Subpage for this topic on my Wikipedia User page
Excerpts related to the burn
Rogovin, Volume 2, Part 3
As Miller was leaving for Harrisburg, he was not aware of the containment pressure spike, nor were Logan or Kunder aware of it, but others were aware of it and thought everyone else knew about it. (page 828, Volume II, part 3)
Herbein did not learn of the containment pressure spike that occurred at 1:50 p.m. until it became general knowledge on Friday morning. (page 834, volume II, Part 3)
A hydrogen burn occurred in the reactor building at 1:50 p.m., shortly before Miller’s departure. This burn resulted in a 28-psi pressure spike and starting of the containment spray pumps. Miller recalled hearing a thud at this time, but not being aware of the pressure spike or of the containment spray pumps running.
Chwastyk recalls diagnosing the containment pressure spike as an explosion associated with venting the pressurizer, and of so informing Miller prior to his departure. Ross, while not thinking of the pressure spike as an explosion, did realize that the pressure spike had occurred and the containment spray pumps had started. He discussed the noise with Miller and believed that Miller also knew that the pressure spike had occurred and the containment spray pumps had started.
Miller, however, recalled only asking about the noise and being told that it was probably caused by a ventilation damper. He did not recall being aware of the pressure spike or the containment spray pump operation, much less of the diagnosis of an explosion. Nor did Herbein recall Miller informing him of these occurrences. These events would all have taken place shortly before Miller left for Harrisburg. If the confusion and rush of leaving was a reason why some of this information did not reach Miller, as it could have been, then leaving was certainly a drawback. (page 837, volume II, Part 3)
One can speculate that, had Miller stayed, he might have recognized the containment pressure spike, and either understood its significance or informed others who might have comprehended its significance. If the significance of this event had been generally conceived on Wednesday, one can further speculate that in the next few days the overall situation would have been safer because (1) actions to handle the hydrogen bubble would have been taken sooner, shortening the time that this presented a potential threat to continued core cooling and (2) contingency planning would have been implemented faster, reducing the risk involved. (page 838, volume II, Part 3)
GPUSC personnel were not aware of the leaking PORV and the throttling of high-pressure injection flow, which had caused the core damage. Nor were they aware of the containment pressure spike (hydrogen burn) that had occurred. The personnel were also unaware of the offsite radiation releases and probably not aware of specific radiation readings in the plant that qualitatively could indicate that the containment dome monitor might be correct. (page 840, volume II, Part 3)
Sometime Thursday night or Friday morning, control room personnel became aware of the containment pressure spike. The significance of this spike became commonly known on Friday morning. (page 842, volume II, Part 3)
On Friday, March 30, several factors combined to make the company management and staff aware of the nature and seriousness of the remaining problems in the plant.
The containment pressure spike on March 28 and its diagnosis as a hydrogen burn became generally known. (page 945, volume II, Part 3)
”’The Pressure Spike”’ (page 902, volume II, Part 3)
At about 1:50 p.m. on Wednesday, March 28, 1979, a control room instrument that displays reactor containment building pressure indicated a sudden, short-lived but dramatic pressure increase in the reactor building. As an operator and at least two supervisors standing at the consoles watched, the pen-recorder on the instrument-which traces the level of building pressure on a slowly moving drum of graph paper with a scale from 0 to 80 pounds per square inch gauge (psig)-jumped almost straight up to about 28 psig then slowly fell back over the next 15 seconds to about 4 psig. At the same time, a sodium hydroxide “sprinkler” system in the reactor building that is triggered automatically (when containment pressure reaches a nominal pressure of 30 psig) came on. Several individuals in the control room heard a sound they later described as “thud.”
At the time, little attention apparently was paid to this “pressure spike.” Because the volume of the containment building is so large, many of those in the control room evidently could imagine no credible phenomenon that would have produced such a large pressure increase in so short a time. Some wrote off the instrument reading in their minds as having probably been caused by a stray electrical pulse or “transient” in the electrical wiring, rather than by an actual pressure surge in the reactor building.
When company employees examined the graph of the pressure spike late Thursday night in the light of evidence that had been collected earlier that day (including a reactor coolant sample that showed massive amounts of radioactive material in the coolant water), they realized that an explosion of flammable hydrogen gas must have taken place in the reactor building Wednesday afternoon.
Key GPU officials and NRC personnel learned of the spike and its significance early Friday morning. The large amounts of hydrogen necessary to support such an explosion or “burn” could only have come from chemical decomposition at very high temperatures of a substantial amount of the reactor “cladding”-the protective outer sheathing of the fuel elements that surrounds and encloses the uranium fuel pellets. The cladding is made of a metal called Zircaloy-4, an alloy containing about 96% zirconium, which reacts rapidly with steam at temperatures above 2200°F to produce hydrogen gas. The gas, in turn, escaped from the reactor coolant system into the reactor building through the stuck-open PORV when the block valve behind the PORV was periodically opened to relieve overpressure in the reactor coolant system Wednesday morning, and again after 11:30 a.m. when the block valve was opened for long periods of time to depressurize the system.
