CN110780593B - Operation scheme autonomous decision method for unattended small pressurized water reactor - Google Patents
Operation scheme autonomous decision method for unattended small pressurized water reactor Download PDFInfo
- Publication number
- CN110780593B CN110780593B CN201911062132.6A CN201911062132A CN110780593B CN 110780593 B CN110780593 B CN 110780593B CN 201911062132 A CN201911062132 A CN 201911062132A CN 110780593 B CN110780593 B CN 110780593B
- Authority
- CN
- China
- Prior art keywords
- objective function
- set value
- function
- operation scheme
- steam
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses an operation scheme autonomous decision method for an unattended small pressurized water reactor, which comprises the following steps: (1) carrying out standardization processing on the data; (2) determining a feasible domain of the running scheme; (3) establishing an objective function for measuring the quality of the operation scheme; (4) establishing a recursion function, and determining the relationship between the set value of the operation scheme and the input parameter of the target function; (5) and finding the set value of the operation scheme which enables the objective function value to be maximum through a Bayesian optimization algorithm, and taking the operation scheme as the final scheme. The invention can obtain a better optimization result by using the exploration times as few as possible; the method adopts a recursive function to obtain target function input data, the target function input data are key thermal hydraulic parameters, dependence on an actual system and a simulator in the process of exploring different operation schemes is avoided, and time required by each optimization is shortened; the intelligent level is improved.
Description
Technical Field
The invention relates to a decision method, in particular to an operation scheme autonomous decision method for an unattended small pressurized water reactor, which can be suitable for an operation scheme decision method for selecting key parameter set values of a primary loop and a secondary loop.
Background
The nuclear energy generated by the reactor core of the reactor is converted into heat energy under the slowing action of a coolant, the heat is transferred to a water working medium on the secondary side of the direct current steam generator through the coolant, the water is heated into steam, and the steam enters the steam turbine to do work to convert the heat energy into mechanical energy and electric energy. The nuclear safety is a life line of a nuclear power device, and a proper operation scheme must be adopted according to external load change and the operation state of the nuclear power device, and particularly, the stable operation of the small pressurized water reactor in a safe state is ensured under a fault state.
At present, an operator generally makes a decision on an operation scheme of a nuclear power plant, the operator needs to comprehensively consider tens of key parameter operation states through artificial decision, a feasible operation scheme is provided according to the current load requirement, the process is time-consuming and prone to misjudgment, loss which cannot be recovered is caused, even if correct judgment is made, the given operation scheme is not optimal, particularly, in a fault state, the operator is afraid of making the judgment easily, and generally conservative shutdown operation is adopted. For an unattended small pressurized water reactor, firstly, no operator is on duty, namely, an operation scheme of artificial decision cannot be obtained, and secondly, the unattended small pressurized water reactor is not allowed to be shut down easily, and continuous energy output capacity needs to be ensured.
Disclosure of Invention
The invention aims to provide an operation scheme autonomous decision method for an unattended small pressurized water reactor, which comprehensively considers the operation states of key parameters of a primary loop and a secondary loop, establishes an objective function based on safety and economy, obtains an operation scheme with the maximum objective function value through a Bayesian optimization algorithm, and takes the operation scheme as an optimal operation scheme. The method does not need human intervention, avoids the possibility of human error, can obtain an optimal operation scheme, and improves the operation capacity and the vitality of the small pressurized water reactor.
In order to solve the technical problems, the invention adopts the following technical scheme:
an operation scheme autonomous decision method for an unattended small pressurized water reactor comprises the following steps:
(1) carrying out standardization processing on data, eliminating the influence of different dimensions, and processing the set value of the operation scheme and the input data of the objective function into a quantity of [0,1 ];
(2) determining a feasible region of an operation scheme, receiving the current system health state information, and determining a value range of a set value of the operation scheme through a rule base;
(3) establishing an objective function for measuring the quality of an operation scheme, simultaneously considering economy and safety, respectively establishing an economy objective function and a safety objective function, and obtaining an overall objective function through the two objective functions;
(4) Determining the relationship between the set value of the operation scheme and the input parameter of the target function through a recursive function;
(5) and finding the set value of the operation scheme which enables the objective function value to be maximum through a Bayesian optimization algorithm, and taking the operation scheme as the final scheme.
