CN117057258B - Black-start overvoltage prediction method and system based on weight distribution correlation coefficient - Google Patents
Black-start overvoltage prediction method and system based on weight distribution correlation coefficient Download PDFInfo
- Publication number
- CN117057258B CN117057258B CN202311318800.3A CN202311318800A CN117057258B CN 117057258 B CN117057258 B CN 117057258B CN 202311318800 A CN202311318800 A CN 202311318800A CN 117057258 B CN117057258 B CN 117057258B
- Authority
- CN
- China
- Prior art keywords
- overvoltage
- correlation coefficient
- input data
- obtaining
- neural network
- 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
- 238000000034 method Methods 0.000 title claims abstract description 110
- 241001672018 Cercomela melanura Species 0.000 title claims abstract description 94
- 238000003062 neural network model Methods 0.000 claims abstract description 77
- 241000283153 Cetacea Species 0.000 claims abstract description 56
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 39
- 125000004122 cyclic group Chemical group 0.000 claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 25
- 210000002569 neuron Anatomy 0.000 claims abstract description 24
- 230000006870 function Effects 0.000 claims description 29
- 239000013598 vector Substances 0.000 claims description 14
- 238000013528 artificial neural network Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000004891 communication Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 7
- 230000000875 corresponding effect Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 2
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000002596 correlated effect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241000288105 Grus Species 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000011524 similarity measure Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention belongs to the technical field of black start overvoltage prediction, and particularly provides a method and a system for predicting black start overvoltage based on a weight distribution correlation coefficient, wherein the method is used for acquiring closing overvoltage and various voltage influence parameters of an idle circuit in a black start process of a power grid, and calculating a first correlation coefficient set and a second correlation coefficient set by using a pearson correlation coefficient and cosine similarity; obtaining an overvoltage predicted value set by using a cyclic neural network model, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set and the two correlation coefficient sets by using a weight distribution correlation coefficient method; obtaining an optimal weight and an optimal threshold of each neuron in the neural network model by adopting an improved whale algorithm so as to obtain a target neural network model; and obtaining target input data based on the various voltage influence parameters and the target correlation coefficient set to input the target neural network model and output a switching-on overvoltage target predicted value. The method improves the accuracy of the black start overvoltage prediction.
Description
Technical Field
The disclosure belongs to the technical field of black-start overvoltage prediction, and particularly relates to a method and a system for predicting black-start overvoltage based on weight distribution correlation coefficients.
Background
The black start is self-rescue and recovery after the whole system is stopped, the power system is in a very weak state in the whole process, and any minor deviation and disturbance can possibly cause failure of the black start. Therefore, the formulated black start scheme needs to be implemented by a strict auditing and simulation verification party. The verification of the black-start overvoltage (i.e. the no-load line closing overvoltage) is an important component of the verification of the black-start scheme.
In the prior art, although an artificial intelligent method is proposed to rapidly predict the no-load closing overvoltage. However, the factors influencing the black start overvoltage are numerous, the influence of factors such as line parameters, a breaker closing initial phase angle and the like on the overvoltage is not considered in the prior art, the prediction result is inaccurate, and the influence degree of different factors on the overvoltage is not considered, so that the prediction result is influenced.
Disclosure of Invention
The present disclosure aims to solve, at least to some extent, one of the technical problems in the related art. Therefore, the disclosure provides a method and a system for predicting black-start overvoltage based on a weight distribution correlation coefficient, and the main purpose is to improve the accuracy of black-start overvoltage prediction.
According to a first aspect of the present disclosure, there is provided a black start overvoltage prediction method based on a weight distribution correlation coefficient, including:
acquiring a closing overvoltage of an idle circuit and various voltage influence parameters in a black start process of a power grid, wherein the various voltage influence parameters comprise a closing circuit length, a compensation value of a circuit shunt reactor, a power supply resistance, a power supply leakage reactance, a circuit resistance per kilometer, a positive sequence reactance per kilometer and a closing initial phase angle;
calculating a first correlation coefficient set of the switching-on overvoltage and the plurality of voltage influence parameters by using a pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and the multiple voltage influence parameters by using cosine similarity;
obtaining model input data based on the various voltage influence parameters, the first correlation coefficient set and the second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method;
constructing a neural network model, taking initial weights and initial thresholds of neurons of the neural network model as position vectors of whales, taking an error function as an fitness function, adopting an improved whale algorithm to obtain optimal weights and optimal thresholds of the neurons, and obtaining a target neural network model based on the optimal weights and the optimal thresholds;
And obtaining target input data based on the various voltage influence parameters and the target correlation coefficient set, and inputting the target input data into the target neural network model to output a switching-on overvoltage target predicted value.
In the black start overvoltage prediction method based on the weight distribution correlation coefficient provided in the first aspect of the present disclosure, the convergence factor of the whale algorithm is optimized by using the initial value and the final value of the convergence factor, the nonlinear equalization coefficient and the maximum iteration number, so as to obtain the improved whale algorithm.
In the method for predicting black-start overvoltage based on the weight distribution correlation coefficient provided in the first aspect of the present disclosure, the neural network model adopts a BP neural network.
In the black-start overvoltage prediction method based on the weight distribution correlation coefficient provided in the first aspect of the present disclosure, the model input data includes first model input data and second model input data, and the overvoltage prediction value set includes a first overvoltage prediction value set and a second overvoltage prediction value set; the obtaining model input data based on the plurality of voltage influence parameters, the first correlation coefficient set and the second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, includes: obtaining first model input data based on the multiple voltage influence parameters and the first correlation coefficient set, and inputting the first model input data into a cyclic neural network model to obtain a first overvoltage predicted value set; and obtaining second model input data based on the multiple voltage influence parameters and the second correlation coefficient set, and inputting the second model input data into a cyclic neural network model to obtain a second overvoltage predicted value set.
In the method for predicting a black start overvoltage based on a weight distribution correlation coefficient provided in the first aspect of the present disclosure, the obtaining, by using a weight distribution correlation coefficient method, a target correlation coefficient set based on the closing overvoltage, the overvoltage predicted value set, the first correlation coefficient set, and the second correlation coefficient set includes: obtaining a first error based on the closing overvoltage and the first overvoltage predicted value set; obtaining a second error based on the closing overvoltage and the second overvoltage predicted value set; and obtaining a target correlation coefficient set based on the first error, the second error, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method.
