Pennsylvania Power Company (Penn Power) petitions for review of an order of the Pennsylvania Public Utility Commission (PUC), dated February 14, 1992, that determined that Penn Power is liable for a refund of the replacement power costs it incurred and passed on to its consumers through an automatic adjustment clause as a result of a 148-day shutdown of the Beaver Valley I Nuclear Generating Station (Beaver Valley I) in 1979 pursuant to a March 13, 1979, order of the United States Nuclear Regulatory Commission (NRC). We affirm.
Factual Background
Beaver Valley I is an 810-megawatt electric generating plant powered by nuclear fission, which went into service in 1976. Beaver Valley I is operated by Duquesne Light Company (DLC), a 47.5% owner. Penn Power is a 17.5% owner. The remainder is owned by Ohio Edison. Beaver Valley I was designed in 1968 and subsequently constructed by the Stone and Webster Engineering Corporation (Stone & Webster). Prior to 1972, Stone & Webster relied on a computer code known as PIPESTRESS, which applied an algebraic summation (algebraic formula) to determine load capacity of the safety-related piping at Beaver Valley I and four other plants. The safety piping must not only withstand the loads imposed by normal operation but also be able to withstand the stress of an earthquake. In 1972 Stone & Webster changed its computer code for seismic analysis to NUPIPE, and applied a square-root sum of the squares formula (square-root formula) to calculate the stress factors of the safety-related pipes. The difference between the older algebraic formula and the newer square-root or non-algebraic formula in determining stress *481calculations for Beaver Valley I’s safety piping lies at the heart of this controversy.
The first critical time period commenced on February 17, 1978, and continued until October 2, 1978. On February 17, 1978, Stone & Webster sent a letter to Westinghouse Corporation (Westinghouse) questioning whether the weight of certain valves (Velan valves) supplied by Westinghouse for use in Beaver Valley I equalled their design weight, because a difference in weight could affect the safety piping. On May 30, 1978, more than three months later, Westinghouse responded with the correct weight of the fifteen valves. The weight of fourteen of the valves differed from the design weight. Although this presented a safety-related issue regarding whether Beaver Valley I’s seismic stress analysis was reliable, DLC failed to pursue a solution. On October 2, 1978, Stone & Webster informed DLC of an error in the hand-calculated stress analysis for certain safety injection lines as a result of computing the corrected actual weight of fourteen of the Velan valves.
On October 26,1978, Stone & Webster informed DLC of an overstressed safety injection line. DLC then notified the NRC by telephone of the problem. The next day, October 27, 1978, DLC again reported this problem in a Licensee Event Report (LER). Based on this history DLC was aware in October of 1978 that an initial review of stress calculations for safety injection systems revealed excessive stress in the safety-related piping and that the as-built condition of the plant differed from the design drawings as to 14 safety injection system check valves (the as-built valves weighed twice as much as their design weight).
The second critical time period occurred between October 2, 1978, and March 13, 1979, the date of the Beaver Valley I shutdown. On November 9, 1978, DLC submitted an Interim LER to the NRC indicating that the October 27, 1978, LER was in error and that the hand-calculated stresses were subsequently computer calculated and found to be acceptable. Meanwhile, the NRC conducted an on-site inspection from October 31, 1978, through November 3, 1978. In the NRC’s *482Inspection Report 78-30, issued December 4, 1978, the NRC asked DLC if assurance could be given that the hand-calculated error applied only to the points in question. Also, Donald Beckman (Beckman), the NRC’s resident inspector, conducted another on-site inspection on December 18-20, 1978, and received a December 6, 1978, LER from DLC (containing a Stone & Webster report), which indicated a difference in stress results calculated by the two formulas. After the inspection, Beckman pursued the matter with the NRC but the information provided by Stone & Webster was insufficient to answer questions regarding the difference in stress results produced by the two different computer codes. Despite a January 26,1978, letter from DLC to Stone & Webster urging them to respond to the NRC inquiries, adequate responses to these questions were not provided until March 8, 1979.