Had the pressure spike been recognized as a hydrogen explosion early Wednesday afternoon, it would clearly have demonstrated that the reactor core was uncovered or had been uncovered for a long period of time. Moreover, the fact that temperatures high enough to produce zirconium-water reaction had been sustained in the reactor core for sufficiently long to create the amount of hydrogen necessary to cause such an explosion would have signalled that the core had been very close to a possible meltdown or had indeed experienced significant melting.
The fact that neither the existence of the spike nor its significance came to the attention of responsible NRC officials until Friday morning has raised the question whether Met Ed personnel intentionally withheld important information from the NRC that would have demonstrated that the accident was much more serious than was generally thought on Wednesday and Thursday. The evidence relative to this matter, including testimony which in some respects is inconsistent, follows.
When the pressure spike occurred, Station Manager Miller was in charge of the control room. At that time, Miller was preparing to leave Three Mile Island to brief Lt. Gov. Scranton at the Capitol i n Harrisburg, approximately 25 minutes away by car. Miller would be accompanied by George Kunder, the Superintendent of Technical Support for Unit 2, who was also in or near the control room at 1:50 p.m. Miller’s “deputy” in the control room (whom he left in charge of the plant when he departed for Harrisburg a few minutes later) was Joe Logan, the Superintendent for Unit 2. Also in the control room area were at least three shift supervisors: Bill Zewe, who had been on shift at 4:00 a.m. when the accident began and remained throughout the day; Brian Mehler, who had arrived at about 6:00 a.m. in anticipation of taking over the normal 7:00 a.m. shift; and Joe Chwastyk, who had assumed responsibility for transmitting operating instructions to the control room operators at about 11:00 a.m. that morning, and was therefore directly “in charge” of the control room consoles. Also in the control room area were Mike Ross, the Supervisor of Operations for Unit 1, who was in charge of operations at the control room (Joe Logan was relatively the least experienced in the operating characteristics of the Unit, among the “management group” present) and Lee Rogers, the B&W site representative. Logan, Ross, and Rogers had been present since early that morning.
Zewe has testified that he was standing directly behind the instrument panel, that he saw the spike “first,” and that “everybody there” saw the spike. He recalls discussing with others, among them Ross and Chwastyk, what could have caused the spike and whether the instrument had registered an actual pressure excursion in the building or had malfunctioned. He recalls specifically some discussion of whether a stray electrical impulse in the circuitry might have caused the instrument indication. “Nobody,” he recalled, “had good answers,” and it was eventually concluded that an electrical transient in the circuit rather than actual pressure increase must have caused the pen-recorder’s movement. Zewe also recalls observing that the sodium hydroxide spray activated and shortly thereafter was ordered turned off when it became clear that the pressure increase, if any, had subsided. Significantly, Zewe does not recall discussing the pressure spike with, or even mentioning its existence to Gary Miller.
Zewe’s testimony is corroborated by Mike Ross, who testified that although he did not personally observe the spike as it occurred, his attention was called to it by a report that the building spray pumps had gone on. Ross recalls discussing the spike with Zewe and concluding that it had been caused by an electrical spike in the circuitry, not by an actual pressure increase in the building. Ross recalled the following:
There were many people there, and it was common knowledge that it happened, but I don’t think anybody ever sat down and analyzed it at the time … I think we reached a hurried conclusion saying that we thought the spike was caused by either a malfunction of some kind, and we just went on taking care of business …
John Flint, a B&W engineer who had also been in the control room since early morning (and who, incidentally, testified in a deposition before the President’s Commission that he personally had concluded in midmorning that the core had been uncovered), 68 said in an interview with a Met Ed investigator on April 20, 1979, that he was aware of the high pressure indication and the spray pump actuation, “and about the same time there was a double thud.” Flint also stated the following:
I personally did not think it was from the reactor containment building. I thought that it was the ventilation dampers cycling. It was very close to that sound, and since we had been in and out of
respirators due to the levels in the control building, I just thought somebody had cycled the ventilation dampers again and related it to that.
Joe Chwastyk, who was in the control room along with Zewe and observed the pressure spike as it occurred, also recalls that there was a “lot of conversation” about the spike and that “it immediately came to mind that we had some kind of instrument problem.” However, unlike the others, Chwastyk soon concluded that the spike could not have been caused by a stray electrical pulse because the spray pumps had come on and “there are two different pressure instruments used, one for the recorder and one for starting the pumps.” In other words, the fact that the reactor building spray pumps had been activated confirmed that an independent instrument had also detected a pressure surge in the reactor building.