According to a second aspect of the present disclosure, there is also provided a black-start overvoltage prediction system based on a weight distribution correlation coefficient, including:
the acquisition module is used for acquiring the switching-on overvoltage of the idle circuit and various voltage influence parameters in the black start process of the power grid, wherein the various voltage influence parameters comprise the length of the switching-on circuit, the compensation value of a parallel reactor of the circuit, the power supply resistance, the leakage reactance of the power supply, the resistance of the circuit every kilometer, the positive sequence reactance of the circuit every kilometer and the switching-on initial phase angle;
The coefficient calculation module is used for calculating a first correlation coefficient set of the switching-on overvoltage and the various voltage influence parameters by using the Pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and the multiple voltage influence parameters by using cosine similarity;
the weight distribution module is used for obtaining model input data based on the various voltage influence parameters, the first correlation coefficient set and the second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method;
the modeling module is used for constructing a neural network model, taking the initial weight and the initial threshold value of each neuron of the neural network model as a position vector of a whale, taking an error function as an fitness function, adopting an improved whale algorithm to obtain the optimal weight and the optimal threshold value of each neuron, and obtaining a target neural network model based on the optimal weight and the optimal threshold value;
and the prediction module is used for obtaining target input data based on the various voltage influence parameters and the target correlation coefficient set, inputting the target input data into the target neural network model and outputting a closing overvoltage target predicted value.
In the black start overvoltage prediction system based on the weight distribution correlation coefficient provided in the second aspect of the present disclosure, in the modeling module, the convergence factor of the whale algorithm is optimized by using the initial value and the final value of the convergence factor, the nonlinear equalization coefficient and the maximum iteration number, so as to obtain the improved whale algorithm.
In the black-start overvoltage prediction system based on the weight distribution correlation coefficient provided in the second aspect of the present disclosure, the model input data includes first model input data and second model input data, the overvoltage prediction value set includes a first overvoltage prediction value set and a second overvoltage prediction value set, and the weight distribution module is specifically configured to: obtaining first model input data based on the multiple voltage influence parameters and the first correlation coefficient set, and inputting the first model input data into a cyclic neural network model to obtain a first overvoltage predicted value set; and obtaining second model input data based on the multiple voltage influence parameters and the second correlation coefficient set, and inputting the second model input data into a cyclic neural network model to obtain a second overvoltage predicted value set.
In the black-start overvoltage prediction system based on the weight distribution correlation coefficient provided in the second aspect of the present disclosure, the weight distribution module is specifically configured to: obtaining a first error based on the closing overvoltage and the first overvoltage predicted value set; obtaining a second error based on the closing overvoltage and the second overvoltage predicted value set; and obtaining a target correlation coefficient set based on the first error, the second error, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method.
According to a third aspect of the present disclosure, there is also provided a black-start overvoltage predicting device based on a weight distribution correlation coefficient, including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform a black-start overvoltage prediction method based on a weight distribution correlation coefficient set forth in an embodiment of the first aspect of the present disclosure.
In one or more aspects of the disclosure, acquiring a closing overvoltage of an idle line and various voltage influence parameters in a black start process of a power grid, wherein the various voltage influence parameters comprise a closing line length, a line shunt reactor compensation value, a power supply resistance, a power supply leakage reactance, a line resistance per kilometer, a line positive sequence reactance per kilometer and a closing initial phase angle; calculating a first correlation coefficient set of a switching-on overvoltage and various voltage influence parameters by using the Pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and various voltage influence parameters by using cosine similarity; obtaining model input data based on various voltage influence parameters, a first correlation coefficient set and a second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method; constructing a neural network model, taking initial weights and initial thresholds of all neurons of the neural network model as position vectors of whales, taking an error function as an fitness function, adopting an improved whale algorithm to obtain optimal weights and optimal thresholds of all neurons, and obtaining a target neural network model based on the optimal weights and the optimal thresholds; and obtaining target input data based on various voltage influence parameters and target correlation coefficient sets, and inputting the target input data into a target neural network model to output a switching-on overvoltage target predicted value. Under the condition, the influence of the length of a closing line, the compensation value of a line shunt reactor, the power supply resistance, the leakage reactance of a power supply, the resistance of each kilometer of the line, the positive sequence reactance of each kilometer of the line and the closing initial phase angle on the closing overvoltage is comprehensively considered, the influence degree of different factors on the overvoltage is fully considered by combining the pearson correlation coefficient, the cosine similarity and the weight distribution correlation coefficient method, the accuracy of the black-start overvoltage prediction is improved, and in addition, the accuracy of a neural network model is improved by adopting an improved whale algorithm, and the accuracy of the black-start overvoltage prediction is further improved.
Additional aspects and advantages of the disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the disclosure.
Drawings
The foregoing and/or additional aspects and advantages of the present disclosure will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:
fig. 1 shows a flowchart of a black-start overvoltage prediction method based on a weight distribution correlation coefficient according to an embodiment of the present disclosure;
fig. 2 is a flowchart illustrating another method for predicting black-start overvoltage based on a weight distribution correlation coefficient according to an embodiment of the present disclosure;
FIG. 3 illustrates a block diagram of a black-start overvoltage prediction system based on weight distribution correlation coefficients provided by an embodiment of the present disclosure;
fig. 4 is a block diagram of a weight distribution correlation coefficient-based black-start overvoltage prediction device used to implement a weight distribution correlation coefficient-based black-start overvoltage prediction method of an embodiment of the present disclosure.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the embodiments of the present disclosure. Rather, they are merely examples of apparatus and methods consistent with aspects of embodiments of the present disclosure as detailed in the accompanying claims.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present disclosure. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.
Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In the description of the present disclosure, the meaning of "a plurality" is at least two, such as two, three, etc., unless explicitly specified otherwise. It should also be understood that the term "and/or" as used in this disclosure refers to and encompasses any or all possible combinations of one or more of the associated listed items.
Embodiments of the present disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and intended for the purpose of explaining the present disclosure and are not to be construed as limiting the present disclosure.
The disclosure provides a black start overvoltage prediction method and a system based on weight distribution correlation coefficients, and the method and the system are mainly used for improving the accuracy of black start overvoltage prediction.
In a first embodiment, fig. 1 is a schematic flow chart of a black-start overvoltage prediction method based on a weight distribution correlation coefficient according to an embodiment of the disclosure. Fig. 2 is a flowchart illustrating another method for predicting black-start overvoltage based on a weight distribution correlation coefficient according to an embodiment of the present disclosure. As shown in fig. 1, the method for predicting the black start overvoltage based on the weight distribution correlation coefficient comprises the following steps:
and S11, acquiring the switching-on overvoltage of an idle circuit and various voltage influence parameters in the black start process of the power grid, wherein the various voltage influence parameters comprise the length of the switching-on circuit, the compensation value of a parallel reactor of the circuit, the power supply resistance, the leakage reactance of the power supply, the resistance of the circuit per kilometer, the positive sequence reactance of the circuit per kilometer and the switching-on initial phase angle.
