Intelligent evaluation and optimization regulation and control method for peak regulation and frequency modulation of nuclear power unit

Abstract

The invention provides an intelligent evaluation and optimization regulation and control method for peak regulation and frequency modulation of a nuclear power unit, and relates to the technical field of evaluation and optimization of a power generation system. Defining a fuzzy set for the operation data of the nuclear power unit, determining a vibration amplitude level, formulating a fuzzy rule, and obtaining a fuzzy evaluation result of the nuclear power unit in the aspects of stability and reliability by using a fuzzy reasoning system according to input real-time data and the fuzzy rule; constructing nodes of the Bayesian network, and determining a probability relationship between the nodes according to historical data and expert knowledge; setting an optimization variable of a simulated annealing algorithm and defining a target function; setting particle swarm particles, and combining a particle swarm optimization algorithm with a simulated annealing algorithm to prevent the particles from falling into a local optimal solution; the nuclear power unit and the intelligent power grid establish real-time communication connection and share operation state information of the two parties. The evaluation accuracy is improved, and the peak regulation and frequency modulation strategy is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of evaluation and optimization of electric power generation systems, and in particular to an intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units. Background Art

[0002] In recent years, with the increasing tension in the world's energy supply and the increasingly severe climate and environmental problems, vigorously developing and absorbing clean energy has become a consensus in the energy and power industry around the world. At present, my country's wind power, nuclear power and other clean energy power generation has been rapidly developed. Large-scale random, intermittent wind power and large-capacity nuclear power bases are connected to the grid, which puts higher requirements on the ramp rate of thermal power and other units. In addition, with the rapid development of industry and the continuous improvement of people's living standards, the peak-to-valley difference of power load has increased year by year, and the traditional power system is facing greater pressure in peak load regulation. The role of nuclear power in power grid peak load regulation has always attracted much attention. Power grid peak load regulation refers to adjusting the output power of generator sets in the power system according to the changes in load to meet the changing needs of load.

[0003] According to a comprehensive risk quantification assessment method for nuclear power units participating in grid peak regulation disclosed in China with publication number "CN111882212A", the method comprises the following steps: S1, analyzing the mechanism of the impact of nuclear power units participating in grid peak regulation on the safe operation of the units; S2, screening the safety risk factors affecting the nuclear power units participating in grid peak regulation; S3, establishing risk quantification indicators, and uniformly integrating and constituting comprehensive risk quantification indicators for nuclear power units participating in grid peak regulation; S4, calculating the annual number of failures and annual power loss of nuclear power units corresponding to the peak regulation mode and the base load mode; S5, calculating the increment of safety risk caused by the nuclear power units participating in grid peak regulation. The present invention can be used for optimization and scheduling calculation of power systems containing nuclear power units, and the parameters required for the calculation are easy to obtain and highly practical.

[0004] The above patent documents and prior art have the following technical problems when used:

[0005] Problem 1: In the above documents and existing technologies, the assessment of the stability and reliability of nuclear power units is relatively simple and static, which is difficult to adapt to the changes in different operating stages and working conditions. There is a lack of effective means to explore the potential modes and laws of nuclear power unit operation, resulting in limited ability to predict and optimize the operating status of the unit, resulting in insufficient accuracy in the assessment of nuclear power unit parameters;

[0006] Problem 2: The above document only considers a single factor, such as power fluctuation or response speed, for peak and frequency regulation of nuclear power units, which results in limited optimization effect and ignores important factors such as operating costs, resulting in inaccurate evaluation results and insufficient strategy optimization.

[0007] Question three: In the above-mentioned documents and existing technologies, the coordination between nuclear power units and smart grids is relatively loose, and information sharing is insufficient, resulting in the inability to synchronize information data in a timely manner, making it difficult to achieve efficient energy management. Summary of the invention

[0008] In view of the deficiencies in the prior art, the present invention provides an intelligent evaluation and optimization control method for peak-shaving and frequency regulation of nuclear power units, which solves the problems of evaluation accuracy and insufficient optimization of peak-shaving and frequency regulation strategies in the prior art.

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units, comprising the following steps:

[0010] Real-time collection of nuclear power unit operation data;

[0011] Defining fuzzy sets for the operation data of the nuclear power unit, determining the vibration amplitude level and formulating fuzzy rules, and using a fuzzy inference system to obtain fuzzy evaluation results of the stability and reliability of the nuclear power unit based on the input real-time data and fuzzy rules;

[0012] Constructing nodes of a Bayesian network according to the operation data of the nuclear power unit, determining the probability relationship between the nodes according to historical data and expert knowledge, and updating and optimizing the probability relationship of the Bayesian network through a learning algorithm and input new observation data;

[0013] According to the operation data of the nuclear power unit, the optimization variables of the simulated annealing algorithm are set and the objective function is defined. The simulated annealing algorithm optimization starts from a random initial solution, combines the needs of the smart grid, accepts a poor solution according to the current temperature and the objective function, and updates the optimal solution;

[0014] Particle swarm particles are set according to the operation data of the nuclear power unit, and the particles are updated according to their own historical optimal positions and the optimal positions of the group. At the same time, the particle swarm optimization algorithm is combined with the simulated annealing algorithm to prevent the particles from falling into a local optimal solution;

[0015] The nuclear power unit establishes a real-time communication connection with the smart grid to share the operating status information of both parties.

[0016] The present invention provides an intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units.

[0017] Beneficial effects:

[0018] 1. The present invention adopts the intelligent evaluation of fuzzy logic and Bayesian network combined with Internet of Things technology. By adopting an adaptive adjustment mechanism for fuzzy sets in fuzzy logic evaluation, the parameter fuzzy set division is dynamically adjusted according to different operating stages and working conditions of nuclear power units, and the multivariable fuzzy rule base, expert experience and data analysis are combined to determine the fine fuzzy rules and priorities, which can more accurately evaluate the stability and reliability of the unit. At the same time, Bayesian network evaluation is used, and a hybrid learning algorithm is used to determine the probabilistic relationship between nodes, and historical data and real-time observation data are fully utilized to explore the potential modes and laws of nuclear power unit operation, further optimize the Bayesian network structure and parameters, and improve the reliability of the evaluation.