Furthermore, DLC conducted its own investigation into the stress calculation problem. DLC’s engineer, J.J. Lynch, met with Stone & Webster at its Boston offices on December 18-19, 1978, to conduct an audit of Beaver Valley I’s design records. In his report issued January 17, 1979 (Lynch Report), Lynch identified concerns regarding the adequacy of the seismic piping design at Beaver Valley I based upon the difference in calculations. On January 22, 1979, DLC sent the Lynch Report to Stone & Webster and requested a response by March 1, 1979. DLC did not mail a copy of the Lynch Report to the NRC until April or May of 1979, after the shutdown.
The NRC remained concerned about the difference in calculated stress results. On March 1, 1979, the NRC requested details of the calculations, which Stone & Webster initially declined to provide in accordance with company policy. A meeting was held on March 8,1979, at which Stone & Webster made available the computer codes and calculations. The NRC staff examined the calculations and found significant discrepancies in the results calculated under the older algebraic formula compared to the results under the newer square-root formula. The NRC became alarmed and requested that additional analyses be performed over the weekend of March *4839-12, 1979, using the more conservative square-root formula. The results of the weekend analyses showed load stresses three times the total allowable. On March 13, 1979, Beaver Valley I and four other plants built by Stone & Webster (Surry I & II, Maine Yankee and Fitzpatrick) were shut down.
The problem with the Stone & Webster plants resulted from a lack of drawings or blueprints reflecting the plant’s actual “as-built” condition as of March 13, 1979. This information was needed to conduct a reanalysis of seismic stresses using a non-algebraic formula. Because Stone & Webster had not maintained and updated its blueprints during construction, they were not able to supply the NRC with accurate as-built drawings until April 20, 1979. The NRC was not satisfied with Stone & Webster’s reanalysis until August, 1979. The shutdown order was lifted on August 8,1979. Beaver Valley I returned to full service on August 17, 1979. However, DLC and Penn Power had to purchase replacement power during the shutdown and the costs for the replacement power were passed on to consumers through the utilities’ net energy clause charges.
Procedural Background
On July 12, 1979, the PUC instituted an investigation into the rate-making consequences of the Beaver Valley I shutdown. Originally, the PUC proceeding included only DLC, as the 47.5% owner and the operator of Beaver Valley I. On August 3, 1979, the PUC joined Penn Power as a party to the investigation because it owned 17.5% of Beaver Valley I.1
On February 21, 1981, after nine days of hearings, at which Penn Power was present as a party, the PUC issued an order finding that DLC did not meet its burden of proof by demonstrating that the shutdown and resulting costs were prudently incurred. DLC appealed that decision to this Court. In Duquesne Light Company v. Pennsylvania, Public Utility Commission, 96 Pa. Commonwealth Ct. 168, 507 A.2d 433 *484(1986), we vacated the PUC’s decision on the ground that DLC’s right to due process was violated, because it was not afforded adequate notice of the PUC’s contention that DLC acted imprudently prior to the shutdown and was thus denied a reasonable opportunity at the PUC’s evidentiary hearing to address the issue of the reasonableness of its conduct preceding the shutdown. Id. at 176-77, 507 A.2d at 437.
This Court remanded the matter to the PUC for a determination whether DLC' acted prudently in response to the NRC’s inquiries into the adequacy of the seismic design of the safety piping at Beaver Valley I by taking all reasonable steps to avoid a shutdown. Id. at 178-79, 507 A.2d at 438. On November 6, 1981, the PUC ordered further hearings to identify the costs resulting to DLC ratepayers resulting from the shutdown. On November 10, 1981, the investigation was reopened with respect to Penn Power. However, in lieu of proceeding to hearing, DLC elected to resolve the case by entering into a settlement, with the approval of the PUC, on June 30, 1986. Pursuant to the terms of the settlement DLC agreed to an aggregate refund of $15,300,000.00 including principal and interest.
By letter dated January 27, 1987, the PUC assigned the Penn Power investigation to an Administrative Law Judge (ALJ). Three days of evidentiary hearings were held at which Penn Power, the Office of Consumer Advocate (OCA) and the PUC’s Office of Trial Staff (OTS) presented witnesses. The record was closed on November 10, 1987. On May 29, 1990, the ALJ issued a recommended decision concluding that DLC and Penn Power did not act imprudently with respect to the Beaver Valley I shutdown.