In step S11, the number of the closing overvoltages of the idle line in the acquired grid black start process may be n. All the acquired closing overvoltages may be referred to as a closing overvoltage set. The switching overvoltage set consisting of n switching overvoltages can be represented by a symbolYAnd (3) representing.
In step S11, the various voltage influencing parameters include 7 voltage influencing parameters including a closing line length, a line shunt reactor compensation value, a power supply resistance, a power supply leakage reactance, a line per kilometer resistance (i.e., a line per km resistance), a line per kilometer positive sequence reactance (i.e., a line per km positive sequence reactance), and a closing initial phase angle.
In step S11, the plurality of voltage influencing parameters are the plurality of voltage influencing parameters corresponding to the closing overvoltage. For example, 7 voltage influence parameters under each closing overvoltage are acquired when the closing overvoltage is acquired. The n closing overvoltages correspond to 7n voltage influencing parameters, wherein 7n voltage influencing parameters comprise 7 voltage influencing parameters, each voltage influencing parameter comprising n such voltage influencing parameters.
In step S11, various voltage influencing parameters may be referred to as influencing parameter raw data sets. For example, 7n influence parameter raw data sets corresponding to voltage influence parameters X(which may be simply referred to as raw data) is an n x 7 matrix. The influence parameter raw data set can be expressed asWherein, the method comprises the steps of, wherein,X 1 for n closing line lengths,X 2 for n lines parallel reactanceThe value of the compensation of the device,X 3 for the n power resistors,X 4 is the leakage reactance of n power supplies,X 5 for a resistance of n lines per kilometer,X 6 positive sequence reactance for n lines per kilometer,X 7 n switching-on initial phase angles.
Step S12, calculating a first correlation coefficient set of a switching-on overvoltage and various voltage influence parameters by using the Pearson correlation coefficient; and calculating a second phase relation number set of the switching-on overvoltage and various voltage influence parameters by using the cosine similarity.
As can be easily understood, the pearson correlation coefficient can measure the wireless correlation and the degree of correlation between 2 features, so the pearson correlation coefficient is used in step S12 to measure the correlation between the black start overvoltage (i.e. the closing overvoltage) and other features (i.e. various voltage influencing parameters), and the pearson correlation coefficient satisfies:
wherein,indicating the correlation strength of the ith voltage influence parameter and the closing overvoltage, +.>The positive value indicates that the i-th voltage influence parameter is positively correlated with the closing overvoltage, and +.>A negative value indicates a negative correlation between the i-th voltage-affecting parameter and the closing overvoltage. i=1, 2, …, 7, i.e. the first set of correlation coefficients +. >Is->. First set of correlation coefficients->Can be seen as a 7 x 1 matrix. />The sum of n values representing the i-th voltage influencing parameter in the raw data set of influencing parameters,/->Representing the sum of n closing overvoltages. E () is covariance.
As can be readily appreciated, cosine similarity measures the degree and degree of wireless correlation between 2 features by calculating the cosine value of the included angle between two vectors. Therefore, in step S12, the correlation between the black start overvoltage and other features is measured by using the cosine similarity, and the cosine similarity satisfies:
in the method, in the process of the invention,the i-th voltage influence parameter is related to the overvoltage of the switch-on voltage, i=1, 2, … and 7, namely a second correlation coefficient set +.>Is->Second set of correlation coefficients->Can be seen as a 7 x 1 matrix. Wherein the method comprises the steps ofRepresenting a cosine function.
And S13, obtaining model input data based on various voltage influence parameters, a first correlation coefficient set and a second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method.
In step S13, considering that the pearson correlation coefficient has a significant disadvantage in that its degree of approach to 1 is correlated with the sample capacity, when the sample capacity is small, the variable is determined only by the correlation coefficient being largeAnd->The error of close linear relation is larger. Cosine similarity is insensitive to the absolute magnitude of a particular value. Therefore, a new correlation coefficient is formed by combining the cyclic neural network model and the weight distribution correlation coefficient method, so that the problem of correlation between the characteristic value (namely various voltage influence parameters) and the closing overvoltage is more accurately mined, and the accuracy of subsequent prediction is further improved.
Specifically, in step S13, the recurrent neural network model may be a GRU (gated recurrent unit ) network model. As will be readily appreciated, the expression for the GRU network model is as follows:
wherein,is thatsReset gate of time->Is in an internal state->For resetting the first parameter in the door, < +.>To reset the second parameter in the door,h s-1 for a relative to the current time (i.esTime of day) the hidden state at the previous time of day (i.es-1 timeOutput of the score), ->The input is GRU at the current moment; />Is thatsUpdate door of time, - >For updating the first parameter in the gate, < >>For updating the second parameter in the gate, < >>To output candidate values after the reset gate processing (i.esOutput candidate value for time of day),to find the output candidate value +.>First parameter in the process,/->To find the output candidate value +.>A second parameter in the process; operator "/)>"means that the array elements are multiplied sequentially. />Is thatsHidden state of time (i.e. output of GRU at present time), ->Is thatsCandidate hidden layers at the moment. In addition, the reset gate and the update gate are a simple neural network, in order to fix the gate output between 0 and 1, the neural network is used for the control of the output of the gateThe activation function via the network employs a sigmoid function.
In step S13, the model input data includes first and second model input data, and the set of overvoltage predictors includes first and second sets of overvoltage predictors. Wherein the first model input data may be usedX’The first set of overvoltage predictors may be represented byY’The second model input data can be represented byX”The second set of overvoltage predictors may be represented byY”And (3) representing.
In step S13, obtaining model input data based on the plurality of voltage influencing parameters, the first correlation coefficient set, and the second correlation coefficient set, and inputting the model input data into the recurrent neural network model to obtain an overvoltage prediction value set, including: acquiring first model input data based on various voltage influence parameters and a first correlation coefficient set, and inputting the first model input data into a cyclic neural network model to acquire a first overvoltage predicted value set; and obtaining second model input data based on the multiple voltage influence parameters and the second correlation data set, and inputting the second model input data into the cyclic neural network model to obtain a second overvoltage predicted value set. Specifically, the first model inputs data X’Satisfy the following requirementsThe second model inputs dataX”Satisfy->Wherein, the method comprises the steps of, wherein,Xrepresenting the influence parameter raw dataset,/->For the first set of correlation coefficients->Is the second set of correlation coefficients. Will beX’AndX”respectively substituting into GRU network model to make prediction to obtain first overvoltage prediction value setY’And a second set of overvoltage predictorsY”(see FIG. 2).