[0019] 2. The present invention adopts a combined control of simulated annealing algorithm and particle swarm optimization to integrate smart grid collaborative technology. The simulated annealing algorithm optimizes the various parameters of the unit by combining the needs of the smart grid, accepting poor solutions with a certain probability and reducing the acceptance probability as the temperature decreases, and constantly updating the optimal solution. This can enable the nuclear power unit to better adapt to the changes in grid load and renewable energy supply during the peak and frequency regulation process, and improve energy utilization efficiency. At the same time, the particles in the particle swarm optimization are updated according to their own historical optimal positions, the optimal positions of the group, and external excellent solutions, and the particle update rules are adjusted in combination with the real-time status of the smart grid. The rapid convergence of the particle swarm optimization algorithm is used to accelerate the search process and improve the optimization efficiency. At the same time, it is combined with the simulated annealing algorithm to avoid falling into the local optimal solution, further optimize the power regulation parameters, and make the peak and frequency regulation strategy of the nuclear power unit more efficient.

[0020] 3. The present invention establishes a real-time communication connection between nuclear power units and smart grids to share operating status information, adjusts peak-shaving and frequency regulation strategies according to the needs of the smart grid, and performs collaborative optimization with other energy systems. Through the coordinated control of the smart grid, the complementarity and collaboration between different energy systems are achieved, and a closer two-way feedback mechanism is established. When the peak-shaving and frequency regulation strategies of the nuclear power units are adjusted, the smart grid not only promptly feeds back adjustment suggestions based on the new operating status, but also transmits the real-time demand information of the power grid to the nuclear power units in more detail. The smart grid can provide accurate power regulation targets and time requirements to the nuclear power units based on load forecasts in different time periods and the output of renewable energy. The nuclear power units further optimize the peak-shaving and frequency regulation strategies based on this information to achieve more efficient energy management. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0022] Figure 2 A data transmission path diagram of the present invention;

[0023] Figure 3 A line graph of sensor data and fused data in data collection and transmission of the present invention;

[0024] Figure 4 A bar graph of sensor weight coefficients in data collection and transmission of the present invention;

[0025] Figure 5 It is a fuzzy set line graph of vibration amplitude in fuzzy logic evaluation of the present invention;

[0026] Figure 6 It is a line graph of stability fuzzy sets in fuzzy logic evaluation of the present invention;

[0027] Figure 7 A histogram of the posterior probability distribution of random nodes for the Bayesian network evaluation of the present invention;

[0028] Figure 8 A line graph showing the change of the objective function value with the number of iterations in the simulated annealing algorithm optimization of the present invention;

[0029] Fig. 9 It is a line graph showing the change of the objective function value with the number of iterations in the particle swarm optimization of the present invention. DETAILED DESCRIPTION

[0030] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] Example 1

[0032] like Figure 1-9 As shown, the intelligent evaluation and optimization control method of peak load and frequency regulation of nuclear power units is as follows:

[0033] Sp1: Data collection and transmission: Various types of high-precision sensors, including vibration sensors, temperature sensors, and pressure sensors, are installed at key parts of nuclear power units through the Internet of Things technology. The sensors collect unit operation data in real time and transmit it to the central evaluation system through wireless networks. In order to improve the reliability of data collection and transmission, a multiple redundant sensor layout strategy is adopted. Multiple sensors of the same type are installed at key parts of the nuclear power unit. The data of multiple sensors are processed through data fusion technology to improve the accuracy and reliability of the data. At the same time, a data verification mechanism is set up to verify the collected data in real time. When abnormal data is found, an alarm signal is issued in time, and data recovery technology is used or switched to a backup sensor to ensure that the central evaluation system receives accurate unit operation data. In addition, a data storage and backup system is established to store and back up the collected data regularly to prevent data loss;

[0034] Sp2: Fuzzy logic evaluation: define fuzzy sets for the collected vibration data and other parameters, divide the vibration amplitude into "very low", "low", "moderate", "high" and "very high" levels, formulate fuzzy rules to evaluate the stability and reliability of the unit, combine the vibration amplitude, temperature, pressure and other parameters to determine the stability state of the unit, and use the fuzzy reasoning system to obtain the fuzzy evaluation results of the unit in terms of stability and reliability based on the input real-time data and fuzzy rules. For the definition of fuzzy sets in fuzzy logic evaluation, an adaptive adjustment mechanism is further adopted to dynamically adjust the fuzzy set division of parameters such as vibration amplitude, temperature, and pressure according to the characteristic changes of nuclear power units in different operating stages and working conditions. For example, in the unit startup stage, the range of fuzzy sets such as "low" and "high" of vibration amplitude is different from that in the stable operation stage. At the same time, the machine learning algorithm is used to analyze the historical operation data and continuously optimize the division boundary of the fuzzy set to improve the accuracy and real-time performance of the fuzzy evaluation. The weights of the fuzzy rules are also dynamically adjusted, and different weight values ​​are assigned according to the changes in the degree of influence of each parameter on the stability and reliability of the unit under different working conditions. To make the evaluation results more consistent with the actual operation conditions, the key equipment performance indicators of nuclear power units are introduced as new considerations for the formulation of fuzzy rules. For example, the operating status of the reactor coolant pump and the pressure stability of the steam generator are included in the input variables of the fuzzy rules. By establishing a multi-variable fuzzy rule base, combined with expert experience and data analysis, more refined fuzzy rules are determined to comprehensively evaluate the stability and reliability of the unit under different operating conditions. At the same time, the priority of the fuzzy rules is set. When multiple rules are triggered at the same time, the final evaluation result is determined according to the priority to improve the accuracy and consistency of the evaluation. In the fuzzy logic evaluation, an adaptive adjustment mechanism is used for the fuzzy set. The parameter fuzzy set division is dynamically adjusted according to the different operating stages and working conditions of the nuclear power unit. The multi-variable fuzzy rule base, expert experience and data analysis are combined to determine the refined fuzzy rules and priorities, which can more accurately evaluate the stability and reliability of the unit. For example, in the unit startup stage, the fuzzy set division of the vibration amplitude is different from that in the stable operation stage. Through adaptive adjustment, it can better adapt to the characteristics of different stages and improve the accuracy of the evaluation.

