LETTER TO THE NRC

last update October 3, 1998

January 8, 1997

Mr. Stuart A. Richards
USNRC
Washington, DC 20555

Dear Sir:

This letter presents my concerns about Reactor Operator training in the important subject of reactor behavior. You indicate that this issue will be addressed in the NRC allegation review process as allegation number NRR-96-A-0129. Previous letters on this topic are dated January 10, January 31, February 28, October 23, October 25, November 13, November 30, and December 22, 1996.

I allege that licensing of Reactor Operator candidates uses erroneous and defective fundamentals training material that fails to provide the understanding of reactor behavior needed to safely operate a nuclear reactor.

This allegation is supported by the following:

1. Technical Errors: My letter dated February 28, 1996, submitted 20 test items from the 1989 Institute of Nuclear Power Operations’ Generic Fundamentals Test Item Catalog as examples of items related to reactor behavior that were technically incorrect. These errors included a misdefinition of k_eff , a misapplication of subcritical multiplication, and a misuse of reactor rate, all important fundamental concepts of reactor behavior. Your letter dated May 14, 1996, concurred that 10 of the submitted items were indeed incorrect but indicated that most were no longer part of the current INPO Catalog. You did not concur that the remaining 10 items were incorrect. My letter dated October 23, reiterated that the remaining 10 questions were technically incorrect, and provided further explanation as to why. My follow letter dated October 25, 1996, listed 10 errors from the 1984 General Electric BWR Academic Series, Reactor Theory, a reference that you cited in support of non-concurrence.

Since that exchange, a detailed review of the current INPO Catalog section on Reactor Theory and a cover-to-cover review of the GE BWR manual reveals that:

• 170 of 650 test items are technically incorrect (26% of total)
• 170 of the remaining 480 test items are defective (26% of total)
• 175 technical errors are contained in the 315 pages of the GE Manual

2. Omissions: Courses constituting Class Room fundamentals training are intended to provide background support and understanding, a knowledge base, and prerequisite education for the plant specific training that follows. Although reactor behavior is part of fundamentals training, it far outranks other subjects because of its direct linkage to the Control Room. There are only two meters in the Control Room that provide information vital to the operator in determining the nuclear status of the reactor. These are the power meter and the period meter. The power meter includes three overlapping nuclear instruments, the Source Range meter, the Intermediate Range meter, and the Power Range meter, which provide a measure of the magnitude of power production from reactor shutdown to full power. The period meter provides indispensable information about the rate of power change with time.

As currently taught, reactor behavior treats the concept of source multiplication, which is a manifestation of the many on going chain reactions and that which determines the magnitude of power production, as a phenomenon that is unique to the subcritical reactor. Since chain reactions are the underlying process for all modes of operation, from shutdown to full power, the omission of source multiplication beyond subcritical conditions is a serious impairment to student understanding of reactor behavior.

Your letter dated May 14, 1996, indicates that the definition of transient reactor period need not include a term representing the rate of reactivity change with time ... that the stable reactor period definition suffices for all operational evolutions. As currently taught at many training centers, the reactor period definition includes three terms, which allow for its application from shutdown to full power, for both the steady state and the transient state. Without doubt, this equation is the most powerful and effective innovation ever developed for student understanding of transient reactor behavior. The omission of reactivity rate makes it impossible to explain response that the student will observe on the period meter in the Control Room, response that operators rely upon in conducting important sensitive evolutions, such as reactor startup.

These two omissions in current teaching of reactor behavior introduce a virtual disconnect between the Class Room and the Control Room meters, creating a grievous obstacle to effective training.

