CN112036568A - Intelligent diagnosis method for damage fault of primary loop coolant system of nuclear power plant - Google Patents

Intelligent diagnosis method for damage fault of primary loop coolant system of nuclear power plant Download PDF

Info

Publication number
CN112036568A
CN112036568A CN202010659408.5A CN202010659408A CN112036568A CN 112036568 A CN112036568 A CN 112036568A CN 202010659408 A CN202010659408 A CN 202010659408A CN 112036568 A CN112036568 A CN 112036568A
Authority
CN
China
Prior art keywords
meta
state
proposition
rule
coolant system
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010659408.5A
Other languages
Chinese (zh)
Other versions
CN112036568B (en
Inventor
余刃
彭俏
王天舒
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Naval University of Engineering PLA
Original Assignee
Naval University of Engineering PLA
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Naval University of Engineering PLA filed Critical Naval University of Engineering PLA
Priority to CN202010659408.5A priority Critical patent/CN112036568B/en
Publication of CN112036568A publication Critical patent/CN112036568A/en
Application granted granted Critical
Publication of CN112036568B publication Critical patent/CN112036568B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/02—Knowledge representation; Symbolic representation
    • G06N5/022—Knowledge engineering; Knowledge acquisition
    • G06N5/025—Extracting rules from data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • G06N5/041—Abduction
    • G—PHYSICS
    • G21—NUCLEAR PHYSICS; NUCLEAR ENGINEERING
    • G21C—NUCLEAR REACTORS
    • G21C17/00—Monitoring; Testing ; Maintaining
    • G21C17/002—Detection of leaks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E30/00—Energy generation of nuclear origin
    • Y02E30/30—Nuclear fission reactors
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Plasma & Fusion (AREA)
  • High Energy & Nuclear Physics (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Monitoring And Testing Of Nuclear Reactors (AREA)

Abstract

The invention discloses an intelligent diagnosis method for damage faults of a primary circuit coolant system of a nuclear power plant, which comprises the following steps: setting a damage fault diagnosis rule of a primary coolant system of a loop of a nuclear power plant, and generating a knowledge matrix according to the diagnosis rule, wherein the knowledge matrix comprises a state vector P, an input transformation matrix I, an output transformation matrix O and a transfer trigger vector DT; and acquiring parameter signals corresponding to the meta propositions in the diagnosis rule in real time according to the diagnosis rule, determining the state of the meta propositions according to the parameter signals, assigning the state vectors to a state vector P, carrying out fault diagnosis reasoning calculation according to a knowledge matrix, and judging whether the obtained result is true or not, wherein if the result is true, the damage of a main coolant system of a loop of the nuclear power plant is indicated, and if the result is not true, the damage of the main coolant system of the loop of the nuclear power plant is indicated. The invention realizes the parallel reasoning of the diagnosis process through the matrix operation, thereby realizing the on-line intelligent diagnosis of the damage fault of the primary circuit coolant system of the nuclear power plant.