Further, the objective function in step (3) specifically includes:
a. establishing a safety objective function, and considering the safe operation of the reactor, the voltage stabilizer and the steam generator;
b. establishing an economic objective function, and considering the influence of the quality and the yield of steam on the economic efficiency;
c. and establishing an overall objective function, and considering safety and economy.
Further, the recursive function in step (4) specifically includes:
a. acquiring training data, and acquiring steady-state operation data under each operation scheme from an actual system or a simulator;
b. and training a recursive function, fitting the training data by adopting a least square method, wherein dependent variables are the outlet temperature of the coolant, the pressure of a pressurizer, the steam pressure, the superheat degree of steam and the steam flow, and independent variables are the thermal power set value of the reactor core and the average temperature set value of the coolant.
Further, the bayesian optimization algorithm in the step (5) specifically includes:
a. Establishing a Gaussian process kernel function, wherein a set value of an operation scheme to be evaluated at present is called a current observation point, and the operation scheme which is already evaluated is called a historical observation point;
b. establishing a covariance vector formed by a current observation point and n historical observation points;
c. predicting a mean function and a variance function, wherein the mean function and the variance function in the Gaussian process obey Gaussian distribution;
d. the next observation point is obtained: establishing an acquisition function, and determining a next observation point according to the mean value and the variance of posterior distribution prediction in the Gaussian process;
e. repeating the steps a-d until the preset iteration times are finished;
f. and applying the final optimal operation scheme to a practical system: and after the preset iteration times are finished, setting values of an optimal operation scheme obtained by Bayesian optimization are used as setting values of a reactor power controller, a voltage stabilizer pressure controller and a steam generator pressure controller.
Further, the step (1) specifically includes:
a. determining the parameters to be standardized: the method comprises the steps of setting a running scheme and inputting a target function; the set values of the operation scheme comprise a reactor core thermal power set value, a coolant average temperature set value, a pressurizer pressure set value and a steam pressure set value, and the target function input comprises a coolant outlet temperature, pressurizer pressure, steam superheat degree and steam flow;
b. And (3) carrying out standardization processing on data: the normalization process of the above parameters is shown as follows:
in the formula, x * -normalized parameters;
x-parameter before normalization;
x max -normalizing the pre-parameter maximum;
x min -pre-normalization parameter minimum.
Further, the step (2) specifically includes:
a. determining a current system health state: the system health state information is used as an external interface of the invention and is provided by an alarm system of the small pressurized water reactor;
b. and establishing a rule base which consists of a plurality of IF-THEN rules, wherein the condition is system health state information, and the conclusion is a feasible region of the set value of the operation scheme, wherein the set value of the operation scheme comprises a core thermal power set value, a coolant average temperature set value, a pressurizer pressure set value and a steam pressure set value.
Further, the step (3) specifically includes:
a. establishing a security objective function: the safety objective function takes into account the safe operation of the reactor, the manostat, and the steam generator as shown in the following equation:
in the formula, X is the input vector of the objective function，X＝[x 1 ,x 2 ,x 3 ,x 4 ,x 5 ]The temperature of a coolant outlet, the pressure of a pressure stabilizer, the steam pressure, the degree of superheat of steam and the flow rate of the steam are respectively;
ω s,i -coefficients of the ith input parameter;
f s (X) -an economic objective function;
σ(x i ) -a function of Sigmoid, to which,
the above formula is normalized as shown in the following formula:
in the formula (f) s,max (X) -a safety objective function maximum before normalization processing;
f s,min (X) -a safety objective function minimum before normalization processing;
f s * (X) -a standardized post-processing security objective function;
b. establishing an economic objective function: the economic objective function considers the quality and the yield of the steam, the superheat degree of the steam is ensured to be within a proper range, and the larger the flow rate of the steam is, the more the steam is beneficial to the economic efficiency, as shown in the following formula:
in the formula, ω e,i -coefficients of the ith input parameter;
f s (X) -an economic objective function;
the above formula is normalized as shown in the following formula:
in the formula (f) e,max (X) -economic objective function maximum before normalization processing;
f e,min (X) -an economic objective function minimum before normalization processing;
f e * (X) -an economic objective function after normalization processing;
c. establishing an overall objective function: the overall objective function takes into account economy and safety as shown in the following equation:
f(X)＝ω 1 f s * (X)+(1-ω 1 )f e * (X)
in the formula, ω 1 -a weight coefficient;
f (X) -overall objective function.