Brian Mehler remembers coming out of the shift supervisor’s office when he saw activity in the control room indicating that “an ES” (an activation of safety systems) had occurred, and observing the pressure spike trace. He too first “thought it could have been an electrical thing. But then looking at the spray pumps, I realized it couldn’t have been.”
In fact, as Mehler recalls explaining to an NRC inspector who was present, the reason the actuation of the spray pumps tended to rule out an electrical transient as the cause of the pressure spike was essentially the same reason suggested by Chwastyk in his deposition. Mehler recognized that it was the “coincident logic” arrangement of the circuitry for the spray pump system that made such a possibility highly unlikely. There are three sensors in the reactor building that record building pressure. To reduce the possibility that the spray pumps will come on simply because one sensor malfunctions (thereby necessitating a time-consuming cleanup of the building), the logic circuitry for the spray pump system requires that at least two of the three sensors simultaneously record a pressure increase to 30 psi or above. Because the sensors are connected to the logic circuit by independent wiring, an electrical transient in the wiring leading to only one of the sensors (which might cause the control room instrument to record a spurious spike) would not activate the spray pumps; only a real increase in building pressure could conceivably have such an effect.
Strangely enough, others in the control room area have testified that while they too heard the “thud” described by Flint, they did not even become aware of the existence of a spike or the activation of the spray pumps on Wednesday. Gary Miller, the Station Manager, has consistently given statements and testimony that he heard a “loud deep noise” while standing in the control room shortly before he left for the Capitol, but that he does not believe he knew on that day either that there had been a spike in the building pressure instrumentation or that the spray pumps had activated. Joe Logan gave similar testimony.
Lee Rogers has also testified that he heard a noise and heard conversation or was told that the noise was probably caused by ventilation dampers in the control room ventilation system slamming shut. (Miller has also testified that he recalls hearing someone comment that the noise might be the ventilation system.) Rogers also denies knowing either of the pressure spike or the spray pumps coming on.
George Kunder, who accompanied Miller to Harrisburg, does not recall hearing a noise or knowing on Wednesday of the pressure spike or the activation of the spray pumps.
Immediately after the pressure spike, Joe Chwastyk recalls “a lot of conversation with just about everybody in the control room” about whether “anybody knew what was happening, because I didn’t at the time.” Nor did Mehler immediately recognize what happened. But Mehler recalls that shortly thereafter, perhaps 10 minutes to an hour later, he and Chwastyk talked the matter over and realized that there might have been an explosion.
Chwastyk does not specifically recall the conversation with Mehler, but acknowledges it is quite possible that such a conversation occurred. According to Chwastyk’s recollection, some time after the event he began to ponder its significance and “it just flashed through [his] mind” that one of the control room foremen, Fred Scheimann, had manipulated a valve at precisely the same time that the pressure spike occurred. What Chwastyk realized was that if the valve position had been changed by the operation of a DC (direct current) motor in the containment building, the motor might have created a spark that could have detonated any flammable gas present. “And I think it was after someone related to me also the noise they had heard that I assumed then it was some sort of hydrogen explosion.”
Chwastyk himself “assumed it was hydrogen” that must have caused any explosion in the reactor building, admitting in testimony before us that he knew of nothing else that might have exploded. Mehler has testified that he does not recall the possibility of hydrogen having come up in his conversation with Chwastyk that afternoon. He remembers thinking only that it might have been a “chemical reaction.” As Mehler recalls, he does not believe he considered hydrogen to have been the cause because he “didn’t think hydrogen could form that quick in the building to that concentration in that period of time.” Mehler testified that the possibility of hydrogen never entered his mind until he read it in the news media.
There is no question that after thinking the matter over, both men became quite concerned. Mehler recalls being “really a little scared.” As Chwastyk puts it, “It scared the hell out of me …” Therefore, at least two supervisors in the control room realized sometime shortly after the event that there had been an actual pressure excursion in the reactor building and that it had probably been caused by an explosion, and both were alarmed by their conclusions. Chwastyk agreed in testimony before us that his conclusion certainly “changed [his] view of how serious the situation had been up to then.” Whether their conclusions were communicated to anyone else in the control room on Wednesday, in particular to their superiors, is much less clear.
Mehler recalls that “we did inform the people in the [shift supervisor’s] office that we did have the pressure spike, and just about everyone in the control room knew it.” He has testified that he believed both Ross and Miller were “in there,” and that it is his recollection that both were informed that the spike indicated an actual pressure increase. However, Mehler does not recall a discussion with anyone other than Chwastyk about the possibility of a “chemical reaction,” although according to Mehler, he and Chwastyk did not discuss whether they should keep that possibility quiet.