Description

Intelligent diagnosis method for damage fault of primary loop coolant system of nuclear power plant
Technical Field
The invention belongs to the technical field of nuclear power plant fault diagnosis, and particularly relates to an online intelligent diagnosis method for a primary coolant system damage fault of a primary circuit in the operation process of a pressurized water reactor nuclear power plant.
Background
According to statistics, more than half of the operation events of nuclear power stations in the world are related to misoperation and misjudgment of operators. The main reasons for the misoperation and misjudgment of operators are the complexity of the nuclear power plant and the huge psychological stress generated by operators when accidents happen. At present, the operating fault of the nuclear power plant is mainly judged by an operator according to the knowledge and experience by observing the operating parameters and the change thereof. To reduce or even avoid the occurrence of misjudgment and misoperation, an effective technical means is to provide real-time online operation fault automatic diagnosis capability for operators in a nuclear power plant instrument control system.
The failure of the primary coolant system of the primary circuit refers to the accidents that the pressure boundary of the primary coolant system of the primary circuit is damaged and the coolant leaks due to the fact that pipelines, valves and the like in the pipelines, the valves and the like of the primary coolant system of the primary circuit of the pressurized water reactor nuclear power device are damaged. The accident has a high probability, and if the accident cannot be found in time, serious consequences can be caused.
The nuclear power plant is a complex nonlinear, strongly coupled, dynamically time-varying system. The fault on-line intelligent diagnosis system realizes automatic analysis and processing of the variation condition of the operation parameters of the nuclear power device, and assists operators to accurately and timely judge the operation state and the faults of the nuclear power device so as to take correct treatment measures. Common fault diagnosis techniques can be divided into three major categories, mechanism model-based, data-driven-based, and knowledge rule-based.
The fault diagnosis method based on the mechanism model needs to establish an accurate mechanism or mathematical model of a diagnosed object. However, the nuclear power plant is very complex, and as the nuclear power plant is used and equipment ages, the characteristics of the nuclear power plant also change to a certain extent, it is difficult to establish accurate and complete mechanism models or mathematical models under normal and accident conditions, and the operation time is long, so that the method has a great gap from the requirements of practical application in real-time performance and accuracy.
The data-driven fault diagnosis method is realized on the basis of having a large amount of sample data in normal and fault states of the diagnosed equipment. Because the nuclear power plant is high in risk in a fault state, the operation data sample in the fault state cannot be obtained through experiments generally, and can only be obtained through simulation, but the true degree is limited. Therefore, although there are many theoretical research results in recent years, the feasibility and accuracy of the fault diagnosis method based on data driving are far from the practical requirements for operation fault diagnosis of the nuclear power plant.
People accumulate abundant operation experience in the long-term operation process of the nuclear power device, and a rich and complete fault judgment rule can be formed by combining theoretical analysis. Therefore, the automatic diagnosis of the operation fault of the nuclear power plant by adopting the knowledge-based artificial intelligence technology, namely an expert system, is a practical and feasible technical approach. A practical fault on-line automatic diagnosis expert system needs to solve two problems: firstly, how to convert the fault judgment rule in the natural language form into an expression form suitable for a computer program language; and how to effectively ensure the real-time performance of the fault judgment reasoning process.
The traditional expert system is based on production rules, which are consistent with the summary and expression mode of people on experience knowledge, so that the traditional expert system is convenient to understand, but is often complicated when expressed by a programming language. In addition, the inference engines of the traditional expert system carry out serial inference according to rule combinations, and when the rules are complex, the searching efficiency is low and the inference speed is low. When there are many knowledge rules, there may be problems such as rule conflict and reasoning circulation.
The invention content is as follows:
in order to overcome the defects of the background technology, the invention provides a set of method which can convert the knowledge of the fault diagnosis rule of the main coolant system of the primary circuit of the nuclear power plant of the pressurized water reactor based on natural language into a knowledge matrix which is easy to store and express by a computer, and realize the parallel reasoning of the diagnosis process through the matrix operation, thereby realizing the rapid automatic diagnosis of the fault of the main coolant system of the primary circuit of the nuclear power plant.
In order to solve the technical problems, the invention adopts the technical scheme that:
an intelligent diagnosis method for a damage fault of a primary circuit coolant system of a nuclear power plant comprises the following steps:
setting a damage fault diagnosis rule of a primary coolant system of a loop of a nuclear power plant, and generating a knowledge matrix according to the diagnosis rule, wherein the knowledge matrix comprises a state vector P, an input transformation matrix I, an output transformation matrix O and a transfer trigger vector DT; acquiring parameter signals corresponding to the meta-propositions in the diagnosis rule in real time, determining the states of the meta-propositions according to the parameter signals, assigning the states to a state vector P, carrying out fault diagnosis reasoning calculation according to a knowledge matrix, and judging whether the obtained result is true or not, wherein if the result is true, the damage of a main coolant system of a loop of the nuclear power plant is represented, and if the result is not true, the damage of the main coolant system of the loop of the nuclear power plant is represented.