[0035] Sp3: Bayesian network evaluation: Using the data collected by the Internet of Things, including sensor data, historical operation data and external environmental data, the nodes of the Bayesian network are constructed, including power output, temperature, pressure, vibration amplitude, and external load demand. According to historical data and expert knowledge, the probability relationship between nodes is determined, and the learning algorithm of the Bayesian network is continuously updated and optimized. When new observation data is input, the Bayesian network is inferred based on the existing probability relationship to predict the probability distribution of unobserved variables. Combined with big data analysis technology, the potential operation modes and laws of nuclear power units are mined, and the Bayesian network structure and parameters are further optimized to improve the evaluation accuracy and reliability. The method for determining the probability relationship between nodes in the Bayesian network evaluation adopts a hybrid learning algorithm. First, the maximum likelihood estimation method is used to make a preliminary estimate of a large amount of historical data to obtain the initial probability relationship between nodes, and then , combined with the Bayesian estimation method, by continuously introducing new observation data, the probability relationship is updated. In the updating process, the Markov Chain Monte Carlo (MCMC) sampling technology is used to improve the calculation efficiency and accuracy. At the same time, in order to improve the robustness of the Bayesian network, a confidence interval is set for the probability relationship between nodes. When the new data deviates greatly from the existing probability relationship, the re-learning mechanism is triggered to adapt to the changes in the operating status of the nuclear power unit. In the Bayesian network evaluation, a hybrid learning algorithm is used to determine the probability relationship between nodes. First, a large amount of historical data is preliminarily estimated using the maximum likelihood estimation to obtain the initial probability relationship. Then, the probability relationship is updated by continuously introducing new observation data in combination with the Bayesian estimation. In this way, historical data and real-time observation data can be fully utilized to explore the potential modes and laws of nuclear power unit operation, further optimize the Bayesian network structure and parameters, and improve the reliability of the evaluation;

[0036] Sp4: Simulated annealing algorithm optimization: The power regulation amplitude and regulation speed parameters of the unit are used as optimization variables, and the objective function is defined, including minimizing power fluctuations and maximizing unit response speed and stability. Starting from a random initial solution, combined with the needs of the smart grid, the needs include changes in grid load and renewable energy supply. According to the current temperature and changes in the objective function, a worse solution is accepted with a certain probability. As the temperature gradually decreases, the probability of accepting a worse solution also gradually decreases. The optimal solution is continuously updated until the stop condition is met. In addition to minimizing power fluctuations and maximizing unit response speed and stability, the definition of the objective function in the simulated annealing algorithm optimization also incorporates the operating cost factor of the nuclear power unit. The operating cost includes fuel consumption, equipment maintenance cost, etc. By establishing a cost model, the cost Factors are incorporated into the objective function so that the optimization results can meet the peak-shaving and frequency-regulating requirements while minimizing the operating cost of the unit. In the execution of the simulated annealing algorithm, cost constraints are set. When the optimized solution exceeds the cost constraint, the solution is rejected with a certain probability to ensure that the optimization result is within the economically feasible range. The simulated annealing algorithm optimization takes the power regulation amplitude and regulation speed parameters of the unit as optimization variables, and defines the objective function to comprehensively consider minimizing power fluctuations, maximizing the unit's response speed and stability, and operating cost factors. By combining the needs of the smart grid, a poor solution is accepted with a certain probability and the acceptance probability is reduced as the temperature decreases. The optimal solution is continuously updated, which can enable nuclear power units to better adapt to changes in grid load and renewable energy supply during peak-shaving and frequency-regulating processes, and improve energy efficiency.

[0037] Sp5: Particle swarm optimization: Each particle in the particle swarm represents a set of possible power regulation parameters. Each particle has a position vector and a velocity vector. The position vector represents the current solution, and the velocity vector determines the direction and speed of the particle movement. The particle is updated according to its own historical optimal position and the group's optimal position. The particle update rule is adjusted in combination with the real-time status of the smart grid. For example, when the grid load suddenly increases, the particles move faster in the direction of increasing power output. The fast convergence of the particle swarm optimization algorithm is used to accelerate the search process and improve the optimization efficiency. At the same time, it is combined with the simulated annealing algorithm to avoid falling into the local optimal solution. The particle update rule in the particle swarm optimization is further improved. In addition to considering its own historical optimal position and the group's optimal position, the guiding role of external excellent solutions is introduced. The external excellent solution can be the optimization of other similar nuclear power units. The solution obtained from experience or the excellent solution obtained by other optimization algorithms is balanced by setting weight coefficients to balance the influence of its own historical optimal position, the group's optimal position and external excellent solutions, so that particles can converge to the global optimal solution faster. At the same time, in order to prevent particles from falling into the local optimal solution, mutation operation is used to randomly perturb the position and speed of particles to increase the search ability of the algorithm. In particle swarm optimization, particles are updated according to their own historical optimal position, the group's optimal position and external excellent solutions, and the particle update rules are adjusted in combination with the real-time status of the smart grid. The rapid convergence of the particle swarm optimization algorithm is used to accelerate the search process and improve the optimization efficiency. At the same time, it is combined with the simulated annealing algorithm to avoid falling into the local optimal solution, further optimize the power regulation parameters, and make the peak and frequency regulation strategy of nuclear power units more efficient;

