CN-121983363-A - Multi-state sensing reactor-loop self-adaptive water filling method

CN121983363ACN 121983363 ACN121983363 ACN 121983363ACN-121983363-A

Abstract

The invention provides a multi-state sensing reactor loop self-adaptive water filling method, which relates to the technical field of reactor cooling and comprises the following steps of collecting water flow related characteristic data in a loop, calculating overall change values of water flow, collecting data of temperature and water level in a reactor, comparing the collected data with initial temperature and water level of the reactor before time T, comparing the total change values of water flow velocity, water flow temperature and water flow pressure with the overall change coefficients of the reactor, and quantifying how the change of each item of data of water flow in the loop affects the temperature and water level of the reactor through continuous comparison analysis. According to the invention, the disturbance source is judged through a dynamic comparison mechanism of the overall change value of the water flow and the overall change value of the reactor, and meanwhile, the system can dynamically switch the control mode according to the comparison result and can adaptively adjust the parameters, so that the water filling adjustment is faster and more accurate.

Inventors

WANG CONG
ZHOU LONGYU
ZHOU ZHAOHUA
HU CHEN
YAN XIAOZHE
LIU GUANGYAN
LIU XIAOJIE
SONG FEIFEI
LI LIANG
YU RAN
LIAO YI
HE CHUAN
LIU FENFEN
WANG HUI
LI REN

Assignees

中国船舶集团有限公司第七一九研究所

Dates

Publication Date: 20260505
Application Date: 20260205

Claims (10)

