resisted on the basis that the information went beyond what

was considered relevant according to Procedural Order No. 2, that the appellants were on a fishing

expedition, and that it would be prejudicial to expand the scope of production at this point in the

proceedings.

[151] In Procedural Order No. 3, the arbitration panel said it saw no prejudice to additional

document requests and reminded the parties that they had a fundamental legal obligation to

disclose all records upon which they rely and to produce all relevant and material records requested

by the opposite party. It also required that counsel expressly confirm that they make their clients

aware of this obligation. The arbitration panel also indicated that the appellants could choose to

leave the issue undecided until the oral hearing and raise it again then: reasons at para 114.

[152] At this point, although ENMAX sought additional documents, it had stopped seeking

access to fleetwide TVA testing records. The Balancing Pool, however, made a further attempt for

Page: 39

production of unredacted non-TVA related documents, as well as an “unredacted” copy of the

Brown email about TVA tests at other TransAlta sites in 2012/2013. The email was said to indicate

that TransAlta had received aberrant TVA readings at other units but did not initiate repairs at

those other sites. In addition, the Balancing Pool sought a direction that Kinectrics records,

including records relating to calibration issues with the TVA probe, were relevant and material

and ought to be produced, to assist the Balancing Pool in applying to court for assistance in

directing Kinectrics to provide documents.

[153] In Procedural Order No. 4, the arbitration panel required TransAlta to produce an

unredacted copy of the Brown email. It noted that TransAlta had advised it if any further

calibration records it would produce them. The arbitration panel determined that it could not order

third party disclosure from Kinectrics, particularly in its absence, and would not consider the extent

to which those documents might be relevant and material beyond what was already produced. It

advised that the Balancing Pool was at liberty to make a court application for those documents and

that the arbitration panel could provide a certified copy of the order if required. No such court

application was made.

[154] The unredacted Brown email noted that “the Iris PPM97 meter reads high, dependent upon

the range selected, and the readings are non-linear”, and provided a summary of the test results at

“all TransAlta sites where the Iris PPM97 meter had been used since its purchase in March 2012.”

[155] Shortly before the start of the hearing, the parties exchanged surreply evidence on a number

of issues. TransAlta’s surreply, from Brown, focused on the issues arising from the Brown email

and the use of the Iris TVA probe. The surreply offered reasons that TransAlta may have responded

differently to the faulty TVA results at other units versus the March 2013 results for the Unit.

[156] As the matter proceeded, therefore, the appellants knew about other test experiences,

although it was admittedly late in the day. However, at that point they chose not to seek further

production.

[157] It is apparent that the arbitration panel was of the view that document production was never

denied, and that the possibility remained for further production as the hearing progressed. At para

164 of the award, the arbitration panel noted:

In his cross-examination, Mr Brown was taken to the October email that lists the

other sites, but no questions were asked about the readings at the other sites, nor

was he asked about his explanations set out in his surreply witness statement.

Instead, Respondents say his evidence on this point offends the best evidence rule

and should be discarded. They say that evidence about the other sites should have

Page: 40

come from TransAlta. No documentation or reports regarding TVA test results

from the other sites were produced at the hearing by TransAlta, nor did the

Respondents seek further production.

[158] This is borne out by the Balancing Pool’s acknowledgement in its final argument that:

The Balancing Pool was satisfied with the state of the evidence following review

of the un-redacted email and did not pursue document requests of Kinectrics in

other forums.

[159] The chambers judge found this to be the case on the record. He pointed out that the

arbitration panel never ruled that the information was irrelevant, never denied production of the

records, and never foreclosed the possibility of further production. Rather, the request was stood

down. He noted that the arbitration panel “demonstrated its will to allow deadlines to be exceeded,

to allow surreply evidence, to allow a new witness from ENMAX very late in the process”: reasons

at para 134. He noted the arbitration panel’s comment that the appellants opted not to cross-

examine Mr Brown about TVA testing at other sites and held “[t]he Panel did not act offside

section 45(1)(f) of the Arbitration Act in expecting the Applicants’ position about TVA testing at

other sites to be put to the one witness who undertook such testing”: reasons at para 131.

[160] It is a fair reflection of the record and the findings of the chambers judge to conclude that

the request for fleet wide records was never definitively determined and the prospect of those

records becoming relevant and material, and therefore producible, remained open. When directly

asked, the arbitration panel ordered production of an unredacted version of the Brown email, but

no further application for additional records were made, either before the arbitration panel or

before the court.

[161] The chambers judge appears to have taken the view that the fleetwide TVA tests might

have become relevant as a result of the Brown surreply, but noted that the appellants did not request

an adjournment or any other procedural remedy to deal with that evidence and the Kinectrics

records. The decision not to cross-examine Brown on his surreply evidence was said to have been

made because that evidence was unsupported recollections and should not be accorded any weight.

The arbitration panel seems to have agreed and gave little weight to the general observations of

witnesses regarding why there might be a different response to high TVA readings at different

sites.