In step S13, a weight distribution correlation coefficient method is adopted to obtain a target correlation coefficient set based on the closing overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set, including: obtaining a first error based on the closing overvoltage and the first overvoltage predicted value set; obtaining a second error based on the closing overvoltage and a second overvoltage predicted value set; and obtaining a target correlation coefficient set based on the first error, the second error, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method.
Specifically, the first error and the second error may be calculated using the average relative error. Wherein the average relative errorThe method meets the following conditions:
wherein,is the actual value, which is the closing overvoltage setYI-th value of>As a predicted value, when->First set of overvoltage predictorsY’The first error +. >When->Second set of overvoltage predictorsY”The second error +.>(see FIG. 2), n is the total number of samples, i.e., the closing overvoltage setYThe number of switching-on overvoltages.
Respectively base by adopting weight distribution correlation coefficient methodAnd obtaining a first weight of the first correlation coefficient set and a second weight of the second correlation coefficient set from the first error and the second error, and obtaining a target correlation coefficient set based on the first weight, the second weight, the first correlation coefficient set and the second correlation coefficient set. Wherein the first weight satisfies:the second weight satisfies:。
the target correlation coefficient set satisfies:
wherein,representing the i-th target correlation coefficient, i=1, 2, …, 7, i.e. target correlation coefficient set +.>Represented asTarget correlation coefficient set->Can be seen as a 7 x 1 matrix. Target correlation coefficient set->Also known as inertial weight correlation coefficients.
Step S14, constructing a neural network model, taking initial weights and initial thresholds of all neurons of the neural network model as position vectors of whales, taking an error function as an fitness function, adopting an improved whale algorithm to obtain optimal weights and optimal thresholds of all neurons, and obtaining a target neural network model based on the optimal weights and the optimal thresholds.
In step S14, the constructed neural network model employs a BP (Back Propagation) neural network.
The BP neural network is a multi-layer feedforward network trained by an error back propagation algorithm, and comprises an input layer, an hidden layer and an output layer, wherein the BP neural network is divided into two stages of forward transmission and back propagation, a signal in the forward transmission is transmitted from the input layer to the output layer through the hidden layer through a series of calculation to calculate a predicted value, and if the error between the predicted value of the output layer and the actual value of the input layer does not reach a threshold value, the weight and the bias of the network are updated through the error back propagation, so that the error is reduced. Each of the input layer, the hidden layer and the output layer contained in the BP neural network comprises at least one neuron, and each neuron has an initial weight and an initial threshold.
In step S14, the modified whale algorithm used is modified on the basis of the conventional whale algorithm.
As will be readily appreciated, conventional whale algorithms (Whale Optimization Algorithm, WOA) generally include encircling a target prey, spiral position updating, and searching for prey.
Wherein, in surrounding the target hunting object, taking the position of the whale of the head closest to the hunting object position as an optimal solution, the rest whales of the head gradually update the self position with the aim of approaching to the position of the optimal solution, wherein, the calculation formula of the mathematical model surrounding the target hunting object is as follows:
In the method, in the process of the invention,Dfor the distance of the whale individual from the current optimal solution,is the globally optimal solution corresponding to the t-th iteration,for the position of whale corresponding to the t-th iteration,Aas a result of the first coefficient of the coefficient,Cis a second coefficient>For the position of whale corresponding to iteration t+1, wherein +.>、/>，/>For the first random vector, +.>For the second random vector,>,all belong to [0,1 ]]，/>For astringing factor, ++>，/>Is the maximum number of iterations.
In the spiral position update, the spiral path is expressed mathematically as follows, in the form of a spiral rise to the prey:
in the method, in the process of the invention,=|X*(t)-X(t)|，/>indicating the distance between whale and prey,bis a constant value, which is set to be a constant value,lis located at [ -1,1]Random numbers in between. When hunting whale, it is assumed that whale is presentThe 50% probability was chosen randomly between constriction and spiral ascent, optimizing the position of whales. The expression is as follows:
in searching for a prey, the mathematical model of the search for a prey stage is calculated as follows:
in the method, in the process of the invention,the positions of whale individuals randomly selected for the current iteration t.
Considering that WOA algorithm is nonlinear in evolutionary search process, convergence factoraThe linear decreasing strategy cannot fully embody the actual optimization searching process of the algorithm, so in step S14, the convergence factor of the whale algorithm is optimized by using the initial value and the final value of the convergence factor, the nonlinear equalization coefficient and the maximum iteration number, so as to obtain an improved whale algorithm.
Specifically, the optimized convergence factora’The method meets the following conditions:
in the method, in the process of the invention,and->Respectively the initial value and the final value of the convergence factor, < >>Is a nonlinear equalization coefficient, whose value is at [0,1]Replacing the convergence factor in the traditional whale algorithm with the optimized convergence factor to obtain an improved whale algorithm (i.e. IWOA),the improvement balances the global and local search capabilities of WOA.
In step S14, the initial weight and the initial threshold of each neuron of the neural network model are used as the position vector of the whale, the error function is used as the fitness function, the improved whale algorithm is adopted to obtain the optimal weight and the optimal threshold of each neuron, and the target neural network model is obtained based on the optimal weight and the optimal threshold.
Specifically, the process of obtaining the target neural network model is as follows:
1) BP neural network initialization: determining an initial structure of the neural network, and connecting initial weights and initial thresholds of neurons;
2) The whale algorithm is initialized, namely the initial weight value and the initial threshold value in the first step are converted into position vectors of whales, population scale, maximum iteration times, initial minimum weight, maximum weight and convergence factors are set, and mean square error is selected as an optimized objective function;
3) Calculating individual fitness value, namely finding and recording the individual position of the optimal fitness value, and taking the individual position as the optimal individual position;
4) Giving the weight and threshold parameters obtained by optimization to the BP neural network;
5) Updating the individual location according to a location update policy;
6) And setting a termination strategy, terminating operation after the maximum iteration times are met or the accuracy requirement is met, and giving an optimal weight and an optimal threshold to the BP neural network to obtain a target neural network model (which can be called as an IWOA-BP model).