Preferably, the failure diagnosis rule for the damage of the primary coolant system of the loop of the nuclear power plant comprises the following steps:
rule one, If p1 and p6 and p5 then p10；
Rule two, If rule one holds and If p9 and p8 and p4 and p3 and p2 and p10 then p11；
Rule three, If rule two holds or If p7 then p11；
Wherein, meta proposition p1When the state of (1) indicates that the average temperature of a loop is not reduced, the meta-proposition p1When the state of (1) is 0, the average temperature of a primary circuit is reduced;
meta proposition p2When the state of (1) indicates that the two-loop water dosage does not exceed the standard, the meta-proposition p2When the state of (1) is 0, the secondary loop water dosage exceeds the standard;
meta proposition p3When the state of (1) indicates that the safety valve of the voltage stabilizer is not opened, the meta-proposition p3When the state of (1) is 0, the safety valve of the voltage stabilizer is opened;
meta proposition p4When the state of (1) indicates that the release valve of the voltage stabilizer is not opened, the meta-proposition p4When the state of (1) is 0, the release valve of the voltage stabilizer is opened;
meta proposition p5When the state of (1) indicates that the water level of the voltage stabilizer is reduced, the meta-proposition p5When the state of (1) is 0, the water level of the voltage stabilizer is not reduced;
meta proposition p6When the state of (1) indicates that a loop drain valve is not opened, the meta-proposition p6When the state of (1) is 0, the primary circuit drain valve is opened;
meta proposition p7When the state of (1) indicates that there is a containment dose alarm, p7When the state of (1) is 0, indicating that no containment dosage is high and alarming;
meta proposition p8When the state of (1) indicates that no equipment has high cooling water dosage and is alarming, p8When the state of (1) is 0, the device cooling water amount is high and an alarm is given;
meta proposition p9When the state of (1) indicates no leakage alarm of the auxiliary system, p9When the state of (1) is 0, indicating that the auxiliary system has a leakage alarm;
meta proposition p10When the state of (1) indicates that a break occurs in the pressure boundary of the primary circuit, p10When the state of (1) is 0, the pressure boundary of the primary circuit is not broken;
meta proposition p11When the state of (1) indicates a failure of the primary coolant system of the primary circuit, p11A state of (1) indicates that the primary coolant system of the primary circuit is not broken.
Preferably, the parameter signal corresponding to each meta-proposition in the diagnosis rule includes: the device comprises a primary circuit average temperature, a secondary circuit water dosage high alarm signal, a safety valve discharge pipe temperature high alarm signal, a release valve discharge pipe temperature high alarm signal, a voltage stabilizer water level, a primary circuit drain valve opening state signal, a containment dosage high alarm signal, a device cooling water dosage high alarm signal and other auxiliary system leakage alarm signals.
Preferably, the method of generating the state vector P comprises: the elements of the state vector P correspond to the current states of all meta-propositions in the rules, the meta-propositions are ordered in the order they appear in turn in the three rules, and are not repeated, the final conclusion meta-proposition is arranged last in the vector P,
P＝[p1，p6，p5，p10，p9，p8，p4，p3，p2，p7，p11]。
preferably, the method for generating the transition trigger vector DT includes:
element DT [ i ] in transition trigger vector DT]Is the ith transition tiThe number of corresponding conditional element propositions is a row vector, and the dimension number corresponds to the number of transitions, namely the number of pieces of the diagnosis rule; the number of conditional meta propositions in rule one is 3, DT [ 1]]3; the number of conditional meta propositions in rule two is 6, DT [ 2]]6; the number of conditional meta propositions in rule three is 1, DT [3]1 is ═ 1; transition trigger vector DT ═ 3,6,1]。
Preferably, the method for generating the input transformation matrix I comprises:
the rows in the input transformation matrix I correspond to 11 meta propositions in a damage fault diagnosis rule of a primary coolant system of a loop; the columns correspond to three transitions in rule one, rule two, and rule three; the ordering of the rows in the input transformation matrix I is strictly consistent with the meta-propositional ordering in P. The ordering of the columns is strictly consistent with the ordering of the branches.
Preferably, the method for generating the output transformation matrix o includes:
the rows in the output transformation matrix O correspond to three transitions in rule one, rule two and rule three, the columns correspond to 11 meta-propositions, and the ordering is strictly consistent with the meta-proposition in P.
Preferably, the method for performing fault diagnosis reasoning calculation according to the knowledge matrix includes the steps of obtaining the state of each meta proposition in the state vector P according to the read input signal value of the parameter corresponding to each meta proposition of the state vector P, assigning the state to the state vector P, and performing fault diagnosis reasoning calculation according to the knowledge matrix:
(1) obtaining the initial state of each element proposition in the state vector P according to the read input signal, and assigning to the initial state vector P0Let P be P ═ P0；
(2) Calculating P multiplied by I, and correcting the result of P multiplied by I according to DT to obtain T;
(3) calculating T multiplied by O, correcting elements larger than 1 in the T multiplied by O result to be 1, and assigning the obtained result to S;
(4) calculating S + P0Adding S + P0Assigning the result of (1) to S;