3. Ramifications: The INPO Generic Fundamentals Test Item Catalog can be likened to the proverbial tail that wags the dog. Class Room instruction of Reactor Operator candidates in the important subject area of reactor behavior is literally dictated by the test items contained in the Catalog, for the following reasons:

• Class Room instruction is tailored to include test items as topics
• Test items are used for example questions and problems in the Class Room
• The subject of reactor behavior is taught so as to pass the Licensing Exam
• Practice Licensing exams are based on selections from the INPO Catalog
• Students are provided with personal copies of the INPO Catalog for self study
• References, such as the 1984 GE BWR manual, are used as course text books

To compound this situation, Reactor Operators actually believe that the material provided in the Class Room is what they need as a knowledge base in the Control Room. Unbeknownst to them, they have been short-changed. Reactor Operators do not understand how a reactor works. Yet, they are delegated the ultimate responsibility of diagnosing complex situations in the Control Room that threaten reactor safety, while their Class Room training fails to provide real understanding for even the simplest of reactor transients.

There are those who maintain that in the hierarchy of reactor safety, of redundancy, of procedures, and of protection in depth, when all else fails, the Reactor Operator is the last line of defense. Whether this be the case, or not, erroneous training is not something to be tolerated. One of the reasons that errors have survived was revealed after Three Mile Island when a regulator declared:

"We screwed up -- and I mean "we" the nuclear community ... If you want to use the term "complacency" to describe our behavior, I won’t quibble with you ... I had the attitude that reactors were fairly forgiving, in the sense that they could withstand a lot of problems without having those problems turn into serious accidents. I don’t feel that way anymore."

4. Safety Risk: In your letter dated December 23, 1996, you state, "To aid us in our review, we request that, to the extent possible, you support your concerns with specific instances where such training problems exist, identify the specific problem and explain why it is a problem and explain how the problem is safety significant, i.e. is there any documentable link to the problems(s) you identify to a safety significant event in a nuclear power plant."

• As indicated under Technical Errors, "where the training problem exists" is in the INPO Generic Fundamentals Test Item Catalog, in the GE BWR Academic Series, Reactor Theory, and in Instructor Lesson Plans and Student Handouts in all BWR training facilities in the United States.

• My letters dated February 28, October 23, and October 25, "identify the specific training problem ... (i.e. the test item errors and errors in an important reference) ... and explain why it is a problem." Simply put, the problem is that the test items and the reference are wrong!

• As to "how the problem is safety significant", I submit that staffing Control Rooms with personnel that have a knowledge base built upon a foundation of misinformation and error about reactor behavior is not conducive to ensuring safety.

• "Documenting linkage of problems identified to safety significant events in a nuclear power plant" is not a valid criteria for determining the merit of this problem. The mere fact that the test items are wrong is full and sufficient reason for corrective action. However, it is not as though the linkage between poor, unsatisfactory training and safety significant events has never been made. We have been forewarned:

And:

"... there were a number of important factors not within the operator’s control that contributed to human failure ... these included inadequate training ... a lack of diagnostic skill ... the operators on duty had not received training adequate to ensure that they would be able to recognize and respond to a serious accident ... there is a clear need to restructure and improve operator training." ... Rogovin Report 1980

Even though these findings were part of a post-mortem on a core meltdown, they have gone unheeded and been ignored, at least as far as the teaching of Reactor Theory in the Class Room is concerned. Interestingly, a corroboration of these findings exists ... one of the charges after the Chernobyl meltdown was:

"The entire Soviet chain of command lacked understanding of reactor behavior." ... G. Medvedev

5. Solution: A possible strategy for correcting this problem might begin with three priorities.

• First -- correct all erroneous test items
• Second -- correct all defective test items
• Third -- develop a standardize Class Room text for reactor behavior

"As for the operational sector of the plant -- there is no doubt in my mind that operators ... can always be trained better ... and the operators are the guys who are always there in the middle of the night." ... Kenneth Carr 1990

In this instant, "trained better" means trained correctly, not trained with misinformation and error. This problem has a solution, albeit not an easy one. The benefits to the industry in quality of training and in enhanced reactor safety are well worth the effort.

Sincerely yours,

Robert G. Stater