And S15, obtaining target input data based on various voltage influence parameters and target correlation coefficient sets, and inputting the target input data into a target neural network model to output a switching-on overvoltage target predicted value.
In step S15, target input data is obtained based on the plurality of voltage influence parameters and the target correlation coefficient set. For inputting data to a targetMRepresenting target input dataMSatisfy the following requirements(see FIG. 2)Inputting the target into the dataMAnd inputting the target voltage target prediction value into an IWOA-BP model to perform prediction to obtain a final prediction result, wherein the final prediction result is the required switching-on overvoltage target prediction value.
In order to verify the effect of the method disclosed by the invention, an equivalent network for calculating the closing overvoltage of the 500 kV no-load line in the black start process is constructed. The effectiveness of the combined model-based black start air charge line statistical overvoltage fast prediction method is verified by using the network as an example. Setting a power supply resistor Rs=10Ω -40Ω with an interval of 10Ω; the length l=260 km-340 km of the closing line, and the interval is 40km; the compensation value Q=50 MVAR-60 MVAR of the shunt reactor, the interval is 5 MVAR; the leakage reactance xs=125Ω -140Ω, 5 Ω interval.
The various voltage influence parameters comprise 7 voltage influence parameters including a closing line length, a line shunt reactor compensation value, a power supply resistance, a power supply leakage reactance, a line resistance per km, a line positive sequence reactance per km and a closing initial phase angle.
Mean absolute error (Mean absolute error, MAE), mean absolute percent error (Mean absolute percentage error, MAPE) and root mean square error (Root mean square error, RMSE) were chosen as model evaluation criteria, where all models used MAE as a loss function during training.
The pearson coefficient-IWOA-BP model, the cosine similarity-IWOA-BP model, and the inertial weight correlation coefficient-WOA-BP were compared with the model of the present disclosure (inertial weight correlation coefficient-IWOA-BP), respectively, at the time of the experiment. The results are shown in Table 1.
Table 1 model evaluation criteria table
| Model | MAE | MAPE | RMSE |
| Piercan coefficient-IWOA-BP | 0.0702 | 0.0392 | 0.267 |
| Cosine similarity-IWOA-BP | 0.0698 | 0.0430 | 0.244 |
| Inertial weight correlation coefficient-WOA-BP | 0.0432 | 0.0211 | 0.201 |
| Inertial weight correlation coefficient-IWOA-BP | 0.0221 | 0.0198 | 0.187 |
From table 1, it can be seen that the model using the inertia weight coefficient is higher than the prediction accuracy using the pearson coefficient or cosine similarity alone, demonstrating the importance of the inertia weight coefficient to improve accuracy. The prediction precision of the inertial weight correlation coefficient-IWAA-BP model is higher than that of the inertial weight correlation coefficient-WOA-BP model, and the fact that the nonlinear convergence factor is adopted can balance the global and local searching capacity of WOA is proved.
In the black start overvoltage prediction method based on the weight distribution correlation coefficient, which is disclosed by the embodiment of the invention, the closing overvoltage of an idle circuit and various voltage influence parameters in the black start process of a power grid are obtained, wherein the various voltage influence parameters comprise the length of the closing circuit, the compensation value of a circuit shunt reactor, a power supply resistance, a power supply leakage reactance, the resistance of the circuit every kilometer, the positive sequence reactance of the circuit every kilometer and the closing initial phase angle; calculating a first correlation coefficient set of a switching-on overvoltage and various voltage influence parameters by using the Pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and various voltage influence parameters by using cosine similarity; obtaining model input data based on various voltage influence parameters, a first correlation coefficient set and a second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method; constructing a neural network model, taking initial weights and initial thresholds of all neurons of the neural network model as position vectors of whales, taking an error function as an fitness function, adopting an improved whale algorithm to obtain optimal weights and optimal thresholds of all neurons, and obtaining a target neural network model based on the optimal weights and the optimal thresholds; and obtaining target input data based on various voltage influence parameters and target correlation coefficient sets, and inputting the target input data into a target neural network model to output a switching-on overvoltage target predicted value. Under the condition, the influence of the length of a closing line, the compensation value of a line shunt reactor, the power supply resistance, the leakage reactance of a power supply, the resistance of each kilometer of the line, the positive sequence reactance of each kilometer of the line and the closing initial phase angle on the closing overvoltage is comprehensively considered, the influence degree of different factors on the overvoltage is fully considered by combining the pearson correlation coefficient, the cosine similarity and the weight distribution correlation coefficient method, the accuracy of the black-start overvoltage prediction is improved, and in addition, the accuracy of a neural network model is improved by adopting an improved whale algorithm, and the accuracy of the black-start overvoltage prediction is further improved.
The method is a black start operation overvoltage prediction method based on a combined prediction model of weight distribution correlation coefficients, the weight distribution correlation coefficient method is adopted, the pearson correlation coefficient method and cosine similarity are adopted to extract the characteristics with high correlation degree respectively, then new characteristics are constructed according to the known characteristics, the new characteristics are respectively put into GRUs for prediction, the weight coefficients are established according to the prediction results to form new correlation coefficients (namely target correlation coefficient sets), and finally the known new correlation coefficients are multiplied with original data (namely various voltage influence parameters) to be input into IWOA-BP for prediction. Therefore, the defect of the method when the pearson correlation coefficient or the cosine similarity is adopted independently is overcome, namely the problem that the correlation between the characteristic value and the overvoltage cannot be accurately mined is solved. In addition, the improved whale algorithm is utilized to overcome the defects of low solution precision, low convergence speed and the like of the traditional whale algorithm when the complex global optimization problem is processed.
The following are system embodiments of the present disclosure that may be used to perform method embodiments of the present disclosure. For details not disclosed in the embodiments of the disclosed system, please refer to the embodiments of the disclosed method.