Penn Power, OCA, and OTS filed exceptions to the ALJ’s decision. OCA and OTS disagreed with the ALJ’s determination that DLC and Penn Power acted reasonably during the pre-shutdown period. Penn Power requested a further finding that DLC’s conduct was not the real or proximate cause of the shutdown. On February 14, 1992, the PUC reached a conclusion contrary to the ALJ’s, and found that Penn Power was liable for the replacement power costs incurred as a result *485of the shutdown of Beaver Valley I. Penn Power subsequently filed a $3.9 million refund plan with the PUC. On March 13, 1992, Penn Power petitioned this Court for review of the PUC’s February 14, 1992, decision.
Statement of Issues
On appeal, Penn Power contends: (1) that the PUC erred in imposing liability on Penn Power for DLC’s allegedly imprudent conduct despite its own finding that DLC’s conduct was not the real or proximate cause of the shutdown; (2) that the PUC erred in concluding that DLC acted imprudently where all the contemporaneous evidence confirms that the NRC considered DLC’s responses to its inquiries to be reasonable; (3) that the PUC erred in disregarding uncontested evidence and by attaching liability for the entire outage period; (4) that the PUC erred in directing refunds while refusing to consider evidence that Penn Power failed to earn its authorized rate of return during the period in question; and (5) that the PUC erred by failing to make necessary findings before ordering a refund as required by 66 Pa.C.S. § 1312.
Issues I & II
Our scope of review is limited to determining whether constitutional rights were violated, whether an error of law was committed or whether the PUC’s findings of fact, determinations or order are supported by substantial evidence. Equitable Gas Company v. Pennsylvania Public Utility Commission, 106 Pa. Commonwealth Ct. 240, 526 A.2d 823 (1987). We will not substitute our evaluation of the evidence for that of the PUC. Id. at 254, 526 A.2d at 829. We also note that the burden of proof in a rate proceeding is squarely upon the utility to establish that the rate is just and reasonable under 66 Pa.C.S. § 1301. Lower Frederick Township Water Company v. Pennsylvania Public Utility Commission, 48 Pa. Commonwealth Ct. 222, 226-27, 409 A.2d 505, 507 (1980).
Penn Power contends that the PUC’s determinations are not supported by substantial evidence. Specifically, Penn Power contends that it is not liable for refunds unless it is *486found that DLC, the operator of Beaver Valley I, acted imprudently and that DLC’s imprudent conduct was the real and proximate cause of the shutdown. Penn Power also contends that the PUC erred in finding that DLC acted imprudently where all the contemporaneous evidence confirmed that the NRC considered DLC’s responses to its inquiries to be reasonable.
In Pennsylvania Public Utility Commission v. Philadelphia Electric Company, 522 Pa. 338, 561 A.2d 1224 (1989) our Supreme Court held that replacement power costs incurred by an electric company resulting from its failure to exercise the high degree of care required at nuclear power plants were due to mismanagement and were not to be passed on to consumers through a rate increase. In Philadelphia Electric the Supreme Court reversed this Court, in part, with regard to the standard of care imposed upon a utility in determining whether management abused its discretion in operating a utility:
In vacating the order of the Commission with respect to the first outage at Salem 1, Commonwealth Court held that the Commission imposed an improper standard of care upon the Appellee. The NRC requires that plant operators exercise a high degree of care in operating and maintaining their facilities. Commonwealth Court agreed with the Appellee that the PUC adopted this heightened degree of care in this [rate] proceeding. Since the primary purpose of a rate case is to evaluate the effect of a rate increase on the customer and investor so as to avoid unreasonable utility costs, Commonwealth Court concluded that the heightened standard of care to ensure safety is inappropriate.
While Commonwealth Court states the various principles of law, we cannot agree with its application of those principles. There was no question that the Commission was concerned with numerous NRC violations. It is also true that the Commission acknowledged the high degree of care required by the NRC. However, in addressing these standards, the Commission did not alter the burden of proof required, but concluded that the Appellee’s failure to adhere *487to the NRC’s requirements resulted in outages and delays in restarting the operations. According to the Commission’s analysis, if the Appellee had operated its plants in accordance with NRC regulations, the mechanical failures would have either been detected or prevented. Because of the Appellee’s lack of a sufficient maintenance policy, the failure of the protection system occurring on February 22, 1983 was not even detected until the second failure. Thus, the Commission was not merely ensuring safety but rather was evaluating whether the replacement costs were prudent given the Appellee’s lack of maintenance and operating practices.