(5) if P is not equal to S, making P equal to S, and turning to the step (2) for circulation; if P is equal to S, obtaining the numerical value of a proposition corresponding element P11 in P, if equal to 1, the result is expressed as true, and judging that a primary circuit coolant system damage fault occurs; otherwise, judging that the damage fault of the primary circuit coolant system does not occur.
The invention has the beneficial effects that: the invention provides a practical and efficient fault diagnosis knowledge expression and reasoning method convenient for computer storage and operation for fault diagnosis of a main coolant system of a loop of a nuclear power plant. A fault diagnosis rule knowledge expression system based on meta propositions and a knowledge matrix and a parallel reasoning algorithm based on matrix operation are designed, so that the expression and storage of diagnosis rule knowledge in a computer are greatly facilitated, the speed of a diagnosis reasoning process can be greatly improved, the technical bottleneck that the reasoning speed is sharply reduced along with the increase of rules of a traditional expert system is overcome, and the real-time performance and the realizability of the on-line intelligent diagnosis system for the damaged fault of a primary coolant system of a nuclear power plant loop are ensured.
The failure diagnosis expert knowledge expression system of the nuclear power plant loop main coolant system based on the failure diagnosis knowledge matrix is convenient for computer storage, calling and expansion, and is convenient for rule conflict and endless loop detection; the designed inference algorithm based on knowledge matrix operation can convert a complex logic inference process into simple matrix operation and has parallel inference capability. The method is applied to the damage fault diagnosis of the primary circuit coolant system of the nuclear power plant, can greatly facilitate the expression and storage of diagnosis knowledge, improve the operation efficiency of a fault diagnosis expert system, and realize the complete stripping of reasoning algorithm and data, thereby being easy to continuously expand new knowledge rules in the use process, continuously improving and enhancing the capability of the fault diagnosis system and having good expandability.
Drawings
FIG. 1 is a schematic diagram of a pressurized water reactor nuclear power plant system;
FIG. 2 is a flow chart of a loop primary coolant system failure determination process according to an embodiment of the present invention;
FIG. 3 illustrates a process for initializing a knowledge matrix according to an embodiment of the present invention;
FIG. 4 is a process for diagnosing operating faults of a nuclear power plant in accordance with an embodiment of the present invention;
fig. 5 is a flowchart of an algorithm for performing fault diagnosis by using the knowledge matrix according to the embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings and examples.
(1) Form a damaged fault diagnosis rule knowledge matrix of a primary coolant system of a primary loop of a pressurized water reactor nuclear power plant
The hydraulic reactor nuclear power plant is shown in fig. 1, the primary coolant system damage fault determination process is shown in fig. 2, and the diagnostic rules are generated as follows:
if (the average temperature of a primary circuit does not drop, the water level of a drain valve of the primary circuit does not start and a pressure stabilizer continuously drops, no equipment cooling water dosage is high, an alarm and auxiliary system does not leak, a release valve of the pressure stabilizer does not start, a safety valve of the pressure stabilizer does not start, and the water dosage of a secondary circuit does not exceed the standard) or a safety shell dosage high alarm) or a primary coolant system of the primary circuit is damaged.
A fault diagnosis expert knowledge expression system based on a knowledge matrix achieves the purpose of easy computer expression and storage of fault diagnosis knowledge of a main coolant system of a nuclear power plant loop through the definition of model elements and the design of a knowledge matrix generation method, and lays a foundation for realizing automatic diagnosis of operation faults.
1) Definition of knowledge matrix constituent elements
a. Meta proposition
The generative rule form of the diagnostic knowledge of the damage fault of the primary circuit coolant system of the nuclear power plant can be expressed as follows:
If(p1 and p2…)or(p3 and p4…)……,then A
wherein p is1、p2、p3… …, etc. are conditional propositions and A is a conclusion proposition.
Proposition piThe configuration of (a) is typically a determination of a condition, such as whether a parameter is greater than, less than, or equal to a value (e.g., the average temperature of a circuit has not dropped), whether a device state has a phenomenon, attribute (e.g., a circuit drain valve has not opened), etc. These propositions have only two states, "true" or "false", both of which are defined herein as meta propositions. Conclusion proposition a is also one of meta propositions, which is "loop main coolant system breakage".
"meta proposition" refers to the smallest logical decision unit in the fault diagnosis generative rule knowledge. It can no longer be broken down into a combination of two or more logical decisions. It can be input meta proposition and conclusion meta proposition (usually having specific physical meaning, such as average temperature of circuit is greater than 300 deg.C, or opening of water discharge valve, etc.), or intermediate result of logic reasoning (i.e. intermediate meta proposition, possibly having no specific physical meaning), using piAnd (4) showing.
The meta-propositions are determined according to the production rule of the damage fault diagnosis knowledge of the primary coolant system of the nuclear power plant loop, and are shown in the table 1.
TABLE 1 Definitions table of damaged fault diagnosis rule elements of main coolant system of loop
Figure BDA0002577969470000091
Figure BDA0002577969470000101
In Table 1, p10Is a meta proposition, here with physical meaning, namely: a break is formed in a pressure boundary of a loop; p is a radical of11The conclusion meta-proposition, others are input meta-propositions.
b. Diagnostic knowledge generation rules
Based on the definition of meta propositions in table 1, the production rule of the loop main coolant system damage fault diagnosis knowledge can be symbolized as:
rule one is as follows: if p1 and p6 and p5 then p10；