[0038] Sp6: Smart grid collaboration: Nuclear power units establish real-time communication connections with smart grids and share operating status information of both parties. Nuclear power units understand the load demand of the power grid and the output of renewable energy. Smart grids obtain information such as the power output capacity and response time of nuclear power units, adjust the peak-shaving and frequency-regulating strategies of nuclear power units according to the needs of smart grids, and coordinate and optimize with other energy systems. Through coordinated control of smart grids, complementarity and collaboration between different energy systems are achieved. In smart grid collaboration, a closer two-way feedback mechanism is established. When the peak-shaving and frequency-regulating strategies of nuclear power units are adjusted, the smart grid not only promptly feeds back adjustment suggestions based on the new operating status, but also transmits the real-time demand information of the power grid to the nuclear power units in more detail. For example, the smart grid can provide nuclear power units with information based on load forecasts and the output of renewable energy in different time periods. Provide precise power regulation targets and time requirements. Nuclear power units further optimize peak-shaving and frequency regulation strategies based on this information to achieve more efficient collaborative optimization. At the same time, establish a secure communication protocol between the smart grid and nuclear power units to ensure the reliability and confidentiality of data transmission. Nuclear power units establish real-time communication connections with smart grids, share operating status information, adjust peak-shaving and frequency regulation strategies according to smart grid needs, and conduct collaborative optimization with other energy systems. Through smart grid coordinated control, different energy systems can be complemented and coordinated to improve the stability and reliability of the entire energy system. For example, when the grid load suddenly increases, the smart grid can promptly transmit information to the nuclear power units, and the nuclear power units quickly respond to increase power output. At the same time, the smart grid coordinates other energy systems to meet load demands and improve the reliability of energy supply.

[0039] A comprehensive real-time monitoring and early warning system is set up for the intelligent evaluation and optimization control method. The real-time monitoring system monitors the operating data, evaluation results and optimization control process of the nuclear power unit in real time, and displays the changes in key parameters and indicators through a visual interface. The early warning system sends out early warning signals in time according to preset thresholds and rules when the evaluation results show that the unit's operating status is abnormal or problems occur in the optimization control process. The early warning signals can include sound, light, text messages and other forms, so that operators can take corresponding measures in time. At the same time, a fault diagnosis module is established. When an early warning occurs, the fault is automatically diagnosed and analyzed, and the cause of the fault and solution suggestions are provided.

[0040] Develop a powerful visualization interface that displays the operating data, evaluation results, and optimization and control processes of nuclear power units in the form of intuitive charts, including real-time data curves, historical data trends, visualization of fuzzy evaluation results, the structure and probability distribution of Bayesian networks, etc. Through interactive operations, operators can easily view the relationship between different parameters, adjust the parameters of evaluation and optimization algorithms, and simulate different operating scenarios. At the same time, the visualization interface also provides data analysis and report generation functions to facilitate managers to make decisions and performance evaluations. In addition, the visualization interface should have good user experience and scalability, and be able to adapt to the needs of different users and system upgrades.

[0041] Example 2

[0042] like Figure 1-9 As shown, according to the contents in the above specific embodiments, the following contents are further disclosed:

[0043] Regarding data collection and transmission, the following contents are further disclosed:

[0044] Data fusion: Data fusion technology is used to process the data of multiple sensors to improve the accuracy and reliability of the data. Suppose the data collected by multiple sensors of the same type are {d1, d2, ..., d n}, the weighted average method can be used for data fusion, and the fused data is:

[0045]

[0046] Among them, w i is the weight coefficient of the sensor, satisfying

[0047] Data verification: The data verification mechanism is used to detect abnormal data in a timely manner to ensure that the central evaluation system receives accurate unit operation data. Suppose the collected data is d, and the preset normal range is [d min , d max ], when d min ≤d≤d max When , the data is normal; otherwise, the data is abnormal;

[0048] When collecting and transmitting data, the data collected by multiple sensors of the same type are {d1, d2, ..., d n}, data fusion is performed by weighted average method, and the data d of each sensor i Multiply it by its corresponding weight coefficient w iThen all the products are added together to get the fused data D. The sum of the weight coefficients is 1, ensuring that the fused data is reasonably weighed between the data of each sensor. For example, if the weight of a sensor is large, it means that it contributes more to the final result in data fusion. When verifying the data, the collected data is verified by setting the normal range of the data. If the data is within d min and d max If the data is within this range, it is considered normal; if it is not within this range, it is considered abnormal, which helps to detect erroneous data in time and ensure that the central evaluation system receives accurate unit operation data.

[0049] The following are further disclosed when evaluating fuzzy logic:

[0050] Each parameter has a corresponding fuzzy set, including vibration amplitude, temperature, pressure and other parameters. A set of evaluation definition fuzzy sets is set. The definition parameters and principles are the same. Taking vibration amplitude as an example, let the vibration amplitude be the variable x, and define the fuzzy sets "very low A1", "low A2", "moderate A3", "high A4", "very high A5". Use membership functions to represent each fuzzy set. Common membership functions include triangular membership functions, trapezoidal membership functions, etc. For triangular membership functions, the "low" fuzzy set A2 can be defined as:

[0051]

[0052] Among them, a, b, and c are parameters that determine the range of the fuzzy set. When x is in the interval [a, b], the membership degree 0 increases linearly from 0 to 1; in the interval [b, c], the membership degree 1 decreases linearly from 0 to 0:

[0053] When x is below a or above c, the membership is 0, indicating that it does not belong to the fuzzy set of "low" at all;

[0054] When x is between a and b, the membership increases linearly from 0 to 1, indicating that as increases, the degree of belonging to the fuzzy set of “low” gradually increases;

[0055] When x is between b and c, the membership degree decreases linearly from 1 to 0, indicating that as it continues to increase, the degree of belonging to the fuzzy set of "low" gradually decreases;

[0056] The fuzzy rule is expressed as follows: if the vibration amplitude is A2 and the temperature is B2, the temperature also has a similar fuzzy set definition, and the pressure is C3, then the unit stability is D1. The stability state of the unit is determined by the fuzzy set combination of multiple parameters;

[0057] Fuzzy reasoning: Using the Mamdan i method in fuzzy reasoning, for the input vibration amplitude x0, temperature y0, and pressure z0, first calculate the membership of each condition, set the membership of vibration amplitude x0 to fuzzy set A2 as μA2(x0), the membership of temperature y0 to fuzzy set B2 as μB2(y0), and the membership of pressure z0 to fuzzy set C3 as μC3(z0), and then take the minimum value of these three memberships as the trigger strength α of the rule, which means that the rule will be triggered only when all conditions are met to a certain extent, and the trigger strength depends on the membership of the parameter that least meets the conditions. The trigger strength of the rule is:

[0058]