Chwastyk, who apparently considered the possibility of a hydrogen explosion more concretely than Mehler, testified that he could not be certain he related this possibility to Miller, but his best recollection was that he had done so, even though he does not specifically recall that hydrogen was mentioned. Chwastyk definitely recalls that upon realizing that an explosion may have occurred, he sought Miller’s permission to begin to try to redraw the “bubble in the pressurizer.” In an unsworn but transcribed interview with Chwastyk on October 11, 1979, he related at one point that during this request he had explained to Miller what he thought had happened with an explosion. Later in the interview, Chwastyk said that he was “pretty sure” he told Miller about the explosion, but he was not sure he could swear to it:
:That’s what I thought. Most definitely I did think that [an explosion had probably occurred]. Now, whether or not I related that to Gary then, now that I think about it, I don’t really remember. I may have
just gone back to Gary and asked permission again to redraw the bubble. I just can’t remember if I related to him my thoughts at the time of the correlation of pressure spike in the operation of the valve.
In a sworn deposition on October 30, when pressed on the issue, Chwastyk testified the following:
:My best recollection of that is that I did relate to Gary that we had some sort of an explosion. Whether I said it was hydrogen or not, I’m not sure. But I remember distinctly putting together the operation of the valve and the spike, and I think I related those thoughts to Gary.
A few minutes later in the deposition, however, when it was pointed out to him that Miller’s best recollection was that he had not learned of such an explosion until Friday morning, Chwastyk testified, “that could very well be true. Again, I can’t absolutely-if Gary said-I may not have told him what I thought at the time, because I really wasn’t certain.”
Asked whether he recalled talking with anyone in the control room on Wednesday, in addition to Miller, about his fear that a hydrogen explosion might have taken place, Chwastyk testified that he “probably” discussed it with Mehler, and felt “sure” he had discussed it with other individuals, but could not recall specifically anyone else with whom he might have discussed it, other than an inspector from the NRC.
In spite of Chwastyk’s testimony, neither our investigation nor any of the other investigatory groups to whose depositions and interview memoranda we have had partial or complete access has been able to identify any Met Ed or B&W personnel who, on March 28, were told of or made the connection themselves between the pressure spike and the closing of a valve, or who considered it a possibility that the pressure spike represented an actual explosion. As pointed out, neither Miller, Kunder, Logan, nor Rogers has admitted that they even knew that there had been a pressure spike. Ross, Flint, and Zewe all remember knowing of the spike and the activation of the spray pumps, but did not conclude that the spike reflected an actual pressure surge nor did they conceive that an explosion could have occurred. Furthermore, an entry in the control room operator log book for the afternoon of March 28 notes that at 1:50 p.m. an engineered safeguards initiation signal was received, the reactor building sprays came on, and the reactor building pressure spiked up to 4 psi.
With respect to the question of Miller’s knowledge, both Chwastyk and Mehler, as noted, believe that Miller was informed of the pressure spike, but Miller insists he did not learn of it until Friday. Ross, in a deposition conducted by the SIG on September 18, 1979, testified that both he and Miller were in the control room when the spray pump actuation and pressure spike were reported and that:
:We, being Miller, I and the group, looked back and said, guess we just felt that it was either one: something we just didn’t understand, and we didn’t associate it with anything else and we just went on.
Ross seemed to think this had occurred after Miller returned from Harrisburg (which would have been after 4:30 p.m., several hours later). As he recalled, Zewe reported the spike to him, and “we came to the conclusion, be it right or wrong, that it was an electrical spike of some kind and not a pressure spike in the building.”
Pressed in a later deposition taken on October 30, Ross said that his recollection about having any conversation with Miller on this subject was “a little vague,” but he did remember being in the control room with Miller looking at some other instruments, the operators reporting that the building spray had come on, and Miller saying the following:
:Did you hear that, or did you feel that? Something to that effect. I’m not sure what that was. And we just kind of went right by it. We looked at it and we told Gary, it’s not time to get nervous now. We’re going to have to go from where we are. And that’s what we did too
Ross further explained that his reference to getting “nervous” was that he thought perhaps Miller was “hearing things” or “imagining things.” Ross did not recall if Miller actually looked at the pressure spike on the chart, and he does not recall any further discussion of the matter with Miller later in the afternoon when Miller returned from Harrisburg.
In previous statements and testimony, Miller has been quite consistent in his account of when he first learned of the pressure spike and the diagnosis of a possible hydrogen explosion. Confronted with the testimony of Chwastyk, Mehler, and Ross in a deposition on October 29, Miller continued to maintain that he did not learn of the pressure spike on March 28. Nor, he said, did he recall either a request from Chwastyk that afternoon to try to redraw a bubble in the pressurizer, or approving such a step in a conversation with Chwastyk.
Furthermore, Miller has pointed out to us in the course of the investigation that he had absolutely no reason to conceal or cause anyone else to conceal this information had he known it on Wednesday. And he testified under oath that at no time did he withhold significant information of any kind or instruct anyone else in the control room to withhold such information from the NRC inspectors in the control room or from the NRC itself.