Further, the step (4) specifically includes:
a. Acquiring training data: acquiring steady-state operation data under each operation scheme from an actual system or a simulator;
b. training a recursive function: the least square method is adopted to fit the training data, the dependent variables are the outlet temperature of the coolant, the pressure of a pressure stabilizer, the steam pressure, the superheat degree of steam and the steam flow, and the independent variables are the thermal power set value of the reactor core and the average temperature set value of the coolant, and the method is based on polynomial fitting and is shown as the following formula:
wherein a, b, c, d, e, f, g, h, i are coefficients;
x 1 ,x 2 -an independent variable;
y is a dependent variable;
c. establishing a relation between a set value of an operation scheme and a target function value: the step (3) establishes the relationship between the objective function input and the objective function value, and the step (4) establishes the relationship between the operation scheme set value and the objective function input, so that the relationship between the operation scheme set value and the objective function value is indirectly established.
Further, the step (5) specifically includes:
a. establishing a Gaussian process kernel function: the set value of the operation scheme to be evaluated at present is called a current observation point, the operation scheme to be evaluated is called a historical observation point, and a ratioloqualative kernel function is adopted, as shown in the following formula:
Wherein, l is the length scale;
α -square scale;
x * -a current observation point;
x j -a historical observation point;
b. establishing a covariance vector: the covariance vector of a current observation point and n historical observation points is shown as follows:
k(x * )＝(k(x * ,x 1 )…k(x * ,x n )) T
c. predicted mean and variance functions: the mean function and variance function of the gaussian process follow a gaussian distribution, and the predicted mean function is given by:
where K — K ═ K (X, X) denotes an n × n matrix of covariances at all historical observation points;
i-unit array;
y is the objective function value corresponding to the current observation point;
the predicted variance function is shown as:
in the formula, k (x) * ,x * ) -representing the variance at the current observation point;
d. the next observation point is obtained: establishing an acquisition function, determining a next observation point according to the mean value and the variance of the posterior distribution prediction in the Gaussian process, wherein the acquisition function is shown as the following formula:
in the formula (f) best -an observed optimum value of the objective function;
Φ (-) a standard normal cumulative distribution function;
phi (-) a standard normal cumulative distribution function;
e [ I (x) ] -the next observation point, the operating scenario set value;
e. Repeating the steps a-d until the preset iteration times are finished;
f. and applying the final optimal operation scheme to a practical system: and after the preset iteration times are finished, setting values of the optimal operation scheme obtained by Bayesian optimization are used as setting values of a reactor power controller, a pressure stabilizer pressure controller and a steam generator pressure controller.
Compared with the prior art, the invention has the beneficial technical effects that:
(1) the optimization algorithm adopted by the invention is an optimization algorithm based on posterior distribution, the Bayesian optimization algorithm can learn and analyze the explored operation scheme and balance the operation scheme explored next time, and compared with other random search algorithms (such as a particle swarm optimization algorithm), the Bayesian optimization algorithm can obtain a better optimization result by using the exploration times as few as possible;
(2) the invention adopts a mode of acquiring target function input data by a recursion function, the target function input data is a key thermotechnical hydraulic parameter, and the key parameter is generally acquired by executing an operation scheme to be explored on an actual system or a simulator;
(3) The method can replace professionals to make operation scheme decisions, automatically generates the optimal operation scheme according to the current system health state, and improves the intelligent level.
Drawings
The invention is further illustrated in the following description with reference to the drawings.
Fig. 1 is a flow chart of the autonomous decision method for the operation scheme of the unattended small pressurized water reactor.
FIG. 2 is a schematic diagram of the power-loss fault of four main pumps of the integrated pressurized water reactor.