[0059] For each output fuzzy set, such as the stability fuzzy set D1, fuzzy synthesis is performed according to the trigger strength α, which is expressed by the triangular membership function as follows:

[0060]

[0061] The result after synthesis is:

[0062]

[0063] First, the fuzzy set D1 is defined using the triangular membership function, where d, e, and f are parameters that determine the range of the fuzzy set. Then, the trigger strength α is multiplied by the function value of the fuzzy set D1 to obtain the synthesized membership degree: This indicates the degree to which the stability of the output belongs to the fuzzy set D1, considering the input parameters;

[0064] Finally, the specific stability evaluation value is obtained by defuzzification method (such as centroid method). The formula of centroid method is:

[0065]

[0066] The result of fuzzy synthesis is defuzzified by the centroid method to obtain a specific stability evaluation value. The numerator is the sum of all possible values ​​multiplied by the corresponding membership after synthesis, which represents the weighted sum of stability values. The denominator is the sum of all memberships after synthesis, which represents the weighted sum. By dividing the two, a specific stability value is obtained, which converts the fuzzy evaluation result into a clear numerical representation.

[0067] Regarding the Bayesian network evaluation, the following contents are further disclosed:

[0068] Definition of node probability: Assume that there are n nodes X1, X2, ..., X in the Bayesian network. n , for each node X i , define its conditional probability table:

[0069] P(X i |Parents(X i ))

[0070] Among them, Parents(X i ) represents node X i The parent node set of a node. In a Bayesian network, each node has a conditional probability table, which indicates the probability of the node taking different values ​​given the parent node value of the node. For example, for node X1, if its parent nodes are X2 and X3, then the conditional probability table P(X1|X2, X3) indicates the probability of X1 taking different values ​​under different combinations of X2 and X3 values.

[0071] The conditional probability table of node X1 can be expressed as:

[0072]

[0073] Among them, p 11 , p 12 , ... are conditional probability values, v1, v2, ... and w1, w2, ... are the values ​​of nodes X2 and X3. For example, when X2 takes the value v1 and X3 takes the value w1, the probability that node X1 takes a certain value is p 11 ,Through such conditional probability table, the probability distribution of a node can be calculated based on the known parent node values;

[0074] Bayesian network reasoning: Given the observed values ​​of some nodes, we can use the Bayesian formula to infer the probability distribution of other unobserved nodes. i , according to the Bayesian formula:

[0075]

[0076] Where E is the set of observations, P(E|X i ) is given by X i The probability of the observed value appearing under the condition of taking the value can be calculated through the structure of the Bayesian network and the conditional probability table, P(X i ) is node X i The prior probability of the nodes is P(X i |E) means that when the observation set E is known, node X i The posterior probability, numerator P(E|X i )P(X i ) indicates that at node X i The probability of the observed value appearing when taking a specific value is multiplied by the node X iThe prior probability of all possible X i The values ​​are summed up, that is, all possible situations are considered. In this way, the probability distribution of unobserved nodes can be inferred based on the known observations, so as to better understand the state of the system.

[0077] The simulated annealing algorithm optimization further includes the following:

[0078] Objective function definition: Let the objective function be f(x1, x2, ..., x m ), where x1, x2, …, x m Optimize variables such as power regulation amplitude and regulation speed for nuclear power units;

[0079] The objective function can be to minimize the power fluctuation, that is:

[0080]

[0081] Among them, P t is the power output at time t, is the average power output, T is the time range, and the objective function is used to measure the optimization effect. In the simulated annealing algorithm, the objective function is minimized by adjusting the optimization variables such as power adjustment amplitude and adjustment speed. The objective function here is the sum of the squares of power fluctuations, that is, the power output P at each moment t With average power output The purpose is to make the power output more stable and reduce fluctuations.

[0082] The simulated annealing algorithm steps are as follows:

[0083] Sp1: Initialization: Randomly select an initial solution x0 = (x 01 , x 02 , …, x 0m ), set the initial temperature T0 and the cooling rate T0;

[0084] Sp2: Iteration process: At the current temperature T, from the current solution x = (x1, x2, ..., x m ) Randomly generate a new solution x′=(x′1,x′2,…,x′ m ), calculate the change in the objective function value Δf = f(x′)-f(x), if Δf<0, accept the new solution as the current solution, if Δf>0, then use the probability The probability of accepting a new solution decreases as the temperature decreases, that is, as the algorithm progresses, it becomes less and less likely to accept a worse solution. This process is repeated a certain number of times to fully explore the solution space;

[0085] Sp3: Cooling: Lower the temperature T = αT, and lower the temperature according to the cooling rate. The temperature reduction simulates the temperature drop in the physical annealing process, making the algorithm more inclined to accept better solutions in the later stage, thus improving the quality of the solution;

[0086] Sp4: Repeat the iteration and cooling process until the stopping condition is met, such as the temperature drops to a low enough level or the objective function value no longer changes significantly. The stopping condition can be that the temperature drops to a low enough level, at which point the algorithm is close to convergence; or the objective function value no longer changes significantly, indicating that a better solution has been found.

[0087] For particle swarm optimization, the following are further included:

[0088] Assume that there are N particles in the particle swarm, and the position of each particle i in the m-dimensional space is represented by x i =(x i1 , x i2 , …, x im ), the velocity of the particle is represented by v i =(v i1 , v i2 , …, v im );

[0089] The particle position and velocity update formula is:

[0090] v ij (t+1)=w×v ij (t)+c1×r1×(pbest ij -x ij (t))+c2× r 2×(gbest j -x ij (t))

[0091] x ij (t+1)=x ij (t)+v ij (t+1)

[0092] Where j = 1, 2, ..., m represents the dimension, t represents the number of iterations, w represents the inertia weight, c1 and c2 represent the learning factors, r1 and r2 represent random numbers in the interval [0, 1], and pbest represents the number of iterations. ij is the value of the historical optimal position of particle i in the jth dimension, gbest j is the value of the historical optimal position of the entire particle swarm in the jth dimension;

[0093] Inertia weight w: controls the degree to which the particle maintains its current speed. A larger w is conducive to global search, allowing particles to explore more extensively in the solution space. A smaller w is conducive to local search, allowing particles to focus more on optimizing the current area. Learning factors c1 and c2: respectively adjust the particle to its historical optimal position pbest ij is the historical best position of the group gbest j The close step size, random numbers r1 and r2 increase the randomness of the search and prevent particles from falling into the local optimal solution. The position update formula determines the next position based on the current position and the updated speed. By continuously iteratively updating the position and speed of the particles, the particle swarm gradually approaches a better solution.