[162] The chambers judge concluded that the appellants made a tactical decision in deciding not

to pursue cross-examination of Brown or further evidence on the point. It may be that the

appellants concluded further evidence on the point could only diminish the effect of the evidence

Page: 41

of TransAlta’s response to TVA testing at other sites by explaining it away. That this was a tactical

decision was an available conclusion on the record, as was the chambers judge’s conclusion that

the absence of the undisclosed records did not constitute manifest unfairness in the proceedings.

[163] It is not unusual in arbitration, or litigation, for the relevance of documents to become more

significant as the proceedings unfold. What might have appeared immaterial can become more

relevant as further information comes to light, and as the parties’ positions and the issues

crystallize. It is incumbent upon the parties themselves to reassert the position as the materiality

of records becomes more likely or more obvious. The disclosure orders made by the arbitration

panel allowed for this very possibility. Production, while not ordered, was not foreclosed. We

agree with the chambers judge that there was no manifest unfairness or inequality in this process,

nor were the appellants prevented from meeting the case of TransAlta.

How crucial were the undisclosed documents to the issues determined by the arbitration panel and

to the appellants’ ability to put forward their case?

[164] Determining whether the prerequisite to setting aside an arbitration award exists because a

party was “treated manifestly unfairly and unequally” or deprived of “an opportunity to present a

case” under section 45(1)(f) of the Act requires consideration of whether the alleged errors

challenge the fundamental validity of the process. Moreover, the unfairness must be “manifest”,

in the sense that it is obvious or apparent. Where the unfairness is said to arise from the exclusion

of evidence, that evidence must be crucial to the party’s case. If the alleged unfairness does not

impact the proceedings in a meaningful way, it cannot rise to the level necessary to meet the criteria

to set aside the arbitration award under section 45(1)(f).

[165] Put briefly, the appellants wished to make the case that it was unreasonable for TransAlta

to rely on the high readings from the TVA tests on the Unit because the TVA test is not widely

used in the industry as a primary diagnostic test, TransAlta and its consultants should have known

that the results were unreliable, and similarly high TVA readings at other sites did not result in

shutdowns, repairs, or claims of force majeure. They say that the undisclosed records were crucial

to that case, and they were unfairly precluded from obtaining them.

[166] As discussed above and as the chambers judge found, production of the records was never

foreclosed by the arbitration panel, and there was no manifest unfairness in the process. But in any

event, the argument regarding the TVA results was of less importance in the context of the

arbitration than the appellants now submit.

[167] First, as was noted in the discussion of the arbitration award, it is important to place the

arbitration panel’s conclusions regarding the TVA tests within the entire context of the award.

Page: 42

That TransAlta’s reliance on the TVA test was reasonable was one sub-issue in the review of

TransAlta’s decision to remove Bar T5.

[168] The arbitration panel concluded that there was no reason to exclude the TVA testing from

TransAlta’s decision making process at the investigation stage. The argument that the TVA test

was not in common use in the industry was found to confuse its appropriateness to determine

fitness for service with its use as part of an investigation to try to determine the cause of the trip

and to gather information on individual bars: award at paras 106-107. Mr. Milano, the appellants’

expert who was familiar with TVA testing, acknowledged that TVA tests could be informative and

could give information regarding the condition of insulation. If results are poor, it may be prudent

to postpone or eliminate other testing.

[169] With respect to the reliability of the TVA test results, the arbitration panel heard evidence

that the TVA meter was miscalibrated, resulting in increased test readings. However, the

arbitration panel concluded that this did not affect the reasonableness of TransAlta’s decision to

replace Bar T5 because, even if the readings were corrected for the miscalibration, Bar T5

remained an outlier and the test results showed a serious problem. The appellants’ argument also

ignores the findings of the arbitration panel that, at the time the decision was made to replace Bar

T5, the concerns raised from the TVA test were confirmed by other information, including a partial

discharge test, and that TransAlta had opinions from both Hitachi and Kinectrics that Bar T5

needed to be replaced or it was likely the relay would trip again.

Do comments made in the award elevate the significance of the undisclosed records (the “about

face” issue)?

[170] The appellants say that the “central unfairness” here emerged after the arbitration panel

issued its award. The essence of the appellants’ concern is that they were denied disclosure of

records under the control of TransAlta, following which the arbitration panel relied on the absence

of that evidence in deciding a key issue. The appellants describe this as an “about face”, arguing

that the arbitration panel’s procedural rulings denied production of the undisclosed records and

prevented the appellants from putting forward the evidence they wanted to call, but in their award

the arbitration panel held the absence of the undisclosed records to be material to their decision.

Were that so, it might indeed result in manifest unfairness or denial of an opportunity to present a

case: see, eg, Nortel Networks at paras 19-20; Ary Homes at paras 46-47; Arbutus Software v ACL

Services Ltd, 2012 BCSC 1834 at para 81.

[171] The appellants point to several passages of the award where they say the arbitration panel

relied on the absence of the undisclosed records to support their conclusions.

Page: 43

[172] At para 162, the arbitration panel referred to the general evidence given by witnesses

regarding the reasons why a particular unit might be dealt with in a particular way as a result of

high TVA test readings. At paras 163 and 164 of the award, the arbitration panel went on to say:

163 It can be argued that TransAlta, to be consistent, should have ignored the TVA

results here, just as they did elsewhere, but this would lead to the need to have a

detailed analysis of the reasons for the high readings for the various other units,

evidence we do not have. Rather we have some evidence from Mr Brown that in a

general way explains why the Keephills situation is different from the others.