Fig. 3 is an iteration result of the bayesian optimization algorithm under the condition of four main pumps having faults.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
As shown in fig. 1, an operation scheme autonomous decision method for an unattended small pressurized water reactor mainly includes the following steps:
(1) the data is standardized, and the specific process is as follows:
a. determining the parameters to be standardized: the method comprises the operation scheme set values (a reactor core thermal power set value, a coolant average temperature set value, a pressurizer pressure set value and a steam pressure set value) and target function input (coolant outlet temperature, pressurizer pressure, steam superheat degree and steam flow).
b. And (3) carrying out standardization processing on data: the normalization process of the above parameters is shown as follows:
In the formula, x * -normalized parameters;
x-parameter before normalization;
x max -normalizing the pre-parameter maximum;
x min -pre-normalization parameter minimum.
(2) Determining a feasible domain of an operation scheme, wherein the specific process comprises the following steps:
a. determining a current system health state: the system health state information is used as an external interface of the invention and is provided by an alarm system of the small pressurized water reactor;
b. and establishing a rule base, wherein the rule base consists of a plurality of IF-THEN rules, the condition (IF) is system health state information, and the conclusion (THEN) is a feasible region of operation scheme set values (a reactor core thermal power set value, a coolant average temperature set value, a pressurizer pressure set value and a steam pressure set value).
(3) Establishing an objective function for measuring the quality of the operating scheme, wherein the specific process comprises the following steps:
a. establishing a security objective function: the safety objective function takes into account the safe operation of the reactor, the manostat, and the steam generator as shown in the following equation:
where X is the input vector of the objective function, X ═ X 1 ,x 2 ,x 3 ,x 4 ,x 5 ]The temperature of a coolant outlet, the pressure of a pressure stabilizer, the steam pressure, the degree of superheat of steam and the flow rate of the steam are respectively;
ω s,i -coefficients of the ith input parameter;
f s (X) -an economic objective function;
σ(x i ) -a function of Sigmoid, to which,
the above formula is normalized as shown in the following formula:
in the formula (f) s,max (X) -a safety objective function maximum before normalization processing;
f s,min (X) -a safety objective function minimum before normalization processing;
f s * (X) -a standardized post-processing security objective function;
b. establishing an economic objective function: the economic objective function considers the quality and the yield of the steam, the superheat degree of the steam is ensured to be in a proper range, and the larger the flow rate of the steam is, the more the economic efficiency is facilitated, as shown in the following formula:
in the formula, ω e,i -coefficients of the ith input parameter;
f s (X) -economic objective function.
The above formula is normalized as shown in the following formula:
in the formula (f) e,max (X) -economic objective function maximum before normalization processing;
f e,min (X) -an economic objective function minimum before normalization processing;
f e * (X) -an economic objective function after normalization processing;
c. establishing an overall objective function: the overall objective function takes into account economy and safety as shown in the following equation:
f(X)＝ω 1 f s * (X)+(1-ω 1 )f e * (X)
in the formula, ω 1 -a weight coefficient;
f (X) -overall objective function.
(4) Determining the relationship between the set value of the operation scheme and the input parameter of the target function through a recursive function, wherein the specific process is as follows:
a. Acquiring training data: acquiring steady-state operation data under each operation scheme from an actual system or a simulator;
b. training a recursive function: the least square method is adopted to fit the training data, the dependent variables are the outlet temperature of the coolant, the pressure of a pressure stabilizer, the steam pressure, the superheat degree of steam and the steam flow, and the independent variables are the thermal power set value of the reactor core and the average temperature set value of the coolant, and the method is based on polynomial fitting and is shown as the following formula:
wherein, a, b, c, d, e, f, g, h, i are coefficients;
x 1 ,x 2 -an independent variable;
y is a dependent variable;
c. establishing a relation between a set value of an operation scheme and a target function value: the step (3) establishes the relationship between the objective function input and the objective function value, and the step (4) establishes the relationship between the operation scheme set value and the objective function input, so that the relationship between the operation scheme set value and the objective function value is indirectly established.