Miller’s April 1979 written statement, prepared for presentation to various congressional subcommittees and NRC’s IE investigators, arose out of a joint effort to reconstruct the events of the first day of the accident by a number of individuals who were in the Unit 2 control room on March 28. After the accident, these individuals collegially attempted to create a reliable chronology of events. Miller has ointed out that those discussions were taped and that the tape would provide a more reliable indication of what various individuals knew or did not know the first day of the accident than their recollections today. He had also pointed out that during those conversations none of the individuals present recalled learning of a hydrogen burn or explosion the first day. We have reviewed a tape recording provided to us by Met Ed of this conversation, and it is consistent with Miller’s testimony. The tape (side 1, about 10 minutes in) indicates that Miller recalled hearing a noise and turning to another individual to ask what it was. The tape suggests that Miller may have heard about the spray system going on at about 5:00 p.m., when he returned from the Capitol, but does not make clear whether that information was linked with the pressure spike.
According to Miller’s recollection, he first learned that there had been a “hydrogen burn” from Met Ed consultant, Bill Lowe, when Miller came into work on Friday morning, March 30. He recalls that Lowe spoke to him about the pressure spike and diagnosis of it, and that Ivan Porter, the lead Instrumentation Control Engineer for Unit 2, showed him a graph of the spike on that same morning.
Porter, who tends to corroborate Miller’s account, testified in a deposition conducted by the SIG that he probably first learned of the pressure spike on the morning of the 30th, and that Gary Miller asked him to look at the charts and see if Porter thought they showed a “real valid indication versus, say, a malfunction of that pressure recorder the instrument supplied or whatever.” Porter undertook this assignment by looking at the strip charts for reactor coolant system pressure, which is measured relative to reactor containment building pressure. Observing that the graph for coolant system pressure de-
creased slightly for a short time at 1:50 p.m. (indicating a rise in building or reference level pressure), Porter concluded that the pressure spike had been a reliable indicator of building pressure, and communicated this information to Miller. Porter stated the following:
:My impression was, I think he had just learned of it, and wanted me to look at the charts and stuff involved and tell him what I saw. I don’t know, but I felt it was like somebody just told him, you guys had a hydrogen explosion, and he wanted me to take a look at the charts and stuff and tell him what I could find out from them.
In this connection it may be significant that both Chwastyk and Mehler recall that they did not really figure out that the pressure spike might have been caused by an explosion until sometime after the event. Miller recalls that it was perhaps 30 minutes after he heard the noise that afternoon that he departed for Harrisburg. But State logs show that Miller, Kunder, and Herbein arrived at the Lieutenant Governor’s office at 2:30 p.m., and Herbein’s own records show his departure from the Observation Center across the river from Three Mile Island at 1:55 p.m. It is probably very close to an hour trip from the plant and into Harrisburg, so Miller must have been virtually on his way out of the control room at the time the pressure spike occurred.
The timing of Chwastyk’s request to Miller to “redraw the bubble” in the pressurizer is therefore rather puzzling. According to Chwastyk he spoke to Miller after he had connected in his own mind the pressure spike and the valve operation. However, this conversation would probably have been after Miller left the control room. Such a conversation could have been held after Miller returned at 4:30 to 5:00 p.m., but charts of various system parameters, including pressurizer level, suggest that Chwastyk may have begun to try to redraw a bubble beginning around 2:00 p.m. that afternoon; his account of his attempt to manipulate pressurizer level fits with the observed parameters in a number of respects. On the other hand, Ross (Chwastyk’s superior after Miller left the control room, from 2:00 until 4:30 p.m.) testified that he recalls no attempt during this entire time to redraw a bubble in the pressurizer.
Did anyone, then, other than Chwastyk, clearly realize on March 28 that a hydrogen explosion may have occurred? At one point in our investigation it appeared that some light might be shed on this question by testimony that an order had been given on Wednesday not to start any electric equipment in the reactor building for fear of the consequences of a spark. In an unsworn but transcribed interview on October 11, Mehler recalled that although he had not connected the pressure spike and simultaneous operation of a valve at 1:50 p.m., somebody else obviously had. Later that afternoon he was told not to start oil lift pumps that must be run before the main reactor coolant pump can be started: “not to do anything that could give an ignition.” Mehler recalled responding to Miller that he had already tested these pumps and they were ready to go.
In a sworn deposition with Mehler several weeks later, he recalled that the instruction was “given in the shift supervisor’s office not to start anything electrical … there were other people in the room. They would have been aware of the instructions. I believe the instruction came from Gary Miller.” Mehler recalls telling Miller that he had already started the pumps, and recalls the comment being made, “Well, then I don’t think we have anything left in the building.” Mehler thinks it was somebody else, not himself, who made that comment.
At the October 11 interview, Mehler was quite certain that the instruction not to start electrical equipment had been issued on March 28, prior to the attempt to restart one of the main reactor coolant pumps Wednesday evening:
:I can say for a fact and will go under oath and I will take a lie detector test, prior to running the reactor core (sic) pump, someone did tell us not to start anything and I remember telling Gary, it’s too late now, I have already started them.