Referring to fig. 3, fig. 3 is a block diagram illustrating a black-start overvoltage prediction system based on a weight distribution correlation coefficient according to an embodiment of the present disclosure. The black start overvoltage prediction system based on the weight distribution correlation coefficient can be implemented as all or part of the system by software, hardware or a combination of the two. The black-start overvoltage prediction system 10 based on the weight distribution correlation coefficient includes an acquisition module 11, a coefficient calculation module 12, a weight distribution module 13, a modeling module 14, and a prediction module 15, wherein:
the acquisition module 11 is configured to acquire a closing overvoltage of an idle line and various voltage influence parameters in a black start process of the power grid, where the various voltage influence parameters include a closing line length, a compensation value of a line shunt reactor, a power supply resistance, a power supply leakage reactance, a resistance of each kilometer of the line, a positive sequence reactance of each kilometer of the line, and a closing initial phase angle;
a coefficient calculating module 12, configured to calculate a first correlation coefficient set of the closing overvoltage and the multiple voltage influence parameters using the pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and various voltage influence parameters by using cosine similarity;
the weight distribution module 13 is configured to obtain model input data based on a plurality of voltage influence parameters, a first correlation coefficient set, and a second correlation coefficient set, input the model input data into a cyclic neural network model to obtain an overvoltage prediction value set, and obtain a target correlation coefficient set based on a closing overvoltage, the overvoltage prediction value set, the first correlation coefficient set, and the second correlation coefficient set by adopting a weight distribution correlation coefficient method;
The modeling module 14 is configured to construct a neural network model, take an initial weight and an initial threshold of each neuron of the neural network model as a position vector of a whale, take an error function as an fitness function, obtain an optimal weight and an optimal threshold of each neuron by adopting an improved whale algorithm, and obtain a target neural network model based on the optimal weight and the optimal threshold;
the prediction module 15 is configured to obtain target input data based on a plurality of voltage influence parameters and a target correlation coefficient set, and input the target input data into the target neural network model to output a target predicted value of the closing overvoltage.
Optionally, in the modeling module 14, the convergence factor of the whale algorithm is optimized with its initial and final values, nonlinear equalization coefficients, and maximum number of iterations to obtain an improved whale algorithm.
Alternatively, in modeling module 14, the neural network model employs a BP neural network.
Optionally, the model input data comprises first model input data and second model input data, and the set of overvoltage predictors comprises a first set of overvoltage predictors and a second set of overvoltage predictors.
Optionally, the weight distribution module 13 is specifically configured to: acquiring first model input data based on various voltage influence parameters and a first correlation coefficient set, and inputting the first model input data into a cyclic neural network model to acquire a first overvoltage predicted value set; and obtaining second model input data based on the multiple voltage influence parameters and the second correlation data set, and inputting the second model input data into the cyclic neural network model to obtain a second overvoltage predicted value set.
Optionally, the weight distribution module 13 is specifically configured to: obtaining a first error based on the closing overvoltage and the first overvoltage predicted value set; obtaining a second error based on the closing overvoltage and a second overvoltage predicted value set; and obtaining a target correlation coefficient set based on the first error, the second error, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method.
It should be noted that, when the black-start overvoltage prediction system based on the weight distribution correlation coefficient provided in the foregoing embodiment performs the black-start overvoltage prediction method based on the weight distribution correlation coefficient, only the division of the foregoing functional modules is used for illustration, and in practical application, the foregoing functional distribution may be completed by different functional modules according to needs, that is, the internal structure of the black-start overvoltage prediction device based on the weight distribution correlation coefficient is divided into different functional modules, so as to complete all or part of the functions described above. In addition, the black-start overvoltage prediction system based on the weight distribution correlation coefficient provided in the above embodiment and the black-start overvoltage prediction method embodiment based on the weight distribution correlation coefficient belong to the same concept, which embody the implementation process in detail and are not described herein.
The foregoing embodiment numbers of the present disclosure are merely for description and do not represent advantages or disadvantages of the embodiments.
In the black start overvoltage prediction system based on the weight distribution correlation coefficient, the acquisition module is used for acquiring the closing overvoltage of the idle circuit and various voltage influence parameters in the black start process of the power grid, wherein the various voltage influence parameters comprise the length of the closing circuit, the compensation value of the parallel reactor of the circuit, the power supply resistance, the leakage reactance of the power supply, the resistance of the circuit every kilometer, the positive sequence reactance of the circuit every kilometer and the closing initial phase angle; the coefficient calculation module is used for calculating a first correlation coefficient set of the switching-on overvoltage and various voltage influence parameters by using the Pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and various voltage influence parameters by using cosine similarity; the weight distribution module is used for obtaining model input data based on various voltage influence parameters, a first correlation coefficient set and a second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method; the modeling module is used for constructing a neural network model, taking the initial weight and the initial threshold value of each neuron of the neural network model as a position vector of a whale, taking an error function as an fitness function, adopting an improved whale algorithm to obtain the optimal weight and the optimal threshold value of each neuron, and obtaining a target neural network model based on the optimal weight and the optimal threshold value; the prediction module is used for obtaining target input data based on various voltage influence parameters and target correlation coefficient sets, inputting the target input data into a target neural network model and outputting a switching-on overvoltage target predicted value. Under the condition, the influence of the length of a closing line, the compensation value of a line shunt reactor, the power supply resistance, the leakage reactance of a power supply, the resistance of each kilometer of the line, the positive sequence reactance of each kilometer of the line and the closing initial phase angle on the closing overvoltage is comprehensively considered, the influence degree of different factors on the overvoltage is fully considered by combining the pearson correlation coefficient, the cosine similarity and the weight distribution correlation coefficient method, the accuracy of the black-start overvoltage prediction is improved, and in addition, the accuracy of a neural network model is improved by adopting an improved whale algorithm, and the accuracy of the black-start overvoltage prediction is further improved.
According to embodiments of the present disclosure, the present disclosure also provides a black-start overvoltage prediction device, a readable storage medium, and a computer program product based on the weight distribution correlation coefficient.
Fig. 4 is a block diagram of a weight distribution correlation coefficient-based black-start overvoltage prediction device used to implement a weight distribution correlation coefficient-based black-start overvoltage prediction method of an embodiment of the present disclosure. The weight distribution correlation coefficient based black-start overvoltage prediction device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The weight distribution correlation coefficient based black-start overvoltage prediction device may also represent various forms of mobile devices such as personal digital processing, cellular phones, smart phones, wearable electronics, and other similar computing devices. The components, connections and relationships of components, and functions of components shown in this disclosure are exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed in this disclosure.
As shown in fig. 4, the black-start overvoltage predicting device 20 based on the weight distribution correlation coefficient includes a calculating unit 21 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 22 or a computer program loaded from a storage unit 28 into a Random Access Memory (RAM) 23. In the RAM 23, various programs and data required for the operation of the black-start overvoltage predicting device 20 based on the weight distribution correlation coefficient may also be stored. The computing unit 21, the ROM 22 and the RAM 23 are connected to each other via a bus 24. An input/output (I/O) interface 25 is also connected to bus 24.