1. A multi-state-aware reactor-loop adaptive water filling method, comprising the steps of: s1, collecting relevant characteristic data of water flow in a loop, wherein the relevant characteristic data comprise water flow speed, water flow rate, water flow temperature and water flow pressure; s2, after preprocessing the related characteristic data, respectively comparing the water flow velocity, the water flow rate, the water flow temperature and the water flow pressure with the values of the water flow velocity, the water flow rate, the water flow temperature and the water flow pressure before the time T, and calculating the total change value of the water flow; S3, collecting data of the temperature and the water level in the reactor, comparing the data with the initial temperature and the water level of the reactor before time T, and calculating the overall change coefficient of the reactor; S4, comparing the overall change value of the water flow with the overall change coefficient of the reactor, and judging the conditions in the loop and the reactor through the comparison result of the overall change value of the water flow and the overall change coefficient in the reactor; S5, through continuous comparison analysis, when abnormality of each item of data of water flow is detected, and when the water level and the temperature of the reactor are not changed significantly, coolant leakage or pump failure which occurs in advance is early warned, and if the temperature of the reactor is increased abnormally, but the water flow parameter is normal, the power in the reactor core is abnormal.
2. A multi-state sensing reactor-loop self-adaptive water filling method according to claim 1 is characterized by comprising the steps of preprocessing water flow velocity, water flow quantity, water flow temperature and water flow pressure of water flow related characteristic data in a loop, sequentially performing physical range validity check on collected original data, marking the data which do not pass verification as invalid or unreliable when the data do not pass verification, triggering an alarm, generating high-reliability characteristic data values for the redundancy data which pass the verification by adopting a weighted fusion algorithm based on historical accuracy and current state, performing real-time digital filtering on the fused data flow to inhibit noise, and performing online compensation and correction on the original readings according to the latest calibration coefficient and the calibration validity period stored in a database.
3. The multi-state aware reactor-loop adaptive water filling method of claim 1, wherein the overall variation of the water flow is calculated by the following formula: wherein v (t) represents the water flow velocity of the first loop at the current moment, q (t) represents the water flow rate of the first loop at the current moment, w (t) represents the water flow temperature of the first loop at the current moment, and p (t) represents the water flow pressure of the first loop at the current moment; Representation of A corresponding value of the flow velocity of the water flow before the time; Representation of A corresponding value of the water flow before time; Representation of A corresponding value of the water flow temperature before time; Representation of The corresponding value of the water flow pressure before time, the total change value of the water flow is as follows: 。
4. A multi-state sensing reactor-loop self-adaptive water filling method is characterized by comprising the steps of carrying out voting and consistency verification on reactor temperature data based on triple redundancy measurement and space distribution logic, eliminating abnormal measuring points, carrying out real-time coupling analysis on a reactor temperature change trend and a neutron fluence rate change trend, marking the temperature data as data to be verified and preferentially adopting heat balance calculation verification data when the reactor temperature change trend is abnormal, carrying out cross verification on reactor water level data based on diversity measurement on the data, carrying out dynamic density compensation on water level reading by combining current system pressure and temperature, carrying out systematic verification on the reactor water level data, and automatically enabling a data reconstruction mechanism based on interpolation of adjacent measuring points and estimation of a physical model to generate high-reliability reactor temperature and water level data when a main measurement channel of parameters is judged to be unreliable.
5. The multi-state sensing reactor-loop self-adaptive water filling method as claimed in claim 1, wherein the change value of each item of data of temperature and water level change in the reactor is calculated by the following formula: wherein m (t) represents the temperature of the reactor at the current moment, h (t) represents the water level of the reactor at the current moment; Representation of A temperature corresponding value of the reactor before time; Representation of A corresponding value of the water level of the reactor before time, wherein the total change value of the reactor is 。
6. The multi-state sensing self-adaptive water filling method for a reactor, as set forth in claim 1, wherein the overall change value of the water flow and the overall change coefficient of the reactor itself include outliers, further processing is required, firstly, eliminating outliers, calculating a standard deviation sigma of the change value, eliminating extreme points exceeding + -3 sigma, setting a threshold value, forcibly setting to zero when the change value is smaller than or equal to the threshold value, and calculating and comparing when the change value is larger than the threshold value.
7. The multi-state-aware reactor-loop adaptive water filling method of claim 1, wherein when the overall change value S of the water flow is greater than the overall change value F of the reactor, the dynamic change of the one-loop coolant system is more remarkable than the dynamic change of the core state of the reactor, S > F corresponds to the rapid change of the flow rate, the temperature or the pressure parameter in the one loop, and the reactor temperature and the pressure stabilizer water level do not change to the same extent due to the thermal inertia and the system inertia; Under the self-adaptive water filling frame, S > F is used as an early warning signal to prompt the control system to intervene in advance, increase the sensitivity of water filling adjustment, reduce the control dead zone or improve the gain of the controller so as to respond to the water level change trend more quickly, simultaneously predict the water level change based on the change trend of the water flow parameters, adjust the water filling flow in advance, reduce overshoot or oscillation, and switch from the conventional feedback control to the compound control mode comprising multi-parameter prediction.
8. The multi-state aware reactor-loop adaptive water filling method of claim 7, wherein when the overall change in water flow S is less than the overall change in reactor F, the dynamic change in reactor core state is more severe than the dynamic change in a loop coolant system, and the system is switched from normal regulation mode to fast response mode; In the fast response mode, firstly, the control dead zone is reduced, the response is immediately made to a small water level deviation, secondly, the gain of the controller is increased, so that the change amplitude of the water filling flow is larger, the deviation is corrected more quickly, and the additional water filling quantity requirement is estimated in advance based on the increase amplitude of F.
9. A multi-state aware reactor-loop adaptive water filling method according to claim 8, wherein when the overall change value S of the water flow is equal to the overall change value F of the reactor, the dynamic change of the loop coolant system is matched with the dynamic change of the core state of the reactor in overall amplitude, and when S=F is detected, whether the cancellation is equal is identified. And observing whether the equation relation of S and F is continuously maintained or instantaneously coincided, continuously maintaining the direction of coordinated transient or system drift, wherein instantaneous coincided is only one point in fluctuation, and when the reason cannot be immediately known, adopting a conservation strategy, maintaining a water filling strategy, improving the monitoring frequency and not performing aggressive regulation.
10. A multi-state sensing reactor-loop self-adapting water filling method as set forth in claim 9 wherein the transient coincidence is that at the sampling time, the calculated S value is exactly equal to the F value, but immediately after the sampling time before the sampling time, the S and F are separated, each wave, no longer maintain the equality relationship, a time duration threshold is set at which the difference between S and F is required to be maintained for 5-10 consecutive sampling periods, equality below the threshold is regarded as noise, no special mode switching or control action is triggered, and the duration is that the S and F are maintained for a significant and continuous time, the numerical equality is maintained or a stable proportional relationship is presented, at which the system response conforms to the expected model.