However, by the time of the October 30 sworn deposition, he had become quite unsure of his recollection as to which day he received the orders. In the deposition, Mehler testified that he believed the orders probably occurred either on Thursday, March 29, when he was also in the control room from about 1:00 p.m. until about 11:00 p.m., or on Friday, March 30.
Why the change in recollection? Mehler testified that between October 11 and 30, he had talked with others who had been in the control room on March 28 (Miller, Ross, Zewe, and Chwastyk), and that none of them recalled the instructions having been given that day. Some, however, had told him they remembered such an instruction being given after March 28. Therefore, Mehler concluded that he must have been mistaken. Mehler acknowledged that his “own recollection, faulty or not, standing alone, has been that it was the 28th,” but stated that it “seems funny, if I would be the only one that remembered it happening on the 28th when there were people in the room that don’t remember it.”
Chwastyk also recalled the instruction not to start electrical equipment in the reactor building because of the spark potential, and that someone had just started the oil lift pumps. Chwastyk testified that he did not believe this could have happened on March 28, because he recalled the instruction having been given while he was physically present in the shift supervisor’s office and he did not recall having been in that office at all on March 28.
No other witnesses recalled such an instruction having been given on Wednesday, March 28. Control room operator Theodore Lilies stated in an NRC investigation report that the hydrogen explosion was discussed in the control room on the 28th. The NRC found that Lilies was apparently mistaken and that the discussion occurred on the 29th. Our review of the evidence indicated that the NRC investigator’s conclusion was correct. Additionally, William Lowe of Pickard, Garrick & Lowe, consultants to Met Ed, recalled in a SIG interview that the recorder trace showing the pressure spike was brought to him after 9:00 p.m. on March 29, and that he then notified Met Ed, GPU, and B&W personnel.
The only contemporaneous documents that might shed light on when such an instruction was given are the control room log book and a set of notes taken throughout the first few days of the accident by two Met Ed employees, Don Barry and Walter “Bubba” Marshall, neither of whom is a licensed reactor operator. The notes, which are extremely sketchy for March 28, do not record such an instruction on any day. The control room log book contains an entry at 9:14 p.m. on Thursday the 29th: “Placed RCP lift pumps in off (minimize sparking potential in reactor building).”
Thus, an instruction not to start electrical equipment in the reactor building on account of an awareness of a potential hydrogen explosion apparently was issued shortly after 9:00 p.m. on Thursday. The log entry does not, however, necessarily reflect an awareness of the earlier explosion on Wednesday afternoon. Fear of the possibility of significant hydrogen in the reactor building due to fuel cladding failure could have given rise to such an order by Thursday night, even without knowledge or appreciation of the earlier pressure spike, because by that time the existence of a possible hydrogen “bubble” was beginning to be postulated.
Mehler recalls that it was Miller who gave the order; however, Miller’s best recollection is that he was not in the control room at 9:15 p.m. on Thursday evening when the instruction was given. In any event, Miller does not recall having given suchan instruction at any time.
Mehler’s recollection that the instruction not to start electrical equipment was given on March 28, was keyed to his recollection that it came just after he had started oil lift pumps for a reactor coolant pump. Mehler originally believed that this happened Wednesday because at that time the reactor coolant pumps were all off, and efforts were being made to restart at least one pump. No pumps were restarted on Thursday. However, preparations were apparently made Thursday or Friday to restart one of the nonoperating pumps in case the operating pump should for some reason trip off, and these preparations would require testing the oil lift pumps.
To summarize, only one person present in the control room on March 28-Chwastyk, the shift supervisor in charge of the consoles-has acknowledged that he realized the pressure surge was real, evaluated that in connection with a valve operation, and concluded that a hydrogen explosion had probably occurred in the reactor containment building. Others present either say they did not know of the pressure spike, or dismissed it as an electrical transient, except for Mehler, who feared that a chemical reaction had taken place but did not believe it was a hydrogen explosion (and therefore did not necessarily have a reason to believe that substantial core uncovering had occurred). Several say they heard a noise but believed it to have been the control room ventilation dampers cycling.
Chwastyk believes he related his conclusion to Miller, but is not certain that he did. Miller does not recall being aware of the pressure spike or the spray pumps coming on. Others testified that Miller probably was informed of these events. It may be significant that Kunder, who accompanied Miller to Harrisburg shortly thereafter, and a number of others in the control room have also testified that they were not aware of the spike.