A plurality of components in the black-start overvoltage predicting device 20 based on the weight distribution correlation coefficient are connected to the I/O interface 25, including: an input unit 26 such as a keyboard, a mouse, etc.; an output unit 27 such as various types of displays, speakers, and the like; a storage unit 28, such as a magnetic disk, an optical disk, or the like, the storage unit 28 being communicatively connected to the computing unit 21; and a communication unit 29 such as a network card, modem, wireless communication transceiver, etc. The communication unit 29 allows the weight-distribution correlation coefficient-based black-start overvoltage prediction device 20 to exchange information/data with other weight-distribution correlation coefficient-based black-start overvoltage prediction devices through a computer network such as the internet and/or various telecommunication networks.
The computing unit 21 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 21 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The calculation unit 21 performs the respective methods and processes described above, for example, performs a black-start overvoltage prediction method based on the weight distribution correlation coefficient. For example, in some embodiments, the black start overvoltage prediction method based on the weight distribution correlation coefficient may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as the storage unit 28. In some embodiments, part or all of the computer program may be loaded and/or installed on the black-start overvoltage prediction device 20 based on the weight distribution correlation coefficient via the ROM 22 and/or the communication unit 29. When the computer program is loaded into the RAM 23 and executed by the calculation unit 21, one or more steps of the above-described black-start overvoltage prediction method based on the weight distribution correlation coefficient may be performed. Alternatively, in other embodiments, the computing unit 21 may be configured to perform the black-start overvoltage prediction method based on the weight distribution correlation coefficient in any other suitable way (e.g. by means of firmware).
Various implementations of the systems and techniques described above in this disclosure may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In this disclosure, a machine-readable medium may be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or black-start overvoltage prediction device based on weight distribution correlation coefficients. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or electronic device, or any suitable combination of the preceding. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical electronic storage device, a magnetic electronic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), the internet, and blockchain networks.
The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service ("Virtual Private Server" or simply "VPS") are overcome. The server may also be a server of a distributed system or a server that incorporates a blockchain.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present disclosure may be performed in parallel, sequentially, or in a different order, so long as the desired result of the technical solution of the present disclosure can be achieved, and the present disclosure is not limited herein.
The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.
Claims (10)
1. The method for predicting the black start overvoltage based on the weight distribution correlation coefficient is characterized by comprising the following steps of:
acquiring a closing overvoltage of an idle circuit and various voltage influence parameters in a black start process of a power grid, wherein the various voltage influence parameters comprise a closing circuit length, a compensation value of a circuit shunt reactor, a power supply resistance, a power supply leakage reactance, a circuit resistance per kilometer, a positive sequence reactance per kilometer and a closing initial phase angle;
Calculating a first correlation coefficient set of the switching-on overvoltage and the plurality of voltage influence parameters by using a pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and the multiple voltage influence parameters by using cosine similarity;
obtaining model input data based on the various voltage influence parameters, the first correlation coefficient set and the second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method;
constructing a neural network model, taking initial weights and initial thresholds of neurons of the neural network model as position vectors of whales, taking an error function as an fitness function, adopting an improved whale algorithm to obtain optimal weights and optimal thresholds of the neurons, and obtaining a target neural network model based on the optimal weights and the optimal thresholds;
and obtaining target input data based on the various voltage influence parameters and the target correlation coefficient set, and inputting the target input data into the target neural network model to output a switching-on overvoltage target predicted value.
2. The method of claim 1, wherein the convergence factor of the whale algorithm is optimized using an initial value and a final value of the convergence factor, a nonlinear equalization coefficient, and a maximum number of iterations to obtain the improved whale algorithm.
3. The method for predicting black-start overvoltage based on weight distribution correlation coefficient according to claim 1, wherein the neural network model adopts a BP neural network.
4. The weight distribution correlation coefficient based black-start overvoltage prediction method according to claim 1, wherein the model input data includes first model input data and second model input data, and the set of overvoltage predictors includes a first set of overvoltage predictors and a second set of overvoltage predictors;
the obtaining model input data based on the plurality of voltage influence parameters, the first correlation coefficient set and the second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, includes:
obtaining first model input data based on the multiple voltage influence parameters and the first correlation coefficient set, and inputting the first model input data into a cyclic neural network model to obtain a first overvoltage predicted value set;
And obtaining second model input data based on the multiple voltage influence parameters and the second correlation coefficient set, and inputting the second model input data into a cyclic neural network model to obtain a second overvoltage predicted value set.
5. The method for predicting black-start overvoltage based on weight distribution correlation coefficient according to claim 4, wherein said obtaining a target correlation coefficient set based on the closing overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by weight distribution correlation coefficient method comprises:
obtaining a first error based on the closing overvoltage and the first overvoltage predicted value set;
obtaining a second error based on the closing overvoltage and the second overvoltage predicted value set;
and obtaining a target correlation coefficient set based on the first error, the second error, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method.
6. A black start overvoltage prediction system based on a weight distribution correlation coefficient, comprising:
the acquisition module is used for acquiring the switching-on overvoltage of the idle circuit and various voltage influence parameters in the black start process of the power grid, wherein the various voltage influence parameters comprise the length of the switching-on circuit, the compensation value of a parallel reactor of the circuit, the power supply resistance, the leakage reactance of the power supply, the resistance of the circuit every kilometer, the positive sequence reactance of the circuit every kilometer and the switching-on initial phase angle;
The coefficient calculation module is used for calculating a first correlation coefficient set of the switching-on overvoltage and the various voltage influence parameters by using the Pearson correlation coefficient; calculating a second phase relation number set of the switching-on overvoltage and the multiple voltage influence parameters by using cosine similarity;
the weight distribution module is used for obtaining model input data based on the various voltage influence parameters, the first correlation coefficient set and the second correlation coefficient set, inputting the model input data into a cyclic neural network model to obtain an overvoltage predicted value set, and obtaining a target correlation coefficient set based on the switching-on overvoltage, the overvoltage predicted value set, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method;
the modeling module is used for constructing a neural network model, taking the initial weight and the initial threshold value of each neuron of the neural network model as a position vector of a whale, taking an error function as an fitness function, adopting an improved whale algorithm to obtain the optimal weight and the optimal threshold value of each neuron, and obtaining a target neural network model based on the optimal weight and the optimal threshold value;
and the prediction module is used for obtaining target input data based on the various voltage influence parameters and the target correlation coefficient set, inputting the target input data into the target neural network model and outputting a closing overvoltage target predicted value.