The use of the term ‘high degree of care’ may have been inappropriate but the Commission’s application of its enunciated degree of care is consistent with our prior holdings. If the Appellee acted imprudently in failing to have appropriate maintenance and operating practices and as the result of that failure the power plants were unable to supply energy to its customers, management has acted arbitrarily and abused its discretion. Such additional costs would not therefore be just and reasonable.
522 Pa. at 344-46, 561 A.2d at 1227 (citations and footnote omitted) (emphasis added). In accordance with Philadelphia Electric, Penn Power has the burden of proof to establish in the present case that DLC prudently and appropriately responded to the NRC’s inquiries and requests for information.
Also, in Philadelphia Electric, the Supreme Court held that where a utility chooses a contractor and negotiates the contract, the utility must bear the initial losses incurred as a result of its contractor’s negligence. 522 Pa. at 346, 561 A.2d at 1228. The Court noted that if it were to hold otherwise the utility would have no incentive to pursue the tortfeasor, having already received full compensation through increased rates. Id. Consequently, DLC, and hence, Penn Power, are responsible for the acts or omissions of Stone & Webster, the contractor who built the Beaver Valley plant.
*488Our examination of the PUC’s February 14, 1992, decision indicates that the PUC made several critical findings regarding DLC’s actions prior to the shutdown. On page 11 of its decision, the PUC summarized the testimony of Dale G. Bridenbaugh (Bridenbaugh), a nuclear power expert called by the OCA. Bridenbaugh stated that between October 2, 1978, and March 8, 1979, there were clear indications that the problems with Stone & Webster’s piping analyses required an organized, thorough, and urgent resolution, but that DLC wanted to operate Beaver Valley I until the next scheduled refueling outage. Reproduced Record (R.R.) at 626a.
Bridenbaugh testified that the NRC performed inspections and questioned DLC about its stress analyses as early as November 1978, and that DLC’s own engineering report identified problems in the analyses and the techniques employed. R.R. at 627a. Bridenbaugh further stated that the NRC’s Regulatory Guide 1.92 (issued in 1974 and revised in 1976) defines more conservative analytical techniques (than those used by Stone & Webster) as being methods acceptable to the NRC for combining responses and loads in seismic analyses. R.R. at 627a. In Bridenbaugh’s opinion, DLC was aware of substantial differences in the results calculated from various Stone & Webster codes before December 1978 and aware that the NRC questioned the codes employed on Beaver Valley I at that time. R.R. at 627a. In light of Bridenbaugh’s testimony the PUC concluded that DLC failed to take timely and reasonable action in order to insure the safety of the plant. PUC’s Decision, February 14, 1992, at 11.
The PUC also concluded that DLC did not adequately respond to critical questions posed by the NRC in its Inspection Report 78-30 concerning calculation errors in Stone & Webster’s analyses and whether steps were being taken to resolve those errors, until March 8, 1979. PUC’s Decision at 14. This finding is supported by the testimony of Russell Beckman (Beckman), the NRC resident inspector at the time of the shutdown, who stated that DLC did not provide any information regarding Stone & Webster’s calculation errors prior to March 8, 1979. R.R. at 766a.
*489Also, the PUC found that DLC failed to aggressively pursue a solution or to respond timely to the concerns raised in its own engineering report and that such failure was imprudent and unreasonable. The Lynch report, issued January 17, 1979, indicated that Stone & Webster’s algebraic formula was less conservative and calculated less stress than the square-root formula. R.R. at 511a. However, neither DLC nor Stone & Webster informed the NRC of the Lynch Report prior to the shutdown.
We further note that the record indicates David Lee Wigginton (Wigginton), the NRC’s project manager, testified that twenty-nine nuclear plants were designed using the algebraic formula. R.R. at 347a. However, only Beaver Valley I and the four other plants built by Stone & Webster were shut down. R.R. at 347a. Wigginton stated that there were significant differences between the as-built condition of the Beaver Valley plant as it existed in the field and the design drawings in existence (which prevented non-algebraic reanalysis as of March 13, 1979), and that such differences required a shutdown. R.R. at 351-52a. Wigginton further stated that (after sufficient reanalysis information had been provided) no subsequent modifications were required due to the as-built condition of Beaver Valley I’s piping. R.R. at 354a. However, Wigginton testified that it was not until August, 1979 that sufficient reanalysis information was provided to the NRC’s satisfaction. R.R. at 346a.