According to documents tending to establish the time Miller and Kunder left for Harrisburg, Miller evidently departed the control room before Mehler and Chwastyk discussed the pressure spike among themselves and realized that there might have been an explosion. Thus, if Chwastyk related his conclusions to Miller, this would have occurred after Miller returned at about 4:30 or 5:00 p.m., at the time when Miller was instructed by Herbein to repressurize the system. However, such a supposition is inconsistent with Chwastyk’s testimony that he recalls getting permission from Miller to redraw a bubble in the pressurizer, and with the charts showing that such efforts may have been made around 2:00 p.m. Neither Miller nor Ross, however, recalls any attempts being made to redraw a bubble during the afternoon.
Except for Chwastyk’s testimony, no other evidence indicates that anyone in the control room realized on March 28 that there might have been a hydrogen explosion in the reactor building, or that Miller was aware of such a possibility. No evidence suggests that Met Ed management officials in the Observation Center or GPUSC officials in New Jersey were informed either about the pressure spike or about the possibility of any explosion. No evidence, either documents or testimony, establishes that NRC employees either at Region I Headquarters in King of Prussia, Pa., or in Washington were aware on Wednesday of the pressure spike or of the possibility that large amounts of hydrogen had probably been generated during the first 10 hours of the accident.
At the time the pressure spike occurred, at least two of the five NRC inspectors that had arrived at about 10:30 a.m., were in the area of the Unit 2 control room. They were James Higgins, a Reactor Operations Inspector, and Don Neely, a Health Physics expert. Two other inspectors who had come onto Three Mile Island, Charles Gallina and Ronald Nimitz, had gone to the Unit 1 control room. The fifth inspector, Karl Plumlee, was performing radiation surveys outside of the plant buildings. A second team, composed of an inspector and an investigator from Region I, Walter Baunack and Raymond Smith, arrived at TMI around noon, but did not get to the Unit 2 control room until approximately 3:00 p.m. or later.
In sworn depositions, neither Higgins nor Neely recalled having been made aware of the pressure spike. Higgins kept a spiral notebook of information to pass back to Region I, which had a direct open line to the Unit 2 control room during most of the afternoon, but his notes neither reflect the pressure spike or a possible pressure excursion in the reactor building, nor the actuation of the spray pump system. Nor have we discovered anyone at Region I or NRC Headquarters who recalls being made aware of the spike or the possibility that an explosion might have occurred.
However, quite a number of others in the control room recall that an inspector was standing in the control room and was in a position to observe the pressure spike directly and the ensuing response to it by control room operators and supervisors. For example, Zewe recalled that an NRC inspector was standing directly behind him when the spike occurred.
The most specific testimony relating to this question was given by Mehler. Mehler recalls that when he came out of the shift supervisor’s office, an NRC inspector followed him out of the office and “was behind us” when Mehler instructed the spray pumps turned off. 118 The inspector, according to Mehler, asked why he had given such an instruction, and Mehler explained to him that there had been a pressure spike but that the pressure had gone down, and that Mehler did not know what had caused the pressure increase. Mehler specifically recalls explaining to the inspector the coincident two-of-three logic of the spray pump system. According to Mehler’s testimony, the inspector did not seem to understand what had happened, and did not make
any further inquiry about it.
Mehler could not identify the inspector, but testified that the individual was not Neely, whom he knows from the TMI postaccident recovery effort. He described the inspector as being medium height, with dark hair and a little gray around the sideburns, aged perhaps 30. Mehler’s description does not fit James Higgins, who is a tall (well over 6 feet), thin man in his midthirties. According to Mehler, he recalls that the individual he talked to was wearing a white hard-hat with “U.S. NRC” emblazoned on it, also suggesting that perhaps the inspector had only recently entered the control room from elsewhere in the plant. (Mehler testified that the only NRC inspector in the control room on the 28th whom he recognized was Donald Haverkamp, the Project Inspector for TMI. Haverkamp was, in fact, in King of Prussia, Pa., on Wednesday and did not arrive at TMI until Thursday, March 29.)
Chwastyk also recalls that “there was an NRC inspector … standing behind Mehler when we shut down the spray pumps.” Chwastyk’s recollection is that the individual was about Mehler’s height (Mehler is about 5 feet, 8 inches tall), but somewhat heavier. Chwastyk did not know the names of any of the NRC personnel in the control room March 28, and does not think he could identify the individual he saw standing behind Mehler if he saw him today. Nor does he recall whether Mehler had a conversation with the inspector.
More important, Chwastyk testified that he had “some recollection of talking to someone from the NRC” about his conclusion that there might have been an explosion in the reactor building. Chwastyk does not recall whether it was the same inspector he observed standing behind Mehler when the spray pumps were turned off, but he does recall the following:
:Relating to someone from the NRC that I think we may have had an explosion in the building, but I wasn’t sure. And that was about it, and I was probably a little curt because I had other things to do.