Rule two: if p9 and p8 and p4 and p3 and p2 and p10 then p11；
Rule three: or Ifp7 then p11。
The above rules can be divided into two categories, namely (1) if the rule header starts with an "or", this indicates that the rule is in an "or" relationship with other rules; (2) if the rule header directly begins with an "if" then this indicates that the rule is an "and" relationship with other rules.
With respect to the three rules above, it can be seen that each rule has an inference conclusion that includes a "then" character. In addition, each rule has only one or operator at most, and if a plurality of or operators exist, the rule can be split into a plurality of rules.
c. Status of state
And (5) referring to the judgment result of the meta proposition. Its value is alpha (p)i) Meaning that there are two values of 1 and 0. When the judgment result of a meta proposition is true, the state value is 1, otherwise, the state value is 0. The state value of the conditional element proposition depends on the judgment result of the nuclear power device operation parameter (or equipment operation state, etc.) contained in the conditional element proposition at the current moment, and the state value of the conclusion or the intermediate element proposition depends on the logic operation result of the preceding element proposition.
d. The state vector P
The elements of the state vector P correspond to the current state of all meta-propositions in the rule. Except that the final proposition of the final proposition is arranged at the end of the vector P, other propositions are ordered according to the sequence which sequentially appears in the rule. If meta-propositions appear multiple times in each rule, they are arranged in the order of first appearance, and there is no duplication of elements in P.
The value of a certain element in the vector is 1, and the meta proposition p corresponding to the element is representediIs true, the state of the meta-proposition is activated (i.e., alpha (p)i) 1); the element value is 0, and the opposite is true (i.e., alpha (p)i) 0). For simplicity of expression, in the following definitions and calculations, p is used directlyiState value alpha (p) representing its corresponding meta propositioni)。
For the fault diagnosis rule of the damage of the main coolant system of the loop, the state of 11 meta propositions (the state of 9 input meta propositions, 1 intermediate meta proposition and 1 conclusion meta proposition) is total. Its state vector is:
P＝[p1，p6，p5，p10，p9，p8，p4，p3，p2，p7，p11] (1)
e. transfer of
Transition represents logical AND operation, which means that the state of some intermediate element proposition or conclusion element proposition changes in the process of diagnosis and inference. A transition may have 1 or more conditional meta-propositions (inference logic conditions, which may be input meta-propositions, or intermediate meta-propositions), and 1 result meta-proposition (inference logic results, which may be intermediate meta-propositions, or final conclusion meta-propositions). When the states of all conditional meta-propositions of a certain transition are 1, the transition is triggered, and the conclusion meta-proposition state thereof becomes 1, indicating that the transition has occurred. Transfer by tiAnd (4) showing.
"then" in a rule represents a "state transition". For the fault diagnosis of the damage of the primary coolant system of the primary loop, 3 rules are provided, which correspond to three transfer t respectively1、t2、t3。
f. Transition trigger vector DT
Element DT [ i ] in transition trigger vector DT]Is the ith transition tiThe number of corresponding input element propositions. Is a row vector and the dimensions correspond to the number of metastases, i.e. the number of pieces of the diagnostic rule.
For a primary coolant system breakage fault diagnosis rule, the transfer trigger vector DT is as follows:
DT＝[3,6,1] (2)
g. input transformation matrix I
The rows in the input transformation matrix I correspond to 11 meta propositions in a damage fault diagnosis rule of a primary coolant system of a loop; the columns correspond to 3 transitions in the inference rule. The value of a certain element in the matrix I is 1, and the element proposition corresponding to the row where the element is located is the input element proposition of the transfer corresponding to the column where the element is located; if the value of a certain element is 0, the meta-proposition corresponding to the row of the element is not the input meta-proposition of the transition corresponding to the column of the element.
The ordering of the rows in the input transformation matrix I should be strictly consistent with the meta-propositional ordering in P. The ordering of the columns is strictly consistent with the ordering of the branches.
For a primary coolant system damage fault diagnosis rule of a loop, the input transformation matrix I is as follows:
Figure BDA0002577969470000131
h. output transformation matrix O
The rows in the output transformation matrix o correspond to 3 transitions in the inference rule, and the columns correspond to 11 meta-propositions in the inference rule. The value of a certain element is 1, which indicates that the meta-proposition corresponding to the column of the element is the output meta-proposition of the transfer corresponding to the row of the element; and when the value of a certain element is 0, the meta-proposition corresponding to the column of the element is not the output meta-proposition of the transition corresponding to the row of the element.
The ordering of the columns in the output transformation matrix o should be strictly consistent with the meta-propositional ordering in P. The ordering of the rows is strictly consistent with the ordering of the branches.
For a primary coolant system damage fault diagnosis rule of a primary loop, an output transformation matrix O is as follows:
Figure BDA0002577969470000132
2) damaged fault diagnosis knowledge matrix of primary coolant system of primary loop
Under the premise of the above definition, the state vector P, the input transformation matrix I, the output transformation matrix o, and the transition trigger vector DT together form a knowledge matrix N of the fault diagnosis rule knowledge network model { P, I, O, DT }, so that the input conditions and inference logic of the fault diagnosis rule knowledge are completely expressed. The knowledge matrix expression mode only needs to store 4 vectors or matrices in the computer, thereby greatly reducing the storage capacity requirement of the computer on the diagnosis knowledge and laying a foundation for the follow-up reasoning algorithm based on the knowledge matrix.