Description

Multi-state sensing reactor-loop self-adaptive water filling method Technical Field The invention relates to the technical field of reactor cooling, in particular to a multi-state sensing reactor-loop self-adaptive water filling method. Background In the operation of a nuclear reactor, the water content control of a loop coolant system is a key safety function for ensuring the cooling of a reactor core and preventing the overheating of fuel, and the loop of the reactor is a highly coupled complex system, and early information of the system state is contained in the multi-parameter dynamic changes such as the flow rate, the temperature, the pressure and the like of the coolant, so that the coupling relation and the influence mechanism among the multi-state parameters need to be fully fused and analyzed, and the loop of the reactor can quickly respond when the working condition is changed rapidly or early faults occur, so that the predictive adjustment based on the overall dynamic state of the system is realized. In the prior art, publication number CN120388771A discloses a pressurized water reactor loop main pump flow prediction and control method and system, the technology comprises the steps of obtaining an original main pump flow signal of the pressurized water reactor loop, using a HOFEM module of NL FCNN to conduct multi-scale decomposition on the main pump flow signal to obtain multi-scale main pump flow, inputting the multi-scale main pump flow component into a ANSA dynamic optimization network structure to conduct feature extraction to obtain a multi-scale feature vector, using NL FCNN to conduct main pump flow signal prediction based on the extracted multi-scale feature vector to obtain an optimized main pump flow prediction value, converting the optimized main pump flow prediction value into an electric signal through a digital-to-analog conversion circuit based on a D ADC algorithm, inputting the electric signal into a main pump frequency converter to drive the main pump to operate in a frequency conversion mode, and controlling the main pump flow of the pressurized water reactor loop. However, in the prior art, the stable loading of the primary loop coolant is the key for guaranteeing the safe cooling of the reactor core, and the traditional water filling control method mainly depends on a fixed threshold value of a single or few parameters such as the reactor water level to respond, has the limitations of lag response, incapability of prejudging trend, difficulty in distinguishing the root of a fault and the like, and particularly when facing complex working conditions such as slow leakage of coolant, degradation of pump performance or abnormal internal power and the like, the traditional method is difficult to timely and accurately adjust the water supplementing amount to influence the safety and economy of the system, so that an intelligent water filling method capable of comprehensively sensing multiple states of the primary loop and the reactor core and realizing early warning and self-adaptive decision by analyzing the dynamic association of the primary loop and the multiple states of the reactor core is urgently needed to improve the active safety capability of coping with transients and faults. The above information disclosed in the background section is only for enhancement of understanding of the background of the disclosure and therefore it includes information that does not form the prior art that is already known to those of ordinary skill in the art. Disclosure of Invention The present invention is directed to a multi-state-aware reactor-loop adaptive water filling method, which solves the problems set forth in the above-mentioned background art. According to the invention, the disturbance source is intelligently judged through three relations of S > F, S < F, S =F by a dynamic comparison mechanism of the overall change value of the water flow and the overall change value of the reactor, so that the primary fault source distinction is realized, and meanwhile, the system can dynamically switch the control mode according to the comparison result and can adaptively adjust the parameters, so that the water filling adjustment is faster and more accurate. In order to achieve the above purpose, the present invention provides the following technical solutions: A multi-state-aware reactor-loop adaptive water filling method, comprising the steps of: And S1, collecting relevant characteristic data of water flow in a loop, wherein the relevant characteristic data comprise water flow speed, water flow quantity, water flow temperature and water flow pressure. S2, after preprocessing the related characteristic data, respectively comparing the water flow velocity, the water flow quantity, the water flow temperature and the water flow pressure with the values of the water flow velocity, the water flow quantity, the water flow temperature and the water flow pressure before the time T, and calculating