Neither Higgins nor Neely recalls any such conversation, nor do Higgins’ notes record such information. It appears that the only other possibility is that Ronald Nimitz, who was stationed in the Unit I control room, may have been in the Unit 2 control room at this time and may have talked to Chwastyk and Mehler about the pressure spike. Nimitz was definitely in and out of the Unit 2 control room during the afternoon, according to his recollection. He does not recall learning about either the pressure spike or the possibility of an explosion, nor do his notes reflect such information. Nimitz is not a reactor operations inspector, but he has had nuclear engineering and reactor operators training, and he testified in a deposition taken by the SIG that had he been informed of the spray pumps actuating, such information would have been very significant to him and he would have recalled it.
When Chwastyk was pressed as to whether he told anyone else about his conclusion that there might have been an explosion in the reactor building, and as to why that fact did not seem to have been common knowledge until late Thursday or Friday morning, Chwastyk responded that he thought he had reported to Miller and the NRC inspector, and that he “must have talked to other people in the control room,” including his counterparts who relieved him late that night. But he could not recall anyone specifically with whom he had talked. He did not recall discussing the matter with Kunder, and was not sure about Mike Ross. Chwastyk denied that he, or he and Mehler, made any decision to hide the possibility that an explosion had occurred, but admitted that they “didn’t just make it general knowledge to everybody in the control room.” As for the NRC inspector, Chwastyk conceded that “maybe I should have explained more [to him] but I just didn’t have the time. At least I didn’t feel I had the time.” He recalled that his primary concern was dealing with the reactor coolant system.
What conclusions can be drawn, then, from this evidence as to whether Met Ed personnel willfully withheld significant information from the NRC? No evidence indicated that company management off the island was aware of the pressure spike or the possibility that an explosion had occurred on Wednesday, March 28. The top official on site, Gary Miller, the Station Manager, has consistently testified that he did not become aware of the spike or the possibility that an explosion had occurred until Friday morning. Testimony from more than one witness indicates that Miller was aware of the pressure surge on Wednesday, but this testimony is less than clearcut; some other testimony tends to corroborate Miller’s account.
Even if Miller knew of the pressure spike, possibly he dismissed it or failed to recognize its significance, because other supervisory employees who were in the control room and their coworkers did not believe that the spike actually reflected a pressure increase. Such an interpretation of the spike appears to have been technically deficient: at least two supervisors recognized that the actuation of the spray pumps probably compelled the conclusion that there had actually been a pressure excursion in the building. On the other hand, in the minds of most or all of those present in the control room, serious consideration of the possibility that a hydrogen explosion had occurred probably would have contradicted all the assumptions under which they were proceeding to try to cope with the accident.
Only Chwastyk’s testimony made Miller aware that an explosion might actually have occurred. Such testimony might give rise to an inference that Miller withheld such information from the NRC, except that Chwastyk is not certain that he did indeed tell Miller, and Chwastyk also testified that he himself informed an NRC inspector present of this possibility. Chwastyk also testified that he explained the pressure spike and its significance to the same or another NRC inspector right after the event.
There is no dispute that at least two NRC inspectors were on hand in the control room observing plant operations, and that no effort was being made to restrict their freedom to move about and ask any questions they wished to ask. The virtually unanimous testimony of Met Ed witnesses is that an NRC inspector was standing directly behind the console when the pressure spike occurred and the spray pumps went on, and was in a position not only to observe these events but also to hear the discussion among operators and supervisors about what had happened. However, none of the three NRC inspectors who may have known or been told about the spike and possibly about its consequences around midafternoon on the 28th presently recalls having been so informed. Chwastyk’s recollection, on the other hand, is that the inspector to whom he spoke did not really seem to have understood what was happening. In any event, none of the inspectors reported the event to Region 1 or to Headquarters in Bethesda.
Of course, if Chwastyk had really told an NRC inspector that a hydrogen explosion might have just occurred, and explained what that meant to him, it is hard to believe any inspector, or anyone else, would have taken such information lightly. Chwastyk was, in a de facto if not official sense, the shift supervisor in charge of the control room consoles. It could be argued that his failure to inform and alert responsible NRC officials of his conclusions and their implications represented a willful withholding of information from the NRC, that merely informing an inspector in the control room should not have been enough. However, it might also be argued that control room personnel can hardly be expected to double as NRC employees in the middle of an emergency, and that as long as NRC inspectors are present observing everything that happens, there can seldom be a “conscious withholding” of information”from the NRC.”
Perhaps most important, it does not appear to us that the legal requirements placed on reactor licensees to report significant safety-related information to the NRC were ever intended as a tool to compel effective flow of information to the NRC in the midst of an accident or disaster situation. If control room supervisors had consciously decided not to tell NRC inspectors on site a significant piece of information because they were afraid of the consequences of doing so, the legal requirements might well have applicability in this situation. We did not find evidence indicating that that occurred here. Moreover, it is unclear what motive Met Ed personnel would have had to hide such information from the NRC, when control room personnel were at the same time encouraging the NRC representatives present to provide any ideas that occurred to them to help cope with the unstable reactor.