In the implementation process of the diagnosis method, firstly, the knowledge matrix is generated according to the constitution of the diagnosed nuclear power plant and the expert knowledge of the damage fault diagnosis of the primary coolant system of the loop, and is stored in a knowledge base of the diagnosis system, so as to complete the initialization work of the diagnosis system. The initialization procedure is shown in fig. 3. This is typically done before the diagnostic system is put into operation and the diagnostic rule knowledge matrix for new fault types may be added continuously during use.
(2) Fault diagnosis reasoning process based on matrix operation
The operation fault diagnosis process of the nuclear power plant is a repeated and cyclic process of collecting required signals according to a certain time period, and performing fault judgment through diagnosis, reasoning and calculation, as shown in fig. 4.
The fault diagnosis process of reasoning one by one according to the generation type diagnosis rule is converted into a simple and efficient processing process based on knowledge matrix operation, so that the operation efficiency of a fault diagnosis expert system can be greatly improved, and the real-time requirement of the operation fault diagnosis of the nuclear power device is met.
After the knowledge matrix N of the failure diagnosis of the primary coolant system of the nuclear power plant is established as { P, I, O, DT }, the inference algorithm steps for performing the failure diagnosis using the knowledge matrix are shown in fig. 5.
According to the damage fault diagnosis rule element proposition definition of the primary coolant system of the primary loop and the signals required to be collected, which are determined in the table 1, two test working conditions of fault occurrence and fault non-occurrence are respectively set for verifying the actual application effect. The initial states of each meta proposition under two conditions are shown in table 2.
TABLE 2 initial state table of damage fault diagnosis element of primary coolant system of primary circuit under two working conditions
Figure BDA0002577969470000151
As previously described.
For a damage fault of a primary coolant system of a loop, a knowledge matrix N { P, I, O, DT } of a diagnosis rule is constructed in an initialization process and is as follows:
state vector P ═ P1，p6，p5，p10，p9，p8，p4，p3，p2，p7，p11]
Inputting a transformation matrix:
Figure BDA0002577969470000161
outputting a transformation matrix:
Figure BDA0002577969470000162
the transition trigger vector DT is [3,6,1 ].
The following is a detailed description of the diagnostic reasoning and calculation process in the implementation process of the damage fault diagnosis of the primary coolant system of the primary loop under two test conditions:
step 1, reading operation parameters of the nuclear power device, determining an initial activation state of each element proposition according to the operation parameter values, assigning an initial value to a state vector P, and assigning the value according to the following method: the first element of the P vector is a meta proposition P1I.e. "the average temperature of the primary circuit has not decreased". During diagnosis, the average temperature of a loop at the current moment is read and compared with the previous oneComparing the time values, if not, then meta proposing p1If true, the first element of the P vector is assigned a value of 1; if it is down, meta-proposition p1False, the first element of the P vector is assigned a value of 0. And the assignment process of other elements is analogized:
assuming the working conditions: p ═ 1, 1, 1, 0, 1, 1, 1, 1, 1, 0;
and under the assumption of a second working condition: p ═ 1, 1, 1, 0, 0, 0;
and step 2, carrying out reasoning operation and outputting a reasoning conclusion.
a. A first working condition calculation step:
taking input signal as initial state vector P0Assignment, P0＝[1，1，1，0，1，1，1，1，1，1，0]And let P be P ═ P0；
(xi) P × I ═ 3, 5, 1], corrected according to DT to give T ═ 1, 0, 1;
③S＝T×O＝[0，0，0，1，0，0，0，0，0，0，1]calculating S as S + P0＝[1，1，1，1，1，1，1，1，1，1，1]If S is not equal to P, the inference is not completed, and the intermediate node P is updated10The state of (1). Let P ═ S ═ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]Entering the next circulation;
t ═ P × I ═ 3,6,1], corrected according to DT to give T ═ 1, 1, 1;
the element with the value larger than 1 indicates that a plurality of transitions occur at the upstream of the node, and because of OR operation, one transition occurs, the propositional state of the downstream element can be changed, so the element with the value larger than 1 in S is corrected to be 1, and the S is [0, 0, 0, 1, 0, 0, 0, 0, 0, 1 ];
calculating S ═ S + P0＝[1，1，1，1，1，1，1，1，1，1，1]If S is P, the reasoning is completed. The element P [11] corresponding to the proposition of the term in P]And (5) outputting a diagnosis conclusion as 1: a failure in the primary coolant system of the primary loop occurs.
b. And calculating a working condition II:
according to the input signal as initial stateState vector P0Assignment, P0＝[1，1，1，0，1，1，1，1，0，0，0]And let P be P ═ P0；
(xi) P × I ═ 3, 4, 0], corrected according to DT to give T ═ 1, 0, 0;
③S＝T×O＝[0，0，0，1，0，0，0，0，0，0，0]calculating S as S + P0＝[1，1，1，1，1，1，1，1，0，0，0]If S is not equal to P, the inference is not completed, and the intermediate node P is updated10The state of (1). Let P ═ S ═ 1, 1, 1, 1, 1, 1, 1, 0, 0, 0]Entering the next circulation;
t ═ P × I ═ 3, 5, 0], corrected according to DT to give T ═ 1, 0, 0;
⑤S＝T×O＝[0，0，0，1，0，0，0，0，0，0，0]；
calculating S ═ S + P0＝[1，1，1，1，1，1，1，1，0，0，0]If S is P, the reasoning is completed. The element P [11] corresponding to the proposition of the term in P]And (5) outputting a diagnosis conclusion as 0: failure of the primary coolant system of the primary circuit does not occur.
Therefore, the inference conclusion is consistent with the preset conclusion of the working condition.
This example verifies the correctness of the algorithm.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.