Also, William T. Russell (Russell), the NRC’s Chief of Technical Support, testified that if Stone & Webster had provided to him all of the reanalysis information contained in their April 20, 1979, report on Beaver Valley I, the shutdown would probably have been avoided. R.R. at 772a. Russell noted that three other nuclear plants (Indian Point No. 3 and Brunswick Nos. 1 and 2) which required similar continued non-algebraic reanalysis were allowed to continue to operate based on the analysis presented. R.R. at 772a.
In view of the foregoing record evidence, the PUC concluded that DLC’s failure to resolve the concerns regarding the safety of Beaver Valley I in a timely manner was imprudent. *490In view of Philadelphia Electric, we agree. Stone & Webster’s failure to provide the NRC with sufficient non-algebraic reanalysis information, including accurate as-built drawings or blueprints prior to March 13, 1979, proximately caused the shutdown. Concededly, DLC did not create the problem. Stone & Webster committed the errors in calculation and failed to maintain accurate as-built drawings of the plant. However, a utility that selects its contractor bears the risk of contractor performance failures. In fact, neither DLC nor Penn Power disputes responsibility for the acts of Stone & Webster.
However, Penn Power contends that the PUC concluded that DLC’s conduct was not the real or proximate cause of the shutdown. Penn Power cites to page 27 of the PUC’s decision where the PUC states:
Our review of the record indicates that it is devoid of a showing that had Duquesne undertaken any action, the shut-down would have been averted. In fact, it is clear from the record that the decision of the NRC to shut down Beaver Valley I, did not consider any action or inaction on the part of Duquesne.
Actually, Penn Power failed to persuade the PUC that DLC’s conduct, prudent or imprudent, was not the real and proximate cause of the shutdown. The cited paragraph pertains only to events after March 8, 1979, and does not pertain to DLC’s conduct in the preceding five months. As noted herein, the PUC made several findings involving DLC’s conduct prior to March 8, 1979. The PUC merely noted that by March 8,1979, the NRC had no confidence in the design of the plant and that because of the lack of as-built information or updated blueprints needed to conduct a non-algebraic reanalysis, there was nothing DLC could do to avoid the shutdown. More importantly, Penn Power, which had the burden of establishing that DLC’s actions prior to the shutdown were prudent, did not produce any evidence to indicate that its conduct and/or responses to NRC inquiries were sufficient to prevent the shutdown. In fact, our review of the record indicates that DLC did not object to the shutdown. Accord*491ingly, we hold that the record supports the PUC’s conclusion that DLC’s actions prior to March 8, 1979, were imprudent and were also the proximate cause of the plant shutdown.
Penn Power also asserts that the PUC erred in rejecting the testimony of Penn Power’s witnesses and accepting the testimony of witnesses produced by OCA because Penn Power’s witnesses had the advantage of being present at the time of the shutdown. It is beyond cavil that the PUC has the authority to determine the weight and credibility of the evidence presented, and our scope of review does not permit us to engage in reweighing evidence. Peoples Natural Gas v. Pennsylvania Public Utility Commission, 122 Pa. Commonwealth Ct. 445, 458, 552 A.2d 1135, 1142 (1989). Our review of the record indicates that the OCA witnesses relied upon by the PUC were well-qualified nuclear experts who carefully reviewed the facts to determine what was known or should have been known by DLC prior to the shutdown. The OCA witnesses then evaluated DLC’s conduct. We believe that this type of analysis is not impermissible hindsight. Although the ALJ chose to believe Penn Power’s witnesses and found Penn Power’s conduct to be prudent, an ALJ’s decision may always be overruled based upon contrary findings by the PUC if the PUC’s findings are based on substantial evidence. Pennsylvania Retailer’s Association v. Pennsylvania Public Utility Commission, 64 Pa. Commonwealth Ct. 491, 503, 440 A.2d 1267, 1272 (1982).