164 In his cross-examination, Mr Brown was taken to the October email that lists

the other sites, but no questions were asked about the readings at the other sites, nor

was he asked about his explanations set out in his surreply witness statement.

Instead, Respondents say his evidence on this point offends the best evidence rule

and should be discarded. They say that evidence about the other sites should have

come from TransAlta. No documentation of reports regarding TVA results from

the other sites were produced at the hearing by TransAlta, nor did the Respondents

seek further production.

[173] As discussed above, these comments reflect the arbitration panel’s view of the disclosure

issue, which does not accord with the issue as framed by the appellants. It is apparent that the

arbitration panel was of the view that document production was never denied, and that it was open

to the appellants to cross-examine Brown on the significance of TVA results from other sites and

to obtain additional documentation on that point, if required. The chambers judge also found this

to be the case on the record and that the appellants’ decision not to cross-examine further was a

tactical one, a conclusion with which we agree.

[174] The comments at paras 169, 171, 172 and 176 of the award deal with the calibration

problem with the TVA probe and the knowledge of TransAlta and Kinectrics with respect to the

reliability of the TVA test results in March 2013, which has been discussed earlier in these reasons.

The arbitration panel said:

169 From the evidence we have about the other sites and the lack of any

correspondence or documentation at the time the Keephills 1 readings were taken

in March, the Panel must conclude that neither Kinectrics nor TransAlta had any

concern about the reliability of the test equipment used to test the Unit in March,

2013, nor any concern at the time that high readings at other sites should case the

Keephills Unit to be dealt with differently that it was.

Page: 44

…

171 There also remains the question as to whether the meter lacked linearity only

between ranges, or also exhibited non-linearity within a given range. The evidence

on this point is conflicting and not at all clear.

172 Mr Brown says he calibrated the meter before its use in March and it was

working correctly. Whether or not this is so, the Panel must conclude on the

evidence that the meter in question had some kind of linearity problem, but neither

Mr Brown nor anyone at Hitachi or TransAlta had any reason to be concerned in

March about the accuracy of the meter used in the TVA testing. The Panel has no

evidence to suggest Mr Brown or TransAlta ought to have known the readings were

faulty at that time.

…

176 Based on the facts available in March, the Panel cannot say TransAlta had

sufficient information to question the reliability of the test results or the opinions

of Hitachi and Kinectrics to replace Bar 5. The TVA tests were clearly appropriate

and could not be ignored. They militated against doing further hipot testing.

[175] A review of the record and the award in its entirety makes clear that the observations

regarding TransAlta’s level of knowledge about the reliability of the probe in March 2013 are a

red herring. As was noted by the arbitration panel, even adjusting for the miscalibration the

readings from the stator Bar T5 were an “outlier”. Even if the problem had been known, it would

not have affected the reasonableness of TransAlta’s decision to replace Bar T5, nor affected the

arbitration panel’s conclusions on the point.

[176] These comments refer to the absence of fleet wide records that might have suggested that

other TVA results from other sites were also high and raised questions regarding why TransAlta

did not respond by shutting down those facilities as it did the Unit. There are a few relevant points.

What evidence there was on the issue was that test results at other sites were not taken during an

emergency but during routine maintenance, and that the TVA test results on the Unit were uniquely

high. Moreover, the TVA tests were one piece of information used in the decision-making process

to replace Bar T5. To suggest these comments reflect unfairness overstates the significance of the

undisclosed records, both to the actual issue they might be used to determine and to the

deliberations of the arbitration panel.

[177] From the arbitration panel’s perspective, the fleetwide test results were not relevant to the

matters at issue during the pre-hearing production, although we agree with the chambers judge’s

Page: 45

conclusion that the arbitration panel never definitively foreclosed production should the records

eventually prove relevant and material. The arbitration panel’s interest in the records arose as a

result of the Brown surreply, as borne out by its comments regarding the failure to cross-examine

him. The chambers judge’s conclusion that this choice was a tactical one is well grounded in the

record. Moreover, that the records may have gained in relevance does not mean that their absence

was crucial to the ability of the appellants to put forward their case.

[178] For the reasons set out above, we find that the chambers judge did not commit a palpable

and overriding error in concluding that the failure to order production of these records earlier in

the process did not render the entire arbitral process manifestly unfair or deprive the appellants of

the opportunity to present their case or respond to that of TransAlta.

Conclusion

[179] The appeal is dismissed.

Appeal heard on January 27, 2022

Memorandum filed at Calgary, Alberta

This 9th day of June, 2022

Paperny J.A.

Rowbotham J.A.

Authorized to sign for:

Strekaf J.A.

Page: 46

Appearances:

D.W. McGrath, Q.C.

M.C. O’Brien

for the Appellant in 1901-0237AC

J.A. Bancroft, Q.C.

M. Gehlen

P.L. Roche

for the Appellant in 1901-0235AC

M.J. Donaldson

S.K. Hayes

for the Respondent