Claims (8)

1. An intelligent diagnosis method for a damage fault of a primary circuit coolant system of a nuclear power plant is characterized by comprising the following steps:
setting a damage fault diagnosis rule of a primary coolant system of a loop of a nuclear power plant, and generating a knowledge matrix according to the diagnosis rule, wherein the knowledge matrix comprises a state vector P, an input transformation matrix I, an output transformation matrix O and a transfer trigger vector DT; and acquiring parameter signals corresponding to the meta-propositions in the diagnosis rule in real time, determining the states of the meta-propositions according to the parameter signals, assigning the states to the state vector P, carrying out fault diagnosis reasoning calculation according to the knowledge matrix, and judging whether the obtained result is true or not, wherein if the result is true, the damage of a main coolant system of a loop of the nuclear power plant is represented, and if the result is not true, the damage of the main coolant system of the loop of the nuclear power plant is represented.
2. The intelligent diagnostic method for the fault of the primary coolant system of the nuclear power plant as claimed in claim 1, wherein the fault diagnostic rule for the fault of the primary coolant system of the nuclear power plant comprises:
rule one is as follows: ifp1andp6andp5thenp10；
Rule two: if rule one holds andIfp9andp8andp4andp3andp2andp10thenp11；
Rule three: if rule two holds or If p7thenp11；
Wherein, meta proposition p1When the state of (1) indicates that the average temperature of a loop is not reduced, the meta-proposition p1When the state of (1) is 0, the average temperature of a primary circuit is reduced;
meta proposition p2When the state of (1) indicates that the two-loop water dosage does not exceed the standard, the meta-proposition p2When the state of (1) is 0, the secondary loop water dosage exceeds the standard;
meta proposition p3When the state of (1) indicates that the safety valve of the voltage stabilizer is not opened, the meta-proposition p3When the state of (1) is 0, the safety valve of the voltage stabilizer is opened;
meta proposition p4When the state of (1) indicates that the release valve of the voltage stabilizer is not opened, the meta-proposition p4When the state of (1) is 0, the release valve of the voltage stabilizer is opened;
meta proposition p5When the state of (1) indicates that the water level of the voltage stabilizer is reduced, the meta-proposition p5When the state of (1) is 0, the water level of the voltage stabilizer is not reduced;
meta proposition p6When the state of (1) indicates that a loop drain valve is not opened, the meta-proposition p6When the state of (1) is 0, the primary circuit drain valve is opened;
meta proposition p7When the state of (1) indicates that a containment vessel is presentHigh dose alarm, p7When the state of (1) is 0, indicating that no containment dosage is high and alarming;
meta proposition p8When the state of (1) indicates that no equipment has high cooling water dosage and is alarming, p8When the state of (1) is 0, the device cooling water amount is high and an alarm is given;
meta proposition p9When the state of (1) indicates no leakage alarm of the auxiliary system, p9When the state of (1) is 0, indicating that the auxiliary system has a leakage alarm;
meta proposition p10When the state of (1) indicates that a break occurs in the pressure boundary of the primary circuit, p10When the state of (1) is 0, the pressure boundary of the primary circuit is not broken;
meta proposition p11When the state of (1) indicates a failure of the primary coolant system of the primary circuit, p11A state of (1) indicates that the primary coolant system of the primary circuit is not broken.
3. The intelligent diagnosis method for the breakage fault of the primary loop coolant system of the nuclear power plant as recited in claim 1 or 2, wherein the parameter signal corresponding to each meta-proposition in the diagnosis rule comprises: the device comprises a primary circuit average temperature, a secondary circuit water dosage high alarm signal, a safety valve discharge pipe temperature high alarm signal, a release valve discharge pipe temperature high alarm signal, a voltage stabilizer water level, a primary circuit drain valve opening state signal, a containment dosage high alarm signal, a device cooling water dosage high alarm signal and auxiliary system leakage alarm signals.
4. The intelligent diagnostic method for the breakage fault of the primary circuit coolant system of the nuclear power plant as recited in claim 2, wherein the method for generating the state vector P comprises: the elements of the state vector P correspond to the current states of all meta-propositions in the rules, the meta-propositions are ordered in the order they appear in turn in the three rules, and are not repeated, the final conclusion meta-proposition is arranged last in the vector P,
P＝[p1，p6，p5，p10，p9，p8，p4，p3，p2，p7，p11]。
5. the intelligent diagnostic method for the breakage fault of the primary circuit coolant system of the nuclear power plant as recited in claim 2, wherein the method for generating the transition trigger vector DT comprises the following steps:
element DT [ i ] in transition trigger vector DT]Is the ith transition tiThe number of corresponding conditional element propositions is a row vector, and the dimension number corresponds to the number of transitions, namely the number of pieces of the diagnosis rule; the number of conditional meta propositions in rule one is 3, DT [ 1]]3; the number of conditional meta propositions in rule two is 6, DT [ 2]]6; the number of conditional meta propositions in rule three is 1, DT [3]1 is ═ 1; transition trigger vector DT ═ 3,6,1]。
6. The intelligent diagnosis method for the damage fault of the primary loop coolant system of the nuclear power plant as claimed in claim 2, wherein the method for generating the input transformation matrix I comprises the following steps:
the rows in the input transformation matrix I correspond to 11 meta propositions in a damage fault diagnosis rule of a primary coolant system of a loop; the columns correspond to three transitions in rule one, rule two and rule three; the ordering of the rows in the input transformation matrix I is strictly consistent with the meta-propositional ordering in P. The ordering of the columns is strictly consistent with the ordering of the branches.
7. The intelligent diagnosis method for the failure fault of the primary coolant system of the nuclear power plant as set forth in claim 2, wherein the method for generating the output transformation matrix o comprises:
the rows in the output transformation matrix O correspond to three transitions in the rule I, the rule II and the rule III, and the columns correspond to 11 meta-propositions and are in the same order as the meta-propositions in P.
8. The intelligent diagnosis method for the breakage fault of the primary circuit coolant system of the nuclear power plant as recited in claim 5, wherein: according to the read input signal values of the parameters corresponding to the meta propositions of the state vector P, the state of each meta proposition in P is determined and assigned to the state vector P, and the method for carrying out fault diagnosis reasoning calculation according to the knowledge matrix comprises the following steps:
(1) determining the initial state of each element proposition in the state vector P according to the read input signal, and assigning to the initial state vector P0Let P be P ═ P0；
(2) Calculating P multiplied by I, and correcting the result of P multiplied by I according to DT to obtain T;
(3) calculating T multiplied by O, correcting elements larger than 1 in the T multiplied by O result to be 1, and assigning the obtained result to S;
(4) calculating S + P0Adding S + P0Assigning the result of (1) to S;
(5) if P is not equal to S, making P equal to S, and turning to the step (2) for circulation; if P is equal to S, obtaining the numerical value of a proposition corresponding element P11 in P, if equal to 1, representing that the result is true, and judging that a primary coolant system damage fault of the loop occurs; otherwise, judging that the damage fault of the primary coolant system of the primary loop does not occur.
CN202010659408.5A 2020-07-09 2020-07-09 Intelligent diagnosis method for damage faults of primary loop coolant system of nuclear power device Active CN112036568B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010659408.5A CN112036568B (en) 2020-07-09 2020-07-09 Intelligent diagnosis method for damage faults of primary loop coolant system of nuclear power device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010659408.5A CN112036568B (en) 2020-07-09 2020-07-09 Intelligent diagnosis method for damage faults of primary loop coolant system of nuclear power device