(5) The operation scheme set value which enables the objective function value to be maximum is found through a Bayesian optimization algorithm, and the specific process is as follows:
a. establishing a Gaussian process kernel function: the set value of the operation scheme to be evaluated at present is called a current observation point, the operation scheme to be evaluated is called a historical observation point, and a ratioloqualative kernel function is adopted, as shown in the following formula:
Wherein, l is the length scale;
α -square scale;
x * -a current observation point;
x j -historical observation points.
b. Establishing a covariance vector: the covariance vector of a current observation point and n historical observation points is shown as follows:
k(x * )＝(k(x * ,x 1 )…k(x * ,x n )) T
c. predicted mean and variance functions: the mean function and variance function of the gaussian process follow a gaussian distribution, and the predicted mean function is given by:
where K — K ═ K (X, X) denotes an n × n matrix of covariances at all historical observation points;
i-unit array;
y is the objective function value corresponding to the current observation point;
The predicted variance function is shown as:
in the formula, k (x) * ,x * ) -representing the variance at the current observation point;
d. the next observation point is obtained: establishing an acquisition function, determining a next observation point according to the mean value and the variance of the posterior distribution prediction in the Gaussian process, wherein the acquisition function is shown as the following formula:
in the formula (f) best -an observed optimum value of the objective function;
Φ (-) a standard normal cumulative distribution function;
phi (-) a standard normal cumulative distribution function;
e [ I (x) ] -the next observation point (operating scenario set value);
e. Repeating the steps a-d until the preset iteration times are finished;
f. and applying the final optimal operation scheme to a practical system: and after the preset iteration times are finished, setting values of the optimal operation scheme obtained by Bayesian optimization are used as setting values of a reactor power controller, a pressure stabilizer pressure controller and a steam generator pressure controller.
Referring to fig. 2, the implementation of the autonomous decision method of the pressurized water reactor operation scheme is illustrated. Fig. 2 is a flow chart of an integrated pressurized water reactor, heat generated by a reactor core of the reactor is driven out of the reactor core through coolant flow, under the drive of four main pumps, the coolant flows into a primary side of 12 direct current steam generators to transfer the heat to feed water at a secondary side, the feed water absorbs the heat and then is converted into superheated steam, and the generated superheated steam can be used for seawater desalination or can be converted into kinetic energy through a steam turbine. The decision process of the autonomous decision method is explained by simulating the fault that the four main pumps lose power at the same time. The decision method comprises the following steps:
(1) and carrying out standardization processing on the data. The values of the minimum value and the maximum value in the process of carrying out data standardization on the set values of the operation scheme (the set value of the thermal power of the reactor core, the set value of the average temperature of the coolant, the set value of the pressure stabilizer and the set value of the steam pressure) and the input of the objective function (the temperature of the outlet of the coolant, the pressure of the pressure stabilizer, the steam pressure, the superheat degree of the steam and the flow rate of the steam) are shown in the following table, and input parameters can be standardized according to a standard calculation formula.
| Unit | Minimum before normalization | Maximum before normalization | ||
| Thermal power set value of | ％ | 0 | 100 | |
| Coolant average temperature setpoint | ℃ | 250 | 350 | |
| Pressure setting value of | MPa | 0 | 20 | |
| Steam pressure set | MPa | 0 | 5 | |
| Coolant outlet temperature | ℃ | 250 | 350 | |
| Pressure | MPa | 0 | 20 | |
| | MPa | 0 | 5 | |
| Degree of superheat of | ℃ | 0 | 90 | |
| Flow rate of steam | kg/ |
0 | 100 |
(2) The feasible fields of the running scheme are determined. The normalized data is dimensionless data of [0,1], according to the corresponding rules in the rule base as follows:
IF4 main pumps lose power, and the feasible regions of the THEN core thermal power set value, the coolant average temperature set value, the voltage stabilizer pressure set value and the steam pressure set value are respectively [0.05,0.4], [0.39,0.58], [0.75,0.75], [0.6,0.6] ".
The operating recipe set point is determined according to the above rules.
(3) And establishing an objective function for measuring the quality of the operating scheme. The safety objective function corresponds to coefficients of [0.5,0.3,0.1,0.1], the economic objective function corresponds to coefficients of [0.8,0.2,0,0], and the overall objective function coefficient is 0.6. The objective function can be determined by using the above coefficients in combination with an objective function formula.
(4) And determining the relation between the set value of the operation scheme and the input parameter of the target function through a recursive function. According to the rules given in the rule base, the pressurizer pressure set point and the steam pressure set point are kept constant at 15MPa and 3MPa, respectively, so that the independent variable of the recursion function only needs to take into account the core thermal power set point and the coolant average temperature set point, the dependent variable only needs to take into account the coolant outlet temperature and the steam superheat, and the recursion function [ a, b, c, d, e, f, g ] for the coolant outlet temperature is [ -25941.376607,47.0968956,249.971100, 132.981481, -0.80314413, -678.819110,0.000863368016], and the recursion function [ a, b, c, d, e, f, g ] for the steam superheat is [19518.9859, -176.556030, -190.206583, d2 is 345.018452, 0.62155049, -325.0675524, -0.000679851177 ]. From the coefficients, a recursive function can be determined.
(5) And finding the set value of the operation scheme which enables the value of the objective function to be maximum through a Bayesian optimization algorithm. As shown in fig. 3, each iteration obtains an objective function of the current observation point, determines the observation point of the next iteration according to the current observation point and the historical observation point, and repeats the above steps until the preset number of iterations (in this example, 29 iterations) is completed, and finally obtains an optimal objective function value of 0.6429, where the optimal operation scheme includes a core thermal power set value of 31.13%, a coolant average temperature set value of 286.0 ℃, a pressurizer pressure set value of 15MPa, and a steam pressure set value of 3 MPa.
The above-described embodiments are merely illustrative of the preferred embodiments of the present invention, and do not limit the scope of the present invention, and various modifications and improvements of the technical solutions of the present invention can be made by those skilled in the art without departing from the spirit of the present invention, and the technical solutions of the present invention are within the scope of the present invention defined by the claims.
Claims (3)
1. An operation scheme autonomous decision method for an unattended small pressurized water reactor is characterized by comprising the following steps:
(1) carrying out standardization processing on the data to eliminate the influence of different dimensions;
(2) determining a feasible region of an operation scheme, receiving the current system health state information, and determining a value range of a set value of the operation scheme through a rule base; the set values comprise a reactor core thermal power set value, a coolant average temperature set value, a pressure stabilizer pressure set value and a steam pressure set value;
(3) Establishing an objective function for measuring the quality of an operation scheme, simultaneously considering economy and safety, respectively establishing an economy objective function and a safety objective function, and obtaining an overall objective function through the two objective functions; the step (3) specifically comprises:
a. establishing a security objective function: the safety objective function takes into account the safe operation of the reactor, the manostat, and the steam generator as shown in the following equation:
wherein, X is an objective function input vector, X ═ X1, X2, X3, X4, X5], and is respectively a coolant outlet temperature, a pressurizer pressure, a steam superheat degree and a steam flow;
ω s,i -coefficients of the ith input parameter;
f s (X) -an economic objective function;
σ(x i ) -a function of Sigmoid, to which,
；
the above formula is normalized as shown in the following formula:
in the formula (f) s,max (X) -a safety objective function maximum before normalization processing;
f s,min (X) -a safety objective function minimum before normalization processing;
f s * (X) -a standardized post-processing security objective function;
b. establishing an economic objective function: the economic objective function considers the quality and the yield of the steam, the superheat degree of the steam is ensured to be in a proper range, and the larger the flow rate of the steam is, the more the economic efficiency is facilitated, as shown in the following formula:
In the formula, ω e,i -coefficients of the ith input parameter;
f e (X) -an economic objective function;
the above formula is normalized as shown in the following formula:
in the formula (f) e,max (X) -economic objective function maximum before normalization processing;
f e,min (X) -an economic objective function minimum before normalization processing;
f e * (X) -an economic objective function after normalization processing;
c. establishing an overall objective function: the overall objective function takes into account economy and safety as shown in the following equation:
in the formula, ω 1 -a weight coefficient;
(X) -overall objective function;
(4) determining the relationship between the set value of the operation scheme and the input parameter of the target function through a recursive function; wherein the step (4) specifically comprises:
a. acquiring training data: acquiring steady-state operation data under each operation scheme from an actual system or a simulator;
b. training a recursive function: fitting the training data by using a least square method, wherein the dependent variables are the outlet temperature of the coolant, the pressure of a pressurizer, the steam pressure, the superheat degree of steam and the steam flow, and the independent variables are the set value of the thermal power of the reactor core and the level of the coolant
The set value of the average temperature is based on polynomial fitting and is shown as the following formula:
wherein a, b, c, d, e, f, g, h, i, j are coefficients;
x 1 ,x 2 -an independent variable;
y is a dependent variable;
c. establishing a relation between a set value of an operation scheme and a target function value: step (3) establishes a relationship between the objective function input and the objective function value, and step (4) establishes a relationship between the operation scheme set value and the objective function input, thereby operating
Indirectly establishing the relationship between the scheme set value and the objective function value;
(5) finding out a set value of an operation scheme which enables the objective function value to be maximum through a Bayesian optimization algorithm, and taking the operation scheme as a final scheme; wherein the step (5) specifically comprises:
a. establishing a Gaussian process kernel function: the set value of the operation scheme to be evaluated at present is called a current observation point, the operation scheme to be evaluated is called a historical observation point, and a ratioloqualative kernel function is adopted, as shown in the following formula:
wherein, l is the length scale;
α -square scale;
x * -a current observation point;
x j -a historical observation point;
b. establishing a covariance vector: the covariance vector of a current observation point and n historical observation points is shown as follows:
c. predicted mean and variance functions: the mean function and variance function of the gaussian process follow a gaussian distribution, and the predicted mean function is given by:
Where K — K ═ K (X, X) denotes an n × n matrix of covariances at all historical observation points;
i-unit array;
y is the objective function value corresponding to the current observation point;
the predicted variance function is shown as:
in the formula, k (x) * ,x * ) -representing the variance at the current observation point;
d. the next observation point is obtained: establishing an acquisition function, determining a next observation point according to the mean value and the variance of the posterior distribution prediction in the Gaussian process, wherein the acquisition function is shown as the following formula:
in the formula (f) best -an observed optimum value of the objective function;
phi (-) a standard normal cumulative distribution function;
e [ I (x) ] -the next observation point, the operating scenario set value;
e. repeating the steps a-d until the preset iteration times are finished;
f. and applying the final optimal operation scheme to a practical system: and after the preset iteration times are finished, setting values of the optimal operation scheme obtained by Bayesian optimization are used as setting values of a reactor power controller, a pressure stabilizer pressure controller and a steam generator pressure controller.
2. The unattended small pressurized water reactor oriented operation scheme autonomous decision method according to claim 1, characterized in that the step (1) specifically comprises:
a. Determining the parameters to be standardized: the method comprises the steps of setting a running scheme and inputting a target function; the set values of the operation scheme comprise a reactor core thermal power set value, a coolant average temperature set value, a pressurizer pressure set value and a steam pressure set value, and the target function input comprises a coolant outlet temperature, pressurizer pressure, steam superheat degree and steam flow;
b. and (3) carrying out standardization processing on data: the normalization process of the above parameters is shown as follows:
in the formula, x * -normalized parameters;
x-parameter before normalization;
x max -normalizing the pre-parameter maximum;
x min -pre-normalization parameter minimum.
3. The unattended small pressurized water reactor oriented operation scheme autonomous decision method according to claim 1, characterized in that the step (2) specifically comprises:
a. determining a current system health state: the system health state information is used as an external interface and is provided by an alarm system of the small pressurized water reactor;
b. and establishing a rule base which consists of a plurality of IF-THEN rules, wherein the condition is system health state information, and the conclusion is a feasible region of the set value of the operation scheme, wherein the set value of the operation scheme comprises a core thermal power set value, a coolant average temperature set value, a pressurizer pressure set value and a steam pressure set value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911062132.6A CN110780593B (en) | 2019-11-02 | 2019-11-02 | Operation scheme autonomous decision method for unattended small pressurized water reactor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911062132.6A CN110780593B (en) | 2019-11-02 | 2019-11-02 | Operation scheme autonomous decision method for unattended small pressurized water reactor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110780593A CN110780593A (en) | 2020-02-11 |
| CN110780593B true CN110780593B (en) | 2022-07-29 |
Family
ID=69388577
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911062132.6A Active CN110780593B (en) | 2019-11-02 | 2019-11-02 | Operation scheme autonomous decision method for unattended small pressurized water reactor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110780593B (en) |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9666313B2 (en) * | 2012-04-17 | 2017-05-30 | Bwxt Mpower, Inc. | Small modular reactor refueling sequence |
| CN105759611A (en) * | 2016-02-29 | 2016-07-13 | 华南理工大学 | Pressurized water reactor (PWR) nuclear power plant reactor core power model predictive control method based on genetic algorithm |
| CN106773666B (en) * | 2016-11-11 | 2020-01-10 | 中国电力科学研究院 | Model parameter obtaining method for pressurized water reactor primary circuit system |
| CN206696629U (en) * | 2017-04-01 | 2017-12-01 | 华南理工大学 | A kind of pressurized water reactor core output control device |
| CN108983602B (en) * | 2018-06-26 | 2021-04-27 | 华北电力大学 | Active disturbance rejection control method for fast reactor power and coolant outlet temperature |
| CN109214605A (en) * | 2018-11-12 | 2019-01-15 | 国网山东省电力公司电力科学研究院 | Power-system short-term Load Probability prediction technique, apparatus and system |
| CN109740757A (en) * | 2019-01-25 | 2019-05-10 | 电子科技大学 | A kind of Bayes's optimization method based on sequential Monte Carlo method |
-
2019
- 2019-11-02 CN CN201911062132.6A patent/CN110780593B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110780593A (en) | 2020-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fu et al. | Multiobjective optimal control of FOPID controller for hydraulic turbine governing systems based on reinforced multiobjective Harris Hawks optimization coupling with hybrid strategies | |
| Wilding et al. | The use of multi-objective optimization to improve the design process of nuclear power plant systems | |
| CN110110434B (en) | Initialization method for probability load flow deep neural network calculation | |
| Dong et al. | Multilayer perception based reinforcement learning supervisory control of energy systems with application to a nuclear steam supply system | |
| Modiri-Delshad et al. | Economic dispatch in a microgrid through an iterated-based algorithm | |
| CN108121215A (en) | Process control loops method of evaluating performance and device based on full loop reconstruct emulation | |
| CN106015082B (en) | A kind of optimum design method for the impeller that the core main pump coasting time can be improved | |
| CN111399556A (en) | Control method and control system for deaerator water level and computer storage medium | |
| CN110780593B (en) | Operation scheme autonomous decision method for unattended small pressurized water reactor | |
| CN106777521B (en) | Generator set grid-related parameter optimization method based on double-chain quantum genetic algorithm | |
| Sambariya et al. | Evaluation of interval type-2 fuzzy membership function & robust design of power system stabilizer for smib power system | |
| Abdelfattah et al. | Adaptive Neuro-Fuzzy Self Tuned-PID Controller for Stabilization of Core Power in a Pressurized Water Reactor | |
| CN110556828A (en) | Online safety and stability emergency control method and system adaptive to equipment power flow change | |
| Patel et al. | Application of invasive weed optimization algorithm to optimally design multi-staged PID controller for LFC analysis | |
| CN117117840A (en) | Method and device for constructing deep-tuning-condition thermal power grid-related performance multi-level evaluation system | |
| CN116536705A (en) | PEM (PEM) electrolyzed water control method and system based on model predictive control | |
| JP6538573B2 (en) | Machine learning device, motor control device, motor control system, and machine learning method for learning values of resistance regeneration start voltage and resistance regeneration stop voltage | |
| CN116070741A (en) | Scheduling optimization decision system based on deep reinforcement learning and storage medium thereof | |
| CN112308208B (en) | Transformer fault diagnosis method based on deep learning model | |
| CN115480129A (en) | Method and system for monitoring state of submarine cable and fault recovery method | |
| Wu et al. | Parameter optimization for AP1000 steam generator feedwater control system using particle swarm optimization algorithm | |
| Malik et al. | Fractional order modelling and robust multi-model intelligent controllers' synthesis for ACP1000 nuclear power plant | |
| CN112989507A (en) | Method, device and system for optimizing parameters of speed regulator of water turbine | |
| Liu et al. | Research on Accident Diagnosis Method of Reactor System Based on XGBoost Using Bayesian Optimization | |
| CN108984979B (en) | Design method of ultra-supercritical unit depth peak regulation controller based on combination of multivariate frequency domain method and heuristic search algorithm |
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 |