[0094] Combined with real-time state adjustment of smart grid: When the load of smart grid changes, parameters such as learning factors c1 and c2, inertia weight w, etc. can be adjusted according to the specific situation, or other items in the speed update formula can be directly adjusted to guide particles to move faster in the direction that adapts to the needs of smart grid. When the real-time state of smart grid load changes changes, the parameters of particle swarm optimization algorithm can be dynamically adjusted according to the specific situation. For example, adjusting the learning factor can change the speed at which particles approach their own optimal and group optimal positions, adjusting the inertia weight can change the search range of particles, or directly adjusting other items in the speed update formula, so that particles can adapt to the needs of smart grid faster and find power regulation parameter solutions that are more suitable for the current grid state. Specific embodiment 3

[0096] like Figure 1-9 As shown, according to the contents in the above specific embodiments, the following contents are further disclosed:

[0097] The operation logic and data processing of the above-mentioned method steps in conjunction with the algorithm for evaluating and optimizing the peak and frequency regulation of nuclear power units are as follows:

[0098] For the data fusion process, the data collected by multiple sensors of the same type are fused through the weighted average method, and different weight coefficients are assigned to different sensors. These weight coefficients reflect the relative importance of each sensor in the data fusion. The data of each sensor is multiplied by the corresponding weight and then added up to obtain the fused data, making the fused data more accurate and reliable. For example, if a sensor performs more stably and accurately under specific circumstances, it can be given a higher weight, so that it occupies a larger proportion in the fused data.

[0099] Data trends and functions: Collect raw data from multiple sensors, and output fused data after weight calculation and accumulation. This fused data will serve as one of the basic data for subsequent evaluation and optimization and regulation, improve the accuracy and reliability of the data, and provide better input data for subsequent fuzzy logic evaluation, Bayesian network evaluation, etc., making the evaluation of the operating status of nuclear power units more accurate.

[0100] For the data verification part, a normal range is set for the collected data, and whether the data is normal is determined by comparing the data with this range. If the data is between the preset minimum and maximum ranges, the data is considered normal; otherwise, the data is judged to be abnormal. For example, for data collected by a temperature sensor, if the preset normal range is [50, 100] degrees Celsius, when the collected temperature value is within this range, the data is considered normal, otherwise the exception handling mechanism is triggered.

[0101] Data trends and functions: Receive the collected data, compare it with the preset normal range, and output the judgment result of whether the data is normal or abnormal, ensuring that the central evaluation system receives accurate unit operation data. Timely detection of abnormal data can trigger alarm signals, data recovery technology or switch to backup sensors to ensure that the system operates and makes decisions based on reliable data.

[0102] The running logic of the fuzzy logic evaluation part is as follows: First, fuzzy sets are defined for parameters such as vibration amplitude, temperature, and pressure, and the scope of each fuzzy set is represented by a membership function. For example, for the "low" fuzzy set of vibration amplitude, the degree of belonging to the fuzzy set under different vibration amplitude values ​​is determined by a triangular membership function. Then, fuzzy rules are formulated to evaluate the stability and reliability of the unit based on the fuzzy set combination of multiple parameters. For example, if the vibration amplitude is "low", the temperature is "moderate", and the pressure is "normal", the unit stability is "good". In fuzzy reasoning, the membership of the input data to each fuzzy set is calculated, and the minimum membership is taken as the trigger strength of the rule. The output fuzzy set is fuzzy synthesized according to the trigger strength, and finally the specific stability evaluation value is obtained through the defuzzification method.

[0103] Data trends and functions: Receive the collected vibration amplitude, temperature, pressure and other data, calculate the data's membership to each fuzzy set through membership functions, perform reasoning and synthesis based on fuzzy rules, output specific stability assessment values, and perform fuzzy assessment on the stability and reliability of the nuclear power unit. This assessment method can handle uncertainty and complexity, and the output assessment results can provide a reference for subsequent optimization and control, helping to determine the operating status and adjustment strategy of the nuclear power unit.

[0104] The operating logic of the Bayesian network evaluation part is as follows: the nodes of the Bayesian network are constructed using the data collected by the Internet of Things, including power output, temperature, pressure, vibration amplitude, external load demand, etc. The conditional probability table between the nodes is determined based on historical data and expert knowledge. When the observed values ​​of some nodes are known, the Bayesian formula is used for reasoning to calculate the probability distribution of other unobserved nodes. The hybrid learning algorithm is used to continuously update the probability relationship between the optimized nodes to adapt to changes in the operating status of the nuclear power unit.

[0105] Data trends and functions: Receive data collected by the Internet of Things, build Bayesian network nodes, determine the probabilistic relationship between nodes, continuously update the probabilistic relationship based on new observation data, output the probability distribution of unobserved nodes, and provide a more comprehensive evaluation of the system by mining the potential operating modes and laws of nuclear power units. Through Bayesian network evaluation, the probability distribution of unobserved variables can be predicted, which helps understand the operating status of nuclear power units under different conditions and provides a basis for optimizing regulation.

[0106] The simulated annealing algorithm optimizes part of the operating logic: the power regulation amplitude, regulation speed and other parameters of the unit are used as optimization variables, and the objective function is defined as minimizing power fluctuations, maximizing unit response speed and stability, and considering adding operating cost factors. Starting from a random initial solution, combined with the needs of the smart grid, according to the current temperature and changes in the objective function, a worse solution is accepted with a certain probability. As the temperature gradually decreases, the probability of accepting a worse solution also gradually decreases. The optimal solution is continuously updated until the stopping condition is met.