7. The weight distribution correlation coefficient based black start overvoltage prediction system according to claim 6, wherein the convergence factor of the whale algorithm is optimized in the modeling module using an initial value and a final value of the convergence factor, a nonlinear equalization coefficient, and a maximum number of iterations to obtain the improved whale algorithm.
8. The black-start overvoltage prediction system according to claim 6, wherein the model input data comprises a first model input data and a second model input data, the set of overvoltage predictors comprises a first set of overvoltage predictors and a second set of overvoltage predictors, the weight distribution module is configured to: obtaining first model input data based on the multiple voltage influence parameters and the first correlation coefficient set, and inputting the first model input data into a cyclic neural network model to obtain a first overvoltage predicted value set; and obtaining second model input data based on the multiple voltage influence parameters and the second correlation coefficient set, and inputting the second model input data into a cyclic neural network model to obtain a second overvoltage predicted value set.
9. The black-start overvoltage prediction system according to claim 8, wherein the weight distribution module is specifically configured to: obtaining a first error based on the closing overvoltage and the first overvoltage predicted value set; obtaining a second error based on the closing overvoltage and the second overvoltage predicted value set; and obtaining a target correlation coefficient set based on the first error, the second error, the first correlation coefficient set and the second correlation coefficient set by adopting a weight distribution correlation coefficient method.
10. A black-start overvoltage predicting device based on a weight distribution correlation coefficient, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the weight distribution correlation coefficient-based black start overvoltage prediction method of any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311318800.3A CN117057258B (en) | 2023-10-12 | 2023-10-12 | Black-start overvoltage prediction method and system based on weight distribution correlation coefficient |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311318800.3A CN117057258B (en) | 2023-10-12 | 2023-10-12 | Black-start overvoltage prediction method and system based on weight distribution correlation coefficient |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117057258A CN117057258A (en) | 2023-11-14 |
| CN117057258B true CN117057258B (en) | 2023-12-12 |
Family
ID=88659472
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311318800.3A Active CN117057258B (en) | 2023-10-12 | 2023-10-12 | Black-start overvoltage prediction method and system based on weight distribution correlation coefficient |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117057258B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117728504B (en) * | 2024-02-18 | 2024-05-07 | 西安热工研究院有限公司 | Black start system and method for diesel-engine combined combustion engine |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015138774A1 (en) * | 2014-03-13 | 2015-09-17 | Sequenom, Inc. | Methods and processes for non-invasive assessment of genetic variations |
| CN115034493A (en) * | 2022-06-24 | 2022-09-09 | 江苏方天电力技术有限公司 | Wind power plant black start path optimization method considering unit operation state |
| CN115859057A (en) * | 2022-12-29 | 2023-03-28 | 山东工商学院 | Sea cucumber culture water quality prediction method based on whale algorithm optimization GRU neural network |
| CN116861694A (en) * | 2023-07-24 | 2023-10-10 | 四川威鹏电缆制造股份有限公司 | Cable temperature prediction method based on artificial intelligence |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112653142B (en) * | 2020-12-18 | 2022-09-02 | 武汉大学 | Wind power prediction method and system for optimizing depth transform network |
-
2023
- 2023-10-12 CN CN202311318800.3A patent/CN117057258B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015138774A1 (en) * | 2014-03-13 | 2015-09-17 | Sequenom, Inc. | Methods and processes for non-invasive assessment of genetic variations |
| CN115034493A (en) * | 2022-06-24 | 2022-09-09 | 江苏方天电力技术有限公司 | Wind power plant black start path optimization method considering unit operation state |
| CN115859057A (en) * | 2022-12-29 | 2023-03-28 | 山东工商学院 | Sea cucumber culture water quality prediction method based on whale algorithm optimization GRU neural network |
| CN116861694A (en) * | 2023-07-24 | 2023-10-10 | 四川威鹏电缆制造股份有限公司 | Cable temperature prediction method based on artificial intelligence |
Non-Patent Citations (2)
| Title |
|---|
| 基于神经网络的黑启动操作过电压快速预测;李膨源;顾雪平;;电网技术(第03期);全文 * |
| 基于鲸鱼优化小波神经网络的飞行航迹缺失数据预测方法;石旭东;姜鸿晔;;计算机应用与软件(第07期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117057258A (en) | 2023-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10460230B2 (en) | Reducing computations in a neural network | |
| JP2021166109A (en) | Fusion sorting model training method and device, search sorting method and device, electronic device, storage medium, and program | |
| CN117057258B (en) | Black-start overvoltage prediction method and system based on weight distribution correlation coefficient | |
| CN107766573B (en) | Commodity recommendation method, commodity recommendation device, commodity recommendation equipment and storage medium based on data processing | |
| CN112580733B (en) | Classification model training method, device, equipment and storage medium | |
| CN107203558B (en) | Object recommendation method and device, and recommendation information processing method and device | |
| CN113553755B (en) | Power system state estimation method, device and equipment | |
| CN110135681A (en) | Risk subscribers recognition methods, device, readable storage medium storing program for executing and terminal device | |
| CN106296434B (en) | Grain yield prediction method based on PSO-LSSVM algorithm | |
| WO2022056841A1 (en) | Neural architecture search via similarity-based operator ranking | |
| CN112949835A (en) | Inference method and device for knowledge graph based on convolution cyclic neural network | |
| Li et al. | Improved LSTM-based prediction method for highly variable workload and resources in clouds | |
| CN116861256A (en) | Furnace temperature prediction method, system, equipment and medium for solid waste incineration process | |
| CN116125279A (en) | Method, device, equipment and storage medium for determining battery health state | |
| CN115345257A (en) | Flight trajectory classification model training method, classification method, device and storage medium | |
| CN114817476A (en) | Language model training method and device, electronic equipment and storage medium | |
| CN114970732A (en) | Posterior calibration method and device for classification model, computer equipment and medium | |
| CN113313049A (en) | Method, device, equipment, storage medium and computer program product for determining hyper-parameters | |
| CN113112092A (en) | Short-term probability density load prediction method, device, equipment and storage medium | |
| CN112925913A (en) | Method, apparatus, device and computer-readable storage medium for matching data | |
| CN113010774A (en) | Click rate prediction method based on dynamic deep attention model | |
| CN115630772B (en) | Comprehensive energy detection and distribution method, system, equipment and storage medium | |
| CN117592550B (en) | Black box attack method and device for graphic neural network model | |
| WO2023065309A1 (en) | Circuit design method and apparatus | |
| CN113033196B (en) | Word segmentation method, device, equipment and storage medium |
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 |