Publications (2)

Publication Number Publication Date
CN112036568A true CN112036568A (en) 2020-12-04
CN112036568B CN112036568B (en) 2023-10-17

Family

ID=73579085

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010659408.5A Active CN112036568B (en) 2020-07-09 2020-07-09 Intelligent diagnosis method for damage faults of primary loop coolant system of nuclear power device

Country Status (1)

Country Link
CN (1) CN112036568B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103699117A (en) * 2013-12-18 2014-04-02 中广核核电运营有限公司 Method and system for diagnosing failure based on actual working conditions of nuclear power plant and simulation system
CN106597965A (en) * 2016-12-07 2017-04-26 中国船舶重工集团公司第七〇九研究所 Nuclear power apparatus running state monitoring system and monitoring method
CN107301884A (en) * 2017-07-24 2017-10-27 哈尔滨工程大学 A kind of hybrid nuclear power station method for diagnosing faults
US20170316315A1 (en) * 2016-04-27 2017-11-02 Foundation Of Soongsil University Industry Cooperation Ontology-based reasoning apparatus and method using knowledge of an expert
CN109445411A (en) * 2018-10-24 2019-03-08 西华大学 A kind of Substation fault diagnosis
CN109491816A (en) * 2018-10-19 2019-03-19 中国船舶重工集团公司第七六研究所 Knowledge based engineering method for diagnosing faults
CN110289112A (en) * 2019-06-14 2019-09-27 清华大学 Nuclear power unit health status diagnostic method based on step analysis and fuzzy evaluation
CN110738274A (en) * 2019-10-26 2020-01-31 哈尔滨工程大学 nuclear power device fault diagnosis method based on data driving

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103699117A (en) * 2013-12-18 2014-04-02 中广核核电运营有限公司 Method and system for diagnosing failure based on actual working conditions of nuclear power plant and simulation system
US20170316315A1 (en) * 2016-04-27 2017-11-02 Foundation Of Soongsil University Industry Cooperation Ontology-based reasoning apparatus and method using knowledge of an expert
CN106597965A (en) * 2016-12-07 2017-04-26 中国船舶重工集团公司第七〇九研究所 Nuclear power apparatus running state monitoring system and monitoring method
CN107301884A (en) * 2017-07-24 2017-10-27 哈尔滨工程大学 A kind of hybrid nuclear power station method for diagnosing faults
CN109491816A (en) * 2018-10-19 2019-03-19 中国船舶重工集团公司第七六研究所 Knowledge based engineering method for diagnosing faults
CN109445411A (en) * 2018-10-24 2019-03-08 西华大学 A kind of Substation fault diagnosis
CN110289112A (en) * 2019-06-14 2019-09-27 清华大学 Nuclear power unit health status diagnostic method based on step analysis and fuzzy evaluation
CN110738274A (en) * 2019-10-26 2020-01-31 哈尔滨工程大学 nuclear power device fault diagnosis method based on data driving

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
彭俏;余刃;: "模糊Petri网专家系统及其在核动力装置故障诊断中的应用", 核动力工程, no. 1, pages 73 - 76 *
王天舒;余刃;刘笑凡;: "核动力装置运行故障诊断系统设计研究", 核动力工程, no. 02, pages 180 - 183 *

Also Published As

Publication number Publication date
CN112036568B (en) 2023-10-17

Similar Documents

Publication Publication Date Title
JPH0522241B2 (en)
WO2023241272A1 (en) Method for automatically generating concrete dam defect image description on basis of graph attention network
CN109472097B (en) Fault diagnosis method for online monitoring equipment of power transmission line
Zhou A fuzzy petri-net approach for fault analysis considering factor influences
Liu et al. Enhanced graph-based fault diagnostic system for nuclear power plants
Liu et al. Diagnosis of break size and location in LOCA and SGTR accidents using support vector machines
Wu et al. Research on rapid source term estimation in nuclear accident emergency decision for pressurized water reactor based on Bayesian network
Do Koo et al. Prediction of nuclear reactor vessel water level using deep neural networks
Park et al. Leak flow prediction during loss of coolant accidents using deep fuzzy neural networks
Groth et al. " Smart Procedures": Using dynamic PRA to develop dynamic context-specific severe accident management guidelines (SAMGs).
CN112036568B (en) Intelligent diagnosis method for damage faults of primary loop coolant system of nuclear power device
CN115564247A (en) Intelligent identification and decision-making method and system for nuclear power plant accidents
CN115712874A (en) Thermal energy power system fault diagnosis method and device based on time series characteristics
Santoso et al. Nuclear plant fault diagnosis using probabilistic reasoning
Liu et al. An accident diagnosis method of pressurized water reactor based on BI-LSTM neural network
CN114580549A (en) Nuclear power plant fault diagnosis method and device based on CNN-LSTM
Zavisca et al. A bayesian network approach to accident management and estimation of source terms for emergency planning
Wu et al. Online fault diagnosis of nuclear power plants using signed directed graph and fuzzy theory
Siqiao et al. Research on artificial intelligent nuclear power plant emergency operating guide
JPH0524521B2 (en)
CN113837535B (en) Method for backtracking severe accident process of nuclear power plant
Lee et al. Concept of robust AI with meta-learning for accident diagnosis
Hanna et al. Benchmarking an AI-Guided Reasoning-Based Operator Support System on the Three Mile Island Accident Scenario
Park et al. Abnormal diagnosis using explainable artificial intelligence in nuclear power plants
Sun et al. Research on Fault Diagnosis of Reactor Coolant Accident in Nuclear Power Plant Based on Radial Basis Function and Fuzzy Neural Network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant