CN116110630A - Reactor core thermal power measurement method and device in DCS system - Google Patents
Reactor core thermal power measurement method and device in DCS system Download PDFInfo
- Publication number
- CN116110630A CN116110630A CN202310122326.0A CN202310122326A CN116110630A CN 116110630 A CN116110630 A CN 116110630A CN 202310122326 A CN202310122326 A CN 202310122326A CN 116110630 A CN116110630 A CN 116110630A
- Authority
- CN
- China
- Prior art keywords
- feature
- thermal power
- core thermal
- optimal
- reactor core
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G21—NUCLEAR PHYSICS; NUCLEAR ENGINEERING
- G21D—NUCLEAR POWER PLANT
- G21D3/00—Control of nuclear power plant
- G21D3/001—Computer implemented control
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Plasma & Fusion (AREA)
- High Energy & Nuclear Physics (AREA)
- Monitoring And Testing Of Nuclear Reactors (AREA)
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
The application relates to a reactor core thermal power measurement method, a reactor core thermal power measurement device and computer equipment in a DCS system. The method comprises the following steps: acquiring core thermal power data and related characteristic variables of the core thermal power; classifying the related feature variables to construct an original feature vector set; the feature importance ranking is carried out on the feature subsets through an extreme gradient lifting algorithm, and an optimal feature subset sequence is determined through an extreme learning machine model; acquiring the super parameters of the core thermal power calculation through a balance optimizer algorithm; determining the optimal combination of modal components of the reactor core thermal power data through an integrated empirical mode decomposition algorithm; and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model. By adopting the scheme, the measurement accuracy can be improved.
Description
Technical Field
The present application relates to the field of power measurement technology, and in particular, to a method, an apparatus, a computer device, a storage medium, and a computer program product for measuring reactor core thermal power in a DCS (Distributed Control System ) system.
Background
In the actual operation process of the nuclear power station, the core thermal power is difficult to directly measure, and a soft measurement method is generally adopted, namely, the core thermal power is indirectly measured through temperature, pressure, flow and other parameters combined with a thermal balance formula. Therefore, the measurement error is affected by both the process parameter error and the calculation method error, and the process parameter error can be eliminated or reduced by the measurement device with higher precision, but no better solution exists for the error generated by the calculation method at present.
At present, a thermal power measurement method based on a thermal balance mechanism model is mainly adopted, and the method is mainly suitable for steady-state working conditions, and is difficult to adapt to transient measurement of unsteady-state working conditions caused by factors such as power conversion, rod position change, temperature change and the like in the actual operation process. Meanwhile, the scheme of the thermal balance mechanism model relates to measurement of a plurality of process parameters, errors generated in the related parameter measurement links have great influence on measurement results, and the problem that nonlinear accumulated errors have influence on core thermal power measurement is difficult to solve.
Therefore, the measurement accuracy of the thermal power measurement method based on the thermal balance mechanism model is low.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, an apparatus, a computer device, a computer readable storage medium, and a computer program product for measuring reactor core thermal power in a DCS system that can improve measurement accuracy.
In a first aspect, the present application provides a method for measuring reactor core thermal power in a DCS system. The method comprises the following steps:
acquiring core thermal power data and related characteristic variables of the core thermal power;
classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables;
performing feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance ranking;
according to the optimal feature subset sequence, parameter optimization is carried out on the extreme learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained;
performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining a modal component optimal combination of the reactor core thermal power data;
and acquiring a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
In one embodiment, the classifying the related feature variables, and constructing an original feature vector set, where before the original feature vector set includes feature subsets obtained by classifying the related feature variables, further includes:
and preprocessing the reactor core thermal power data and the variable values of the related characteristic variables.
In one embodiment, the feature importance ranking of the feature subsets by the extreme gradient lifting algorithm and determining the optimal feature subset sequence by the extreme learning machine model according to the feature subsets of the importance ranking include:
calculating segmentation gains after feature variable iteration in each feature subset through an extreme gradient lifting algorithm;
according to the segmentation gain, sorting the feature importance of each feature subset;
and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance sequence.
In one embodiment, the determining, by the extreme learning machine model, the optimal feature subset sequence for each feature subset ordered according to importance includes:
calculating fitting errors of feature variables in each feature subset in each iteration through an extreme learning machine model according to the feature subsets with the importance ranking, wherein each feature subset deletes features with the feature importance ranking preset last position in each iteration;
And determining an optimal feature subset sequence according to the fitting error.
In one embodiment, the performing parameter optimization on the extreme learning machine model according to the optimal feature subset sequence through a balance optimizer algorithm, and obtaining the super-parameters of the core thermal power calculation includes:
acquiring an fitness function value of each iteration of the optimal feature subset sequence of the extreme learning machine model through a balance optimizer algorithm;
and determining the super-parameters of the core thermal power calculation according to the fitness function value.
In one embodiment, the determining the optimal combination of modal components of the core thermal power data by integrating an empirical mode decomposition algorithm to perform modal component extraction on the core thermal power data includes:
performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm to obtain a modal component combination mode of the reactor core thermal power data;
obtaining fitting errors of the mode combination modes of the modal components;
and determining the optimal combination of modal components of the reactor core thermal power data according to the fitting error.
In a second aspect, the application also provides a reactor core thermal power measurement device in the DCS system. The device comprises:
The data acquisition module is used for acquiring core thermal power data and related characteristic variables of the core thermal power;
the original feature vector set construction module is used for classifying the related feature variables to construct an original feature vector set, and the original feature vector set comprises feature subsets obtained by classifying the related feature variables;
the optimal feature subset sequence determining module is used for carrying out feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance ranking;
the super-parameter acquisition module is used for carrying out parameter optimization on the extreme learning machine model through a balance optimizer algorithm according to the optimal feature subset sequence to acquire super-parameters of core thermal power calculation;
the modal component optimal combination determining module is used for extracting modal components of the reactor core thermal power data through an integrated empirical mode decomposition algorithm and determining the modal component optimal combination of the reactor core thermal power data;
and the reactor core thermal power calculation module is used for acquiring the parameter value corresponding to the super parameter, and calculating the reactor core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:
acquiring core thermal power data and related characteristic variables of the core thermal power;
classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables;
performing feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance ranking;
according to the optimal feature subset sequence, parameter optimization is carried out on the extreme learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained;
performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining a modal component optimal combination of the reactor core thermal power data;
and acquiring a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
acquiring core thermal power data and related characteristic variables of the core thermal power;
classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables;
performing feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance ranking;
according to the optimal feature subset sequence, parameter optimization is carried out on the extreme learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained;
performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining a modal component optimal combination of the reactor core thermal power data;
and acquiring a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the steps of:
acquiring core thermal power data and related characteristic variables of the core thermal power;
classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables;
performing feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance ranking;
according to the optimal feature subset sequence, parameter optimization is carried out on the extreme learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained;
performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining a modal component optimal combination of the reactor core thermal power data;
and acquiring a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
The method, the device, the computer equipment, the storage medium and the computer program product for measuring the reactor core thermal power in the DCS system are used for acquiring the reactor core thermal power data and the relevant characteristic variables of the reactor core thermal power; classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables; feature importance ranking is carried out on each feature subset through an extreme gradient lifting algorithm, and an optimal feature subset sequence is determined through an extreme learning machine model according to each feature subset of the importance ranking; according to the optimal feature subset sequence, parameter optimization is carried out on the limit learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained; performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining the optimal combination of modal components of the reactor core thermal power data; and obtaining a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination. In the scheme, the optimal feature subset sequence is determined through an extreme gradient lifting algorithm and an extreme learning machine model, then, the super parameters are obtained through a balance optimizer algorithm, and errors generated by non-important parameters on the measurement of the core thermal power and errors caused by the core thermal power due to own multi-frequency components can be reduced through the optimal combination of the super parameters and the modal components, so that the problem of nonlinear error accumulation caused by multi-process parameters in the measurement of the core thermal power is solved, and the measurement accuracy of the core thermal power is improved.
Drawings
FIG. 1 is an environmental diagram of an application of a method for measuring reactor core thermal power in a DCS system in one embodiment;
FIG. 2 is a flow chart of a method of measuring reactor core thermal power in a DCS system in one embodiment;
FIG. 3 is a schematic diagram of a specific flow of a method for measuring reactor core thermal power in a DCS system;
FIG. 4 is a flow chart of a method of measuring reactor core thermal power in a DCS system in another embodiment;
FIG. 5 is a schematic diagram of a backward feature selection flow;
FIG. 6 is a schematic diagram of a balance optimizer optimization algorithm flow;
FIG. 7 is a schematic diagram of a determination flow of optimal modal component reconstruction;
FIG. 8 is a block diagram of a reactor core thermal power measurement device in a DCS system in one embodiment;
fig. 9 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The method for measuring the reactor core thermal power in the DCS system can be applied to an application environment shown in FIG. 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. The terminal 102 obtains a reactor core thermal power calculation request, sends the reactor core thermal power calculation request to the server 104, the server 104 obtains reactor core thermal power calculation request, then obtains reactor core thermal power data and relevant feature variables of the reactor core thermal power, classifies the relevant feature variables, builds an original feature vector set, the original feature vector set comprises feature subsets obtained by classifying the relevant feature variables, performs feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, determines an optimal feature subset sequence according to the feature subsets of the importance ranking through an extreme learning machine model, performs parameter optimization on the extreme learning machine model according to the optimal feature subset sequence through a balance optimizer algorithm, obtains super parameters of reactor core thermal power calculation, performs modal component extraction on reactor core thermal power data through an integrated empirical mode decomposition algorithm, determines a modal component optimal combination of the reactor core thermal power data, obtains parameter values corresponding to the super parameters, and calculates the reactor core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter values, the super parameters and the modal component optimal combination. The server 104 may be implemented as a stand-alone server or a server cluster including a plurality of servers.
In one embodiment, as shown in fig. 2, a method for measuring reactor core thermal power in a DCS system is provided, and the method is applied to the server 104 in fig. 1 for illustration, and includes the following steps:
s100, acquiring core thermal power data and related characteristic variables of the core thermal power.
Wherein the related characteristic variables comprise characteristics of temperature, flow, pressure and the like associated with the thermal power of the core.
Specifically, core thermal power data and relevant characteristic variables such as temperature, flow and pressure are obtained, and the core thermal power data and the relevant characteristic variable data are preprocessed to meet the requirement of model input data, such as outlier correction, missing value filling, dimension normalization and the like.
S200, classifying the related feature variables to construct an original feature vector set.
Specifically, the various factors that affect the core thermal power measurement are classified into three major categories: time variable characteristics, dependent variable characteristics, independent variable characteristics. Taking month, day, hour and minute as 5 basic time characteristics in the time independent variables; in the dependent variable characteristics, 8 variables of the acquired inlet temperature, outlet temperature, inlet pressure, outlet pressure, inlet flow, outlet flow, pipeline length and pipeline caliber are taken as process parameter characteristics; in the independent variable characteristics, 8 variables of the front 1s power, the front 15s power, the front 30s power, the front 45s power, the front 60s power, the daily highest power, the daily lowest power and the daily average power are taken as historical core thermal power characteristics, and the expression of a characteristic vector set and the core thermal power is as follows:
P core (t)＝f(T iv (t),X iv (t),P iv (t))
Wherein P is core (t) is the core thermal power real-time value; t (T) iv (t) is a time argument comprising: the method comprises the steps of collecting 5 variables of month (m), day (d), time (h), minute (min) and second(s) corresponding to data; x is X iv (t) is an influencing factor argument comprising: the collected inlet temperature, outlet temperature, inlet pressure, outlet pressure, inlet flow, outlet flow, pipeline length and pipeline caliber are 8 variables in total; p (P) iv (t) is a historical core thermal power argument comprising: the power of the first 1s, the power of the first 15s, the power of the first 30s, the power of the first 45s, the power of the first 60s, the power of the highest day, the power of the lowest day and the average power of the day are 8 variables.
S300, feature importance ranking is carried out on each feature subset through an extreme gradient lifting algorithm, and an optimal feature subset sequence is determined through an extreme learning machine model according to each feature subset of the importance ranking.
Wherein, each feature subset is obtained according to the classification of related feature variables, and comprises a time variable feature subset, a dependent variable feature subset, an independent variable feature subset and the like; the optimal feature subset sequence is the optimal feature sequence in the feature subset obtained by the extreme learning machine model through each feature subset of importance ranking.
Specifically, the extreme gradient lifting (eXtreme Gradient Boosting, XGBoost) algorithm is composed of a plurality of classification regression tree-based learners, and the kernel of the XGBoost algorithm for processing classification or regression problems is that a tree is added for each iteration, so as to correct the deviation between the predicted value and the actual value in the previous iteration, and the newly generated base learner is utilized to continue the iteration.
And calculating the segmentation gain of the feature variable in each feature subset after each iteration through an extreme gradient lifting algorithm, obtaining the weight coefficient of the corresponding variable, and sequencing the feature importance of each feature subset according to the size of the weight coefficient of each variable as a feature sequencing basis. The XGBoost algorithm uses a sum of k functions to make output predictions, calculated as follows:
wherein: x is the argument, k is the number of trees, and F is a feature in feature space F.
Is based on all trees f k (x i ) A predicted estimate of the result of (a).
The complexity function omega and the error function term L together form a loss function Obj, i.e. an objective function, of the algorithm, which is used to measure the degree of fit of the model to the training data.
Obj(θ)＝L(θ)+Ω(θ)
The training process of the model is equivalent to finding the optimal parameter θ to arrive at the training data x i And test data y i The optimal mapping between the two is:
wherein y is i Is a true value of the code,
is a predicted value. In the process of carrying out t-step iteration on the algorithm, the predicted value obtained in each step is expressed as:
the loss function expression of the model is:
where Obj is the loss function; i j All leaf nodes; f is a feature in feature space F; t is the number of leaf nodes, w j The leaf node weight; 
The first and second derivatives of the regional samples respectively; y is i Is a true value, < >>
Is a fitting value.
The input data of the extreme learning machine model is a feature subset importance ranking result obtained through an extreme gradient lifting algorithm, one feature at the last of feature importance ranking is deleted in each iteration, fitting errors of each iteration are calculated, and each feature subset corresponding to the minimum fitting error is the optimal feature subset sequence.
And S400, carrying out parameter optimization on the limit learning machine model through a balance optimizer algorithm according to the optimal feature subset sequence to obtain the super-parameters of the core thermal power calculation.
The parameter optimization is that the optimal feature subset sequence obtained through importance sorting is a local optimal parameter, and a global optimal parameter is needed to be found by using an optimization algorithm; the super-parameters are global optimal parameters obtained through a balance optimizer algorithm to obtain optimal parameters of a thermal power measurement method.
Specifically, the balance optimizer algorithm is iterated for a plurality of times, the current optimal parameter is obtained after each iteration, then the fitness function value of the current balance optimizer model is calculated, and if the current fitness function value is stable at a smaller value, the current optimal parameter is the global optimal parameter, namely the super parameter of the core thermal power calculation.
S500, performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining the optimal combination of modal components of the reactor core thermal power data.
The optimal modal component combination is the optimal modal component combination selected according to the prediction result of the modal component reconstruction after all modal component combinations are obtained through the modal component reconstruction.
Specifically, an integrated empirical mode decomposition algorithm firstly adds a group of 0.15 times standard deviation white noise to an original reactor core thermal power data sequence to obtain a total body, and decomposes each original reactor core thermal power sequence added with noise to obtain an intrinsic mode component function of each reactor core thermal power sequence; after different white noises are added to the reactor core thermal power sequence, reactor core thermal power modal components under each group of noises are respectively obtained, the influence of the white noises added before is eliminated by adopting a weighted average mode on a plurality of obtained reactor core thermal power intrinsic modal component functions, and the finally obtained reactor core thermal power intrinsic modal component functions have the following expression:
wherein: c j (t) is an intrinsic mode component function, m is the decomposition times of an empirical mode decomposition algorithm, c ij (t) the jth modal component resulting from the ith introduction of white noise.
The end condition of the integrated empirical mode decomposition algorithm is that intrinsic mode components cannot be continuously separated from the original reactor core thermal power, and the corresponding residual sequence expression is as follows:
wherein:
the IMF is the eigenmode component and n is the eigenmode component number for the original core thermal power.
And obtaining a mode of mode component combination through mode component reconstruction by an integrated empirical mode decomposition algorithm, calculating fitting errors of all mode component combinations, and determining the mode component optimal combination of the reactor core thermal power data.
S600, obtaining parameter values corresponding to the super parameters, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter values, the super parameters and the modal components.
Specifically, the parameter value corresponding to the super parameter after pretreatment is obtained, the optimal combination of the parameter value and the modal component is used as the input of a thermal power measurement method based on a thermal balance combination model, and the thermal power of the reactor core is calculated. The specific flow of the whole reactor core thermal power measurement method is shown in fig. 3, and a thermal power measurement method module based on a thermal balance combination model is integrated into a DCS system.
The evaluation index of the thermal power measurement method comprises the following steps: the optimal Feature selection number (feature_n), the optimal modal component reconstruction number (mimf_n), and the Runtime (run time) of the model. The optimal feature selection quantity is the quantity of feature variables corresponding to the smallest model fitting error, the optimal modal component reconstruction quantity is the quantity of components corresponding to the smallest model component reconstruction fitting error, and the running time is the total time of the model after training and fitting measurement.
In the method for measuring the thermal power of the reactor core in the DCS system, the thermal power data of the reactor core and the related characteristic variables of the thermal power of the reactor core are obtained, the related characteristic variables are classified, an original characteristic vector set is constructed, the original characteristic vector set comprises characteristic subsets obtained by classifying the related characteristic variables, the characteristic subsets are subjected to characteristic importance ranking through an extreme gradient lifting algorithm, the characteristic subsets ranked according to importance are determined through an extreme learning machine model, an optimal characteristic subset sequence is determined through an extreme learning machine model, the parameter optimization is carried out on the extreme learning machine model according to the optimal characteristic subset sequence through a balance optimizer algorithm, the super-parameters of the thermal power calculation of the reactor core are obtained, the modal component of the thermal power data of the reactor core is extracted through an integrated empirical modal decomposition algorithm, the optimal combination of the modal components of the thermal power data of the reactor core is determined, the parameter values corresponding to the super-parameters are obtained, the thermal power is calculated according to the parameter values, the super-parameters and the optimal combination of the modal components, the thermal power measurement method based on the thermal power of the thermal balance combination model is adopted, the problem that nonlinear errors are accumulated due to the multi-process parameters in the thermal power measurement of the reactor core can be solved, and the measurement accuracy of the thermal power of the reactor core is improved.
In one embodiment, classifying the related feature variables, before constructing the original feature vector set, further includes:
and preprocessing the variable values of the reactor core thermal power data and related characteristic variables.
The preprocessing is to fill and correct missing values and abnormal values of data.
Specifically, the missing value processing can be padded by using corresponding data several weeks before and after the missing data. The data usually has certain similarity on similar days of different periods by taking one week time as one period, so that missing data filling can be performed by solving an average value of similar day data of a plurality of weeks before and after.
The abnormal value processing can be transversely compared and repaired through the characteristic of the continuity of the heat power of the reactor core, namely, the data collected at adjacent moments are used as references, a maximum variation range is set, and when the collected data are out of the set range, namely:
the data is smoothed by adopting the core thermal power average value at the front and rear moments, namely:
wherein:
is a threshold value; y (d, t) is the core thermal power at time t; y (d, t+1) is the core thermal power at time t+1; y (d, t-1) is the core thermal power at time t-1.
The original data have different dimensions and units, normalization processing is usually needed to reduce the characteristic difference, and the normalization formula is as follows:
Wherein: y is max ，y min Respectively the maximum value and the minimum value in the original data,
is normalized data.
In this embodiment, the reliability of the core thermal power data and the variable values of the relevant characteristic variables can be improved by preprocessing the core thermal power data and the variable values of the relevant characteristic variables.
In one embodiment, as shown in fig. 4, S300 includes:
s320, calculating the segmentation gain after feature variable iteration in each feature subset through an extreme gradient lifting algorithm.
And S340, sorting the feature importance of each feature subset according to the segmentation gain.
S360, determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets ordered by importance.
When the node information gain of the tree is calculated, the gain calculated by the segmentation point is segmented to obtain the weight coefficient of each characteristic variable.
Specifically, the extreme gradient lifting adopts a greedy method to divide and calculate the gain of the decision tree, the gain of each division is required to be maximized as much as possible, and the expression of the division gain which is expected to be maximized is as follows:
wherein: g L ，H L Respectively representing the sum of the first and second derivatives of the left subtree of the current node, G R ，H R Representing the first and second derivative sums, respectively, of the right subtree of the current node.
The weight coefficient of the corresponding feature variable can be obtained by calculating the segmentation gain after each feature variable iteration, the feature importance ranking is carried out on each feature subset according to the size of the weight coefficient of each feature variable as a feature ranking basis, and the optimal feature subset sequence is determined through the extreme learning machine model.
In this embodiment, the segmentation gain after feature variable iteration in each feature subset is calculated by using an extreme gradient lifting algorithm, feature importance ranking is performed on each feature subset according to the segmentation gain, and an optimal feature subset sequence is determined by using an extreme learning machine model according to each feature subset of the importance ranking, so that parameter optimization can be performed on each optimal feature subset sequence in the following steps.
In one embodiment, determining the optimal feature subset sequence by the extreme learning machine model based on the importance ranked feature subsets comprises:
calculating fitting errors of feature variables in each feature subset in each iteration through an extreme learning machine model according to the feature subsets of the importance sequence, and deleting features of the feature importance sequence preset last position in each iteration of each feature subset; and determining the optimal feature subset sequence according to the fitting error.
The fitting error refers to an error between the model predicted value and the actual value.
Specifically, the input-output mapping expression of the extreme learning machine model is:
h i (x)＝g(w i ,b i ,x)＝g(w i x+b i )
wherein w is i ＝[w i1 ,w i1 ,...,w in ] T For the weight vectors of the input layer and the hidden layer, beta i ＝[β i1 ,β i2 ,...,β im ] T B is the weight vector of the hidden layer and the output layer i Is a node threshold; h is a i (x) Represents the output of the i-th hidden layer node, g (w i ,b i X) is a sigmoid excitation function.
The input data of the extreme learning machine model is a backward feature selection result of an extreme gradient lifting model vector sequence, a backward feature selection flow is shown in fig. 5, each feature subset iteratively deletes one feature with the last feature of feature importance ranking, and an expression of an iterative process is as follows:
…
where k, k-1, where k-n is the number of features, F is the feature in feature space F,
is according to all f k (x i ) Fitting values obtained from the results of (a),>
to according to all f k-1 (x i-1 ) Fitting values obtained from the results of (a) and so on, < +.>
To according to all f k-n (x i-n ) Fitting values are obtained from the results of (a).
The fitting error evaluation of each iteration adopts root mean square error, the feature set with the minimum fitting error is selected by each feature subset to be the optimal feature subset sequence, and the calculation formula of the fitting error is as follows:
in the method, in the process of the invention,
to fit value, y i N is the number of samples for the actual value.
In this embodiment, the fitting error of each iteration of the feature variable in each feature subset is calculated through the extreme learning machine model according to each feature subset ordered according to the importance, each feature subset deletes the feature of the preset last position of the feature importance ranking each iteration, and the optimal feature subset sequence is determined according to the fitting error, so that the optimal feature in each feature subset can be obtained.
In one embodiment, performing parameter optimization on the limit learning machine model through a balance optimizer algorithm according to the optimal feature subset sequence, and obtaining the super-parameters of the core thermal power calculation includes:
acquiring an adaptability function value of each iteration of the optimal feature subset sequence of the extreme learning machine model through a balance optimizer algorithm; determining the super-parameters of the core thermal power calculation according to the fitness function value;
specifically, as shown in fig. 6, for the optimization problem, the balance optimizer algorithm generally performs multiple iterations, and the current optimization solution obtained after each iteration can be expressed as:
wherein C is eq Representing the optimal solution currently found by the algorithm, C n-1 Representing the optimal solution obtained in the last iteration, C n Representing the newly generated optimization solution; f represents turnover rate, and the calculation formula is as follows:
F＝a 1 sign(r-0.5)(e -λt -1)
wherein the parameters lambda and r are [0,1 ]]Random numbers of (a); a, a 1 For the constant coefficient of global search, the interval is set as [1,2]Sign isThe sign function, t, represents a decreasing parameter.
And calculating an fitness function value corresponding to the current optimal solution, obtaining a current optimal parameter according to the current optimal solution, and obtaining an optimal parameter corresponding to the fitness function value when the fitness function value is stabilized at a smaller value, namely a global optimal parameter, and obtaining a super parameter of the reactor core thermal power calculation according to the global optimal parameter. The fitness function expression adopted by the balance optimizer model is as follows:
wherein, fit is the fitness function,
and->
Fitting values and test values of normalization processing are respectively adopted, and n is the number of samples.
In this embodiment, the fitness function value of each iteration of the optimal feature subset sequence of the extreme learning machine model is obtained through the balance optimizer algorithm, and the super parameter of the core thermal power calculation is determined according to the fitness function value, so that the error of the non-important parameter on the core thermal power measurement can be reduced through the super parameter.
In one embodiment, the determining the optimal combination of modal components of the core thermal power data by integrating an empirical mode decomposition algorithm includes:
Carrying out modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm to obtain a modal component combination mode of the reactor core thermal power data; obtaining fitting errors of the mode combination modes of the modal components; and determining the optimal combination of modal components of the core thermal power data according to the fitting error.
Specifically, as shown in fig. 7, the optimal modal component reconstruction determining flow is shown in fig. 7, the integrated empirical mode decomposition algorithm modal component reconstruction mode is to perform n times of decomposition on the original electrical load data, treat all the obtained modal components from the modal component of the serial number 1 to the modal component of the serial number n as high-frequency modal components, treat residual components as low-frequency components, and reconstruct the result expression as follows:
P core ＝RMSE min (Mimf_n)
wherein P is core As core thermal power, mimf_n is a processing method of superposing modal components of sequence numbers 1 to n as high-frequency modal components and residual components as low-frequency components.
And obtaining all mode component combination modes through mode component reconstruction, calculating fitting errors of all mode component combination modes, and combining the mode components with the smallest fitting errors into a mode component optimal combination.
In this embodiment, by integrating an empirical mode decomposition algorithm, mode component extraction is performed on the core thermal power data, a mode component combination mode of the core thermal power data is obtained, fitting errors of all mode component combination modes are obtained, and according to the fitting errors, an optimal mode component combination of the core thermal power data is determined, so that errors of the core thermal power due to own multifrequency components can be reduced, and measurement accuracy of the core thermal power is improved.
It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides a thermal power measuring device for realizing the method for measuring the thermal power of the reactor core in the DCS system. The implementation of the solution provided by the device is similar to that described in the above method, so the specific limitation in the embodiments of the thermal power measurement device or devices provided below may be referred to the limitation of the method for measuring thermal power of the reactor core in the DCS system hereinabove, and will not be described herein.
In one embodiment, as shown in FIG. 8, there is provided a reactor core thermal power measurement apparatus in a DCS system, comprising: the system comprises a data acquisition module 100, an original feature vector set construction module 200, an optimal feature subset sequence determination module 300, a super-parameter acquisition module 400, a modal component optimal combination determination module 500 and a core thermal power calculation module 600, wherein:
the data acquisition module 100 is used for acquiring core thermal power data and related characteristic variables of the core thermal power;
the original feature vector set construction module 200 is configured to classify related feature variables to construct an original feature vector set, where the original feature vector set includes feature subsets obtained by classifying the related feature variables;
the optimal feature subset sequence determining module 300 is configured to rank the feature importance of each feature subset by using an extreme gradient lifting algorithm, and determine an optimal feature subset sequence by using an extreme learning machine model according to each feature subset ranked by the importance;
the super-parameter obtaining module 400 is configured to perform parameter optimization on the limit learning machine model through a balance optimizer algorithm according to the optimal feature subset sequence, so as to obtain a super-parameter of core thermal power calculation;
The modal component optimal combination determining module 500 is configured to perform modal component extraction on the core thermal power data by integrating an empirical modal decomposition algorithm, and determine a modal component optimal combination of the core thermal power data;
the reactor core thermal power calculation module 600 is configured to obtain a parameter value corresponding to the super parameter, and calculate the reactor core thermal power by using a thermal power measurement method based on a thermal balance combination model according to an optimal combination of the parameter value, the super parameter and the modal component.
In one embodiment, the optimal feature subset sequence determining module 300 is further configured to calculate, by using an extreme gradient lifting algorithm, a segmentation gain after feature variable iteration in each feature subset; according to the segmentation gain, sorting the feature importance of each feature subset; and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance sequence.
In one embodiment, the optimal feature subset sequence determining module 300 is further configured to calculate, through the extreme learning model, a fitting error of each iteration of feature variables in each feature subset according to each feature subset of the importance ranking, and delete, for each iteration, a feature of a feature importance ranking preset end position of each feature subset; and determining the optimal feature subset sequence according to the fitting error.
In one embodiment, the super-parameter obtaining module 400 is further configured to obtain, by using a balance optimizer algorithm, an fitness function value of each iteration of the optimal feature subset sequence of the extreme learning machine model; and determining the super-parameters of the core thermal power calculation according to the fitness function value.
In one embodiment, the modal component optimal combination determination module 500 is further configured to perform modal component extraction on the core thermal power data by integrating an empirical modal decomposition algorithm, so as to obtain a modal component combination mode of the core thermal power data; obtaining fitting errors of the mode combination modes of the modal components; and determining the optimal combination of modal components of the core thermal power data according to the fitting error.
All or part of each module in the reactor core thermal power measuring device in the DCS system can be realized by software, hardware and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 9. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer equipment is used for storing the reactor core thermal power data and the related characteristic variable data of the reactor core thermal power. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by the processor, implements a method for measuring reactor core thermal power in a DCS system.
It will be appreciated by those skilled in the art that the structure shown in fig. 9 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the computer device to which the present application applies, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:
acquiring core thermal power data and related characteristic variables of the core thermal power; classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables; feature importance ranking is carried out on each feature subset through an extreme gradient lifting algorithm, and an optimal feature subset sequence is determined through an extreme learning machine model according to each feature subset of the importance ranking; according to the optimal feature subset sequence, parameter optimization is carried out on the limit learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained; performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining the optimal combination of modal components of the reactor core thermal power data; and obtaining a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
In one embodiment, the processor when executing the computer program further performs the steps of:
and preprocessing the variable values of the reactor core thermal power data and related characteristic variables.
In one embodiment, the processor when executing the computer program further performs the steps of:
calculating segmentation gains after feature variable iteration in each feature subset through an extreme gradient lifting algorithm; according to the segmentation gain, sorting the feature importance of each feature subset; and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance sequence.
In one embodiment, the processor when executing the computer program further performs the steps of:
calculating fitting errors of feature variables in each feature subset in each iteration through an extreme learning machine model according to the feature subsets of the importance sequence, and deleting features of the feature importance sequence preset last position in each iteration of each feature subset; and determining the optimal feature subset sequence according to the fitting error.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring an adaptability function value of each iteration of the optimal feature subset sequence of the extreme learning machine model through a balance optimizer algorithm; and determining the super-parameters of the core thermal power calculation according to the fitness function value.
In one embodiment, the processor when executing the computer program further performs the steps of:
carrying out modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm to obtain a modal component combination mode of the reactor core thermal power data; obtaining fitting errors of the mode combination modes of the modal components; and determining the optimal combination of modal components of the core thermal power data according to the fitting error.
In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring core thermal power data and related characteristic variables of the core thermal power; classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables; feature importance ranking is carried out on each feature subset through an extreme gradient lifting algorithm, and an optimal feature subset sequence is determined through an extreme learning machine model according to each feature subset of the importance ranking; according to the optimal feature subset sequence, parameter optimization is carried out on the limit learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained; performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining the optimal combination of modal components of the reactor core thermal power data; and obtaining a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
In one embodiment, the computer program when executed by the processor further performs the steps of:
and preprocessing the variable values of the reactor core thermal power data and related characteristic variables.
In one embodiment, the computer program when executed by the processor further performs the steps of:
calculating segmentation gains after feature variable iteration in each feature subset through an extreme gradient lifting algorithm; according to the segmentation gain, sorting the feature importance of each feature subset; and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance sequence.
In one embodiment, the computer program when executed by the processor further performs the steps of:
calculating fitting errors of feature variables in each feature subset in each iteration through an extreme learning machine model according to the feature subsets of the importance sequence, and deleting features of the feature importance sequence preset last position in each iteration of each feature subset; and determining the optimal feature subset sequence according to the fitting error.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring an adaptability function value of each iteration of the optimal feature subset sequence of the extreme learning machine model through a balance optimizer algorithm; and determining the super-parameters of the core thermal power calculation according to the fitness function value.
In one embodiment, the computer program when executed by the processor further performs the steps of:
carrying out modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm to obtain a modal component combination mode of the reactor core thermal power data; obtaining fitting errors of the mode combination modes of the modal components; and determining the optimal combination of modal components of the core thermal power data according to the fitting error.
In one embodiment, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:
acquiring core thermal power data and related characteristic variables of the core thermal power; classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables; feature importance ranking is carried out on each feature subset through an extreme gradient lifting algorithm, and an optimal feature subset sequence is determined through an extreme learning machine model according to each feature subset of the importance ranking; parameter optimization is carried out on the limit learning machine model through a balance optimizer algorithm, and the super-parameters of the core thermal power calculation are obtained; performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining the optimal combination of modal components of the reactor core thermal power data; and obtaining a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
In one embodiment, the computer program when executed by the processor further performs the steps of:
and preprocessing the variable values of the reactor core thermal power data and related characteristic variables.
In one embodiment, the computer program when executed by the processor further performs the steps of:
calculating segmentation gains after feature variable iteration in each feature subset through an extreme gradient lifting algorithm; according to the segmentation gain, sorting the feature importance of each feature subset; and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance sequence.
In one embodiment, the computer program when executed by the processor further performs the steps of:
calculating fitting errors of feature variables in each feature subset in each iteration through an extreme learning machine model according to the feature subsets of the importance sequence, and deleting features of the feature importance sequence preset last position in each iteration of each feature subset; and determining the optimal feature subset sequence according to the fitting error.
In one embodiment, the computer program when executed by the processor further performs the steps of:
acquiring an adaptability function value of each iteration of the optimal feature subset sequence of the extreme learning machine model through a balance optimizer algorithm; and determining the super-parameters of the core thermal power calculation according to the fitness function value.
In one embodiment, the computer program when executed by the processor further performs the steps of:
carrying out modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm to obtain a modal component combination mode of the reactor core thermal power data; obtaining fitting errors of the mode combination modes of the modal components; and determining the optimal combination of modal components of the core thermal power data according to the fitting error.
It should be noted that, user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the various embodiments provided herein may include at least one of relational databases and non-relational databases. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.
Claims (10)
1. A method for measuring reactor core thermal power in a DCS system, the method comprising:
acquiring core thermal power data and related characteristic variables of the core thermal power;
classifying the related feature variables to construct an original feature vector set, wherein the original feature vector set comprises feature subsets obtained by classifying the related feature variables;
Performing feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance ranking;
according to the optimal feature subset sequence, parameter optimization is carried out on the extreme learning machine model through a balance optimizer algorithm, and the super-parameters of core thermal power calculation are obtained;
performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm, and determining a modal component optimal combination of the reactor core thermal power data;
and acquiring a parameter value corresponding to the super parameter, and calculating the core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
2. The method of claim 1, wherein classifying the related feature variables to construct an original feature vector set, wherein before the original feature vector set includes feature subsets obtained by classifying the related feature variables, the method further comprises:
and preprocessing the reactor core thermal power data and the variable values of the related characteristic variables.
3. The method of claim 1, wherein ranking the feature subsets for feature importance by an extreme gradient lifting algorithm and determining an optimal feature subset sequence by an extreme learning machine model based on the feature subsets ranked for importance comprises:
calculating segmentation gains after feature variable iteration in each feature subset through an extreme gradient lifting algorithm;
according to the segmentation gain, sorting the feature importance of each feature subset;
and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance sequence.
4. The method of claim 3, wherein determining the optimal feature subset sequence by an extreme learning machine model for each feature subset ordered according to importance comprises:
calculating fitting errors of feature variables in each feature subset in each iteration through an extreme learning machine model according to the feature subsets with the importance ranking, wherein each feature subset deletes features with the feature importance ranking preset last position in each iteration;
and determining an optimal feature subset sequence according to the fitting error.
5. The method of claim 1, wherein the performing parameter optimization on the extreme learning machine model according to the optimal feature subset sequence by a balance optimizer algorithm, and obtaining the super-parameters of the core thermal power calculation comprises:
Acquiring an fitness function value of each iteration of the optimal feature subset sequence of the extreme learning machine model through a balance optimizer algorithm;
and determining the super-parameters of the core thermal power calculation according to the fitness function value.
6. The method of claim 1, wherein the determining the optimal combination of modal components of the core thermal power data by modal component extraction of the core thermal power data by an integrated empirical modal decomposition algorithm comprises:
performing modal component extraction on the reactor core thermal power data through an integrated empirical mode decomposition algorithm to obtain a modal component combination mode of the reactor core thermal power data;
obtaining fitting errors of the mode combination modes of the modal components;
and determining the optimal combination of modal components of the reactor core thermal power data according to the fitting error.
7. A reactor core thermal power measurement device in a DCS system, the device comprising:
the data acquisition module is used for acquiring core thermal power data and related characteristic variables of the core thermal power;
the original feature vector set construction module is used for classifying the related feature variables to construct an original feature vector set, and the original feature vector set comprises feature subsets obtained by classifying the related feature variables;
The optimal feature subset sequence determining module is used for carrying out feature importance ranking on the feature subsets through an extreme gradient lifting algorithm, and determining an optimal feature subset sequence through an extreme learning machine model according to the feature subsets of the importance ranking;
the super-parameter acquisition module is used for carrying out parameter optimization on the extreme learning machine model through a balance optimizer algorithm according to the optimal feature subset sequence to acquire super-parameters of core thermal power calculation;
the modal component optimal combination determining module is used for extracting modal components of the reactor core thermal power data through an integrated empirical mode decomposition algorithm and determining the modal component optimal combination of the reactor core thermal power data;
and the reactor core thermal power calculation module is used for acquiring the parameter value corresponding to the super parameter, and calculating the reactor core thermal power by adopting a thermal power measurement method based on a thermal balance combination model according to the parameter value, the super parameter and the modal component optimal combination.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310122326.0A CN116110630A (en) | 2023-02-07 | 2023-02-07 | Reactor core thermal power measurement method and device in DCS system |
| PCT/CN2023/090025 WO2024164434A1 (en) | 2023-02-07 | 2023-04-23 | Method and apparatus for measuring thermal power of nuclear reactor core in dcs |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310122326.0A CN116110630A (en) | 2023-02-07 | 2023-02-07 | Reactor core thermal power measurement method and device in DCS system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116110630A true CN116110630A (en) | 2023-05-12 |
Family
ID=86254009
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310122326.0A Pending CN116110630A (en) | 2023-02-07 | 2023-02-07 | Reactor core thermal power measurement method and device in DCS system |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN116110630A (en) |
| WO (1) | WO2024164434A1 (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024164434A1 (en) * | 2023-02-07 | 2024-08-15 | 中广核研究院有限公司 | Method and apparatus for measuring thermal power of nuclear reactor core in dcs |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4297581A (en) * | 1979-01-15 | 1981-10-27 | The Babcock & Wilcox Company | Method for the fast and accurate identification of core power in nuclear reactors |
| CN1118831C (en) * | 2000-06-21 | 2003-08-20 | 北京和利时系统工程股份有限公司 | Method for measuring power of pressurized-water reactor core for nuclear power station |
| CN103617816B (en) * | 2013-10-29 | 2016-06-08 | 中国广核集团有限公司 | The measuring method of reactor core power distribution |
| CN112259271B (en) * | 2020-09-28 | 2023-09-12 | 台山核电合营有限公司 | Reactor core thermal power calculation method and device for nuclear power station DCS |
| CN114611808A (en) * | 2022-03-18 | 2022-06-10 | 河海大学 | Short-term offshore wind power prediction method based on CEEMDAN-SSA-BilSTM |
| CN116110630A (en) * | 2023-02-07 | 2023-05-12 | 中广核研究院有限公司 | Reactor core thermal power measurement method and device in DCS system |
-
2023
- 2023-02-07 CN CN202310122326.0A patent/CN116110630A/en active Pending
- 2023-04-23 WO PCT/CN2023/090025 patent/WO2024164434A1/en unknown
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024164434A1 (en) * | 2023-02-07 | 2024-08-15 | 中广核研究院有限公司 | Method and apparatus for measuring thermal power of nuclear reactor core in dcs |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2024164434A1 (en) | 2024-08-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA3088899C (en) | Systems and methods for preparing data for use by machine learning algorithms | |
| US11650968B2 (en) | Systems and methods for predictive early stopping in neural network training | |
| CN110427654B (en) | Landslide prediction model construction method and system based on sensitive state | |
| Abrahamsen et al. | Machine learning in python for weather forecast based on freely available weather data | |
| CN113505534B (en) | Load prediction method considering demand response | |
| CN112925909B (en) | Graph convolution literature classification method and system considering local invariance constraint | |
| CN114490065A (en) | Load prediction method, device and equipment | |
| CN111178986A (en) | User-commodity preference prediction method and system | |
| CN116110630A (en) | Reactor core thermal power measurement method and device in DCS system | |
| CN115965160A (en) | Data center energy consumption prediction method and device, storage medium and electronic equipment | |
| CN113901679B (en) | Reliability analysis method and device for power system and computer equipment | |
| US20220076121A1 (en) | Method and apparatus with neural architecture search based on hardware performance | |
| CN117913796A (en) | Power economic coupling coordination relation determining method, device, equipment and storage medium | |
| CN117521511A (en) | Granary temperature prediction method based on improved wolf algorithm for optimizing LSTM | |
| CN117152528A (en) | Insulator state recognition method, insulator state recognition device, insulator state recognition apparatus, insulator state recognition program, and insulator state recognition program | |
| CN116706888A (en) | Intelligent power grid load prediction method based on federal learning | |
| CN114784795A (en) | Wind power prediction method and device, electronic equipment and storage medium | |
| CN114549233A (en) | Floating population prediction method based on combination of LGB algorithm and ARIMA algorithm | |
| CN114021619A (en) | Classification method and device based on graph structure evolution | |
| CN113111948A (en) | Time sequence data classification method and system based on feature recalibration mechanism | |
| CN117892166B (en) | Sparse data identification method, intelligent computing subsystem and intelligent computing platform | |
| CN115986746B (en) | Prediction method and device based on soft sensor, computer equipment and storage medium | |
| CN113673591B (en) | Self-adjusting sampling optimization image classification method, device and medium | |
| CN118779752A (en) | Health state assessment method, apparatus, device, storage medium, and program product | |
| Sun | Incomplete Big Data Filling Algorithm based on Machine Learning |
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 |