[0107] Data trends and functions: Receive relevant parameters of nuclear power units and smart grid demand information, and output optimized power regulation amplitude, regulation speed and other parameters through random initialization, iterative update and solution acceptance judgment. By optimizing the power regulation parameters of nuclear power units, nuclear power units can better adapt to the needs of smart grids during peak and frequency regulation. At the same time, considering the operating costs, it can reduce the operating costs while meeting the peak and frequency regulation needs, and improve energy utilization efficiency.

[0108] Partial operation logic of particle swarm optimization: Each particle in the particle swarm represents a set of possible power regulation parameters. The particle is updated according to its own historical optimal position, the group's optimal position and external excellent solutions. The particle update rules are adjusted in combination with the real-time status of the smart grid. The position and speed of the particle are continuously adjusted through the speed and position update formula. The mutation operation is used to avoid falling into the local optimal solution. It is combined with the simulated annealing algorithm to improve the optimization efficiency.

[0109] Data trends and functions: Receive relevant parameters of nuclear power units and real-time status information of smart grids, output optimized power regulation parameters through particle update and optimization process, further optimize power regulation parameters, make peak and frequency regulation strategies of nuclear power units more efficient, and adjust particle update rules in combination with real-time status of smart grids, so as to better adapt to changes in grid load and improve the response speed and stability of nuclear power units.

[0110] The purpose of the final output data is to provide accurate evaluation and optimization results for the peak and frequency regulation of nuclear power units, so that nuclear power units can better adapt to the needs of smart grids, improve energy utilization efficiency and system stability and reliability. Through real-time monitoring and early warning systems, the output data can promptly detect abnormal situations and take corresponding measures to ensure the safe operation of nuclear power units. The visual interface displays the output data, which is convenient for operators to view system status and parameter relationships, adjust evaluation and optimization algorithm parameters, and provide support for managers to make decisions and performance evaluations.

[0111] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprising a reference structure" do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0112] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units, characterized in that: include: Real-time collection of nuclear power unit operation data; Defining fuzzy sets for the operation data of the nuclear power unit, determining the vibration amplitude level and formulating fuzzy rules, and using a fuzzy inference system to obtain fuzzy evaluation results of the stability and reliability of the nuclear power unit based on the input real-time data and fuzzy rules; Constructing nodes of a Bayesian network according to the operation data of the nuclear power unit, determining the probability relationship between the nodes according to historical data and expert knowledge, and updating and optimizing the probability relationship of the Bayesian network through a learning algorithm and input new observation data; According to the operation data of the nuclear power unit, the optimization variables of the simulated annealing algorithm are set and the objective function is defined. The simulated annealing algorithm optimization starts from a random initial solution, combines the needs of the smart grid, accepts a poor solution according to the current temperature and the objective function, and updates the optimal solution; Particle swarm particles are set according to the operation data of the nuclear power unit, and the particles are updated according to their own historical optimal positions and the optimal positions of the group. At the same time, the particle swarm optimization algorithm is combined with the simulated annealing algorithm to prevent the particles from falling into a local optimal solution; The nuclear power unit establishes a real-time communication connection with the smart grid to share the operating status information of both parties.

2. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 is characterized in that: In the step of collecting nuclear power unit operation data, During data collection and transmission, a multiple redundant sensor layout strategy is adopted. Multiple sensors of the same type are installed in the nuclear power unit. The data of multiple sensors are processed by data fusion technology. Suppose the data collected by multiple sensors of the same type are {d1, d2, …, d n }, the weighted average method is used for data fusion, and the fused data is: Among them, w i is the weight coefficient of the sensor, satisfying Set up a data verification mechanism to perform real-time verification on the collected data. Assume the collected data is d, and the preset normal range is [d min , d max ], when d min ≤d≤d max When , the data is normal; otherwise, the data is abnormal; When data anomalies are detected, an alarm signal is issued and data recovery technology is used or switching to a backup sensor is performed.

3. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 is characterized in that: The fuzzy set definition adopts an adaptive adjustment mechanism, dynamically adjusts the fuzzy set division of parameters according to the characteristic changes of the nuclear power unit in different operation stages and working conditions, determines the fuzzy rules by establishing a multi-variable fuzzy rule base, combines expert experience and data analysis, and sets the priority of the fuzzy rules. When multiple rules are triggered at the same time, the final evaluation result is determined according to the priority. The Mamdani method in fuzzy reasoning is used. For the input nuclear power unit parameters, the membership degree of each parameter condition is first calculated, and then the minimum value of each membership degree is taken as the triggering strength of the rule. When all conditions are met, the rule will be triggered, and the triggering strength depends on the membership degree of the parameter that meets the least conditions. For each output fuzzy set D1, fuzzy synthesis is performed according to the trigger strength α, which is expressed by the triangular membership function as follows: The result after synthesis is: First, the fuzzy set D1 is defined using the triangular membership function, where d, e, and f are parameters that determine the range of the fuzzy set. Then, the trigger strength α is multiplied by the function value of the fuzzy set D1 to obtain the synthesized membership degree: This indicates the degree to which the stability of the output belongs to the fuzzy set D1, considering the input parameters; Finally, the specific stability evaluation value is obtained by defuzzification method. The centroid method formula is: The result of fuzzy synthesis is defuzzified by the centroid method to obtain a specific stability evaluation value. The numerator is the sum of all possible values ​​multiplied by the corresponding membership after synthesis, which represents the weighted sum of stability values. The denominator is the sum of all memberships after synthesis, which represents the weighted sum. By dividing the two, a specific stability value is obtained, which converts the fuzzy evaluation result into a clear numerical representation.

4. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 is characterized in that: The method for determining the probability relationship between nodes adopts a hybrid learning algorithm. First, a large amount of historical data is preliminarily estimated using the maximum likelihood estimation method to obtain the initial probability relationship between the nodes. Then, the probability relationship is updated by continuously introducing new observation data in combination with the Bayesian estimation method. Definition of node probability: Assume that there are n nodes X1, X2, ..., X in the Bayesian network. n , for each node X i , define its conditional probability table: P(X i |Parents(X i )) Among them, Parents(X i ) represents node X i The parent node set of a node. In a Bayesian network, each node has a conditional probability table, which indicates the probability of the node taking different values ​​given the value of the node's parent node. For node X1, if its parent nodes are X2 and X3, then the conditional probability table P(X1|X2, X3) indicates the probability of X1 taking different values ​​under different combinations of X2 and X3 values. The conditional probability table of node X1 is expressed as: Among them, p 11 , p 12 , ... are conditional probability values, v1, v2, ... and w1, w2, ... are the values ​​of nodes X2 and X3. When X2 takes the value v1 and X3 takes the value w1, the probability that node X1 takes a certain value is p 11 ,Through such a conditional probability table, the probability distribution of the node is calculated according to the known parent node value; Bayesian network reasoning: Given the observed values ​​of some nodes, we can use the Bayesian formula to infer the probability distribution of other unobserved nodes. i , according to the Bayesian formula: Where E is the set of observations, P(E|X i ) is given by X i The probability of the observed value appearing under the condition of taking the value is calculated through the structure of the Bayesian network and the conditional probability table, P(X i ) is node X i The prior probability of the nodes is P(X i |E) means that when the observation set E is known, node X i The posterior probability, numerator P(E|X i )P(X i ) indicates that at node X i The probability of the observed value appearing when taking a specific value is multiplied by the node X i The prior probability of all possible X i The values ​​are summed up, that is, all possible situations are considered. In this way, the probability distribution of unobserved nodes is inferred based on the known observed values.

5. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 is characterized in that: The definition of the objective function in the simulated annealing algorithm optimization not only considers minimizing power fluctuations and maximizing unit response speed and stability, but also incorporates the operating cost factor of the nuclear power unit. The operating cost includes fuel consumption and equipment maintenance cost. By establishing a cost model, the cost factor is incorporated into the objective function. During the execution of the simulated annealing algorithm, cost constraints are set. When the optimized solution exceeds the cost constraint, the solution is rejected with a certain probability.

6. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 or 5, characterized in that: In the simulated annealing algorithm optimization, Objective function definition: Let the objective function be f(x1, x2, ..., x m ), where x1, x2, …, x m is the optimization variable of the nuclear power unit; The objective function is to minimize the power fluctuation, that is: Among them, P t is the power output at time t, is the average power output, I is the time range, and the objective function is used to measure the effect of optimization. In the simulated annealing algorithm, the objective function is minimized by adjusting the optimization variables. Here, the objective function is the sum of the squares of power fluctuations, that is, the power output P at each moment. t With average power output Sum the squares of the differences; The simulated annealing algorithm steps are as follows: Sp1: Initialization: Randomly select an initial solution x0 = (x 01 , x 02 , …, x 0m ), set the initial temperature T0 and the cooling rate T0; Sp2: Iteration process: At the current temperature T, from the current solution x = (x1, x2, ..., x m ) Randomly generate a new solution x′=(x′1,x′2,…,x′ m ), calculate the change in the objective function value Δf = f(x′)-f(x), if Δf<0, accept the new solution as the current solution, if Δf>0, then use the probability The probability of accepting a new solution decreases as the temperature decreases, that is, as the algorithm progresses, it becomes less and less likely to accept a worse solution. This process is repeated a certain number of times; Sp3: Cooling: lower the temperature T = αT, and lower the temperature according to the cooling rate; Sp4: Repeat the iteration and cooling process until the stopping condition is met.

7. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 is characterized in that: The updating rule of particles in the particle swarm optimization is further improved. In addition to considering its own historical optimal position and the group's optimal position, the guiding role of external excellent solutions is introduced. The external excellent solutions are solutions obtained from the optimization experience of other similar nuclear power units, or excellent solutions obtained through other optimization algorithms. By setting weight coefficients, the influence of its own historical optimal position, the group's optimal position and external excellent solutions is balanced. At the same time, mutation operations are used to randomly perturb the position and velocity of particles.

8. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 or 7, characterized in that: In the particle swarm optimization, Assume that there are N particles in the particle swarm, and the position of each particle i in the m-dimensional space is represented by x i =(x i1 , x i2 , …, x im ), the velocity of the particle is represented by v i =(v i1 , v i2 , …, v im ); The particle position and velocity update formula is: v ij (t+1)=w×v ij (t)+c1×r1×(pbest ij -x ij (t))+c2×r2×(gbest j -x ij (t)) x ij (t+1)=x ij (t)+v ij (t+1) Where j = 1, 2, ..., m represents the dimension, t represents the number of iterations, w represents the inertia weight, c1 and c2 represent the learning factors, r1 and r2 represent random numbers in the interval [0, 1], and pbest represents the number of iterations. ij is the value of the historical optimal position of particle i in the jth dimension, gbest j is the value of the historical optimal position of the entire particle swarm in the jth dimension; The position update formula determines the next position based on the current position and the updated speed. By continuously iteratively updating the position and speed of the particles, the particle swarm gradually approaches a better solution. Adjust the learning factors c1 and c2, the inertia weight w parameter, or directly adjust other terms in the velocity update formula to guide the particles to move in the direction that adapts to the needs of the smart grid.

9. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 is characterized by: The nuclear power unit and the smart grid establish a two-way feedback mechanism. When the peak-shaving and frequency-regulating strategies of the nuclear power unit are adjusted, the smart grid not only promptly feeds back adjustment suggestions based on the new operating status, but also transmits the real-time demand information of the power grid to the nuclear power unit. The smart grid provides power regulation targets and time requirements to the nuclear power unit based on load forecasts in different time periods and the output of renewable energy. The nuclear power unit further optimizes the peak-shaving and frequency-regulating strategies based on this information.

10. The intelligent evaluation and optimization control method for peak and frequency regulation of nuclear power units according to claim 1 is characterized in that: It also includes real-time monitoring and early warning. Real-time monitoring monitors the operating data, evaluation results and optimization and control processes of nuclear power units in real time, and displays the changes in key parameters and indicators through a visual interface. The early warning system sends out early warning signals based on preset thresholds and rules when the evaluation results show that the operating status of the unit is abnormal or problems occur in the optimization and control process. At the same time, a fault diagnosis module is established. When an early warning occurs, the fault is automatically diagnosed and analyzed, and the cause of the fault and solution suggestions are provided.

Images