Likewise, the PUC did not err in concluding that the ALJ improperly disregarded the testimony of four NRC witnesses called by OTS, given in the first phase (Phase I) of the proceedings against DLC in 1980. Penn Power contends that the PUC never explained what testimony it did regard. To the contrary, the PUC specifically referred to the testimony of Beckman, the resident inspector at the time of the shutdown. PUC’s decision at 14. The NRC’s witnesses were critical in establishing the sequence of events leading up to the shutdown. Penn Power was a party to those proceedings and had the opportunity to cross-examine the NRC witnesses. Ac*492cordingly, we conclude that the PUC sufficiently identified the testimony from the Phase I hearing that the ALJ disregarded.
Issues III, TV & V
Penn Power also contends that the PUC erred in three additional areas by holding DLC, and hence Penn Power, liable for the entire 148-day outage: first, that it did not consider the intervening accident at Three Mile Island (TMI); second, that it failed to consider uncontested evidence regarding Penn Power’s earnings deficiency during the pertinent period; and third, that it did not make necessary findings regarding an earnings deficiency as required by 66 Pa.C.S. § 1312.
Penn Power contends that it should not be held liable for the entire outage because the TMI accident on March 28, 1979, had several immediate and major effects on NRC regulation, which served to lengthen the time Beaver Valley I remained out of service. Penn Power also contends that the TMI accident diverted substantial NRC resources away from Beaver Valley I. However, assuming arguendo that the TMI accident interfered with the resolution of the Beaver Valley I problem, Penn Power’s ratepayers cannot be held responsible. In Philadelphia Electric, our Supreme Court affirmed a PUC decision, Re: Salem Nuclear Generating Station, 60 Pa. PUC 249, 70 PUR 4th 568 (1985), holding the utility liable for the entire length of the Salem I shutdown even though the utility’s actions were not found to be imprudent regarding a shellfish infestation that extended the outage. The Court noted that the energy replacement costs of the infestation were not chargeable to the ratepayers. Philadelphia Electric, 522 Pa. at 346 n. 1, 561 A.2d at 1227-28 n. 1. In the present case, the NRC remained unsatisfied that there was no danger to the public health and safety prior to August 8,1979. Penn Power did not establish any facts to the contrary. Accordingly, we hold that Penn Power’s ratepayers cannot be charged for any part of the Beaver Valley I shutdown.
*493Penn Power next contends that the PUC erroneously failed to consider uncontested evidence that Penn Power experienced an earnings deficiency during the period in question. In the same vein, Penn Power asserts that the PUC erred in directing a refund without making findings as required by 66 Pa.C.S. § 1312(a) as to whether the replacement costs were (1) unjust and unreasonable or (2) in excess of the rates contained in the utility’s applicable tariff filing.
In the present case Penn Power did recover the replacement costs for the Beaver Valley I shutdown under its net energy (automatic adjustment) clause without review by the PUC as to whether such charges were excessive due to managerial imprudence. As a result, the PUC could have directed a refund under either 66 Pa.C.S. § 1312 (general refund section) or 66 Pa.C.S. § 1307 (automatic adjustment clause) if the replacement costs were subsequently determined to be unreasonable. Under either section, imprudent costs are not a “just and reasonable” element of rates as required by 66 Pa.C.S. § 1301. In Equitable Gas Co. v. Pennsylvania Public Utility Commission, 113 Pa. Commonwealth Ct. 68, 536 A.2d 846 (1988), we held that the PUC may deny recovery of costs found to be the result of managerial imprudence so long as the PUC’s findings are supported by substantial evidence. Id. at 74, 536 A.2d at 849. Having found the replacement power costs to be imprudently incurred, the PUC did not err in directing a refund, regardless of any earnings deficiency experienced by Penn Power.
The February 14, 1992, decision of the PUC is affirmed.
ORDER
AND NOW, this 6th day of May, 1993, the order of the Pennsylvania Public Utility Commission in the above-captioned case is affirmed.
. As co-owner of Beaver Valley I, Penn Power is liable for DLC’s conduct in operating the plant. See Re: Salem Nuclear Generating Plant, 60 Pa. PUC 249, 70 PUR 4th 568 (1985).