A steam turbine unit heat rate evaluation system and method based on artificial intelligence

Abstract

This invention relates to the technical field of heat rate assessment, and discloses an artificial intelligence-based system and method for assessing the heat rate of a steam turbine unit. First, the invention preprocesses the parameters of the steam turbine unit's heat rate, calculates the heat rate, and corrects it. Second, by clustering the operating parameters of the steam turbine unit and combining them with the corrected heat rate, a backpropagation (BP) neural network is trained to identify leakage data. Then, using the steam volume parameters during the turbine unit's operation, a support vector machine (SVM) regression model is trained, and the parameters of the SVM regression model are optimized using the Tianying optimization algorithm to predict the steam consumption of the steam turbine unit. Finally, the heat rate of the steam turbine unit is assessed. This invention achieves accurate and objective assessment of the steam turbine unit's heat rate by combining heat rate assessment parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of heat rate evaluation, and in particular to a heat rate evaluation system and method for a steam turbine unit based on artificial intelligence. Background Art

[0002] Chinese patent CN109858810A discloses a method for calculating the pure condensing heat consumption rate of a steam turbine unit under heating conditions. The method specifically includes: obtaining the thermal parameters of the heating conditions, the purpose of which is to calculate the pure condensing heat consumption rate of the steam turbine unit, and then obtaining multiple sets of exhaust flow rates and exhaust specific volumes of the low-pressure cylinder under different exhaust pressures under the heating conditions, and after fitting calculation, obtaining the relationship between the exhaust flow rate and the exhaust specific volume under different exhaust pressures; secondly, obtaining the pure condensing heat consumption rate of the steam turbine unit, including returning the heating exhaust flow rate of the steam turbine unit to the low-pressure cylinder to obtain the initial exhaust flow rate of the low-pressure cylinder, and then obtaining the exhaust pressure rating of the low-pressure cylinder to calculate the exhaust specific volume of the low-pressure cylinder under the exhaust flow rate; using the exhaust specific volume to calculate a new exhaust flow rate, using the new exhaust flow rate to replace the old exhaust flow rate, and recalculating the exhaust flow rate of the low-pressure cylinder until the exhaust volume results are the same; finally, using the new exhaust flow rate to calculate the exhaust specific volume, and obtaining the pure condensing heat power of the steam turbine unit. This method directly uses the exhaust flow rate and the exhaust specific volume, and there is an error in the calculation.

[0003] The traditional steam turbine unit heat rate calculation method has problems such as inaccurate calculation data and unpredictable thermal characteristics, which brings troubles to the heat rate evaluation of steam turbine units; at the same time, the traditional steam turbine unit heat rate evaluation method does not use technologies such as neural networks, and cannot obtain accurate and objective evaluation results. Summary of the invention

[0004] In response to the problems in the related art, the present invention provides a steam turbine unit heat rate evaluation system and method based on artificial intelligence to overcome the above-mentioned technical problems existing in the existing related technology.

[0005] In order to solve the above technical problems, the present invention is achieved through the following technical solutions:

[0006] The present invention is a method for evaluating the heat rate of a steam turbine unit based on artificial intelligence, comprising the following steps:

[0007] S1. Acquire parameters of the heat rate of the steam turbine unit, preprocess the parameters of the heat rate of the steam turbine unit, calculate the heat rate of the steam turbine unit, and perform data correction on the heat rate of the steam turbine unit to obtain a corrected heat rate of the steam turbine unit;

[0008] S2. Obtain the operating parameters of the steam turbine unit in previous years, perform clustering processing using the fuzzy C-means clustering algorithm to obtain a processed set of steam turbine unit operating parameters, train a BP neural network in combination with the corrected steam turbine unit heat rate and the processed set of steam turbine unit operating parameters to identify leakage data, and record the tightness of the steam turbine unit as the first steam turbine unit heat rate evaluation parameter;

[0009] S3. Use the steam volume parameters during the operation of the steam turbine unit in previous years to train the support vector machine regression to obtain a support vector machine regression model, and optimize the support vector machine regression model parameters based on the Tianying optimization algorithm to obtain the predicted steam consumption of the steam turbine unit, and record the predicted steam consumption of the steam turbine unit as the heat rate evaluation parameter of the second steam turbine unit;

[0010] S4. Evaluate the heat rate of the steam turbine unit by combining the first steam turbine unit heat rate evaluation parameter and the second steam turbine unit heat rate evaluation parameter.

[0011] The invention first obtains the parameters of the heat rate of the steam turbine unit, performs outlier deletion processing and data filling processing on the parameters of the heat rate of the steam turbine unit, then calculates the heat rate of the steam turbine unit, and introduces a correction factor to correct the heat rate of the steam turbine unit to obtain the corrected heat rate of the steam turbine unit; secondly, the operating parameters of the steam turbine unit are obtained, and the steam turbine unit is divided into low, medium and high load intervals, and fuzzy C-means clustering processing is performed in different intervals, and the membership is continuously iterated to classify the normal state and leakage state data, and the BP neural network is trained to obtain the final BP neural network, so as to realize the recognition of leakage data, and record the tightness of the steam turbine unit as the first steam turbine unit heat rate evaluation parameter; the neural network has a high recognition rate and a high learning speed. Fast; then use the steam volume parameters during the operation of the steam turbine unit, clean and normalize the steam volume parameters during the operation of the steam turbine unit, train the support vector machine regression to obtain the support vector machine regression model, and optimize the penalty factor and kernel function parameters of the support vector machine regression model based on the Eagle optimization algorithm, output the predicted steam consumption of the steam turbine unit, and record the predicted steam consumption of the steam turbine unit as the heat rate evaluation parameter of the second steam turbine unit; the model can effectively reduce the error and improve the prediction accuracy. The algorithm imitates the behavior of the Eagle in the hunting process, with fast convergence speed and strong optimization ability; finally, the heat rate evaluation parameters of the first steam turbine unit and the heat rate evaluation parameters of the second steam turbine unit are combined to evaluate the heat rate of the steam turbine unit.

[0012] Preferably, the S1 comprises the following steps:

[0013] S11. Obtaining parameters of the heat rate of the steam turbine unit, wherein the parameters of the heat rate of the steam turbine unit include generator power, main steam, main feed water, reheated steam, high-pressure cylinder exhaust, superheater cooling, and reheater cooling. Generating a generator power set according to the parameter data of the heat rate of the steam turbine unit , Main steam assembly , main water supply assembly , Reheat steam assembly , High pressure cylinder exhaust assembly 、Superheater desuperheating assembly and reheater desuperheater assembly , we get the heat rate parameter set of the steam turbine unit, recorded as ;

[0014] The steam turbine unit heat rate parameter set is preprocessed, and the specific process is as follows:

[0015] S111, performing outlier deletion processing on the steam turbine unit heat rate parameters in the steam turbine unit heat rate parameter set, dividing the steam turbine unit heat rate parameters in the steam turbine unit heat rate parameter set into 4 equal data parts, respectively recorded as the first divided data set, the second divided data set, the third divided data set and the fourth divided data set, calculating the average value of the heat rate parameter of the rightmost steam turbine unit in the first divided data set and the heat rate parameter of the leftmost steam turbine unit in the second divided data set, and recording it as the upper quantile; calculating the average value of the heat rate parameter of the rightmost steam turbine unit in the third divided data set and the heat rate parameter of the leftmost steam turbine unit in the fourth divided data set, and recording it as the lower quantile, then the parameter range calculation formula is as follows,

[0016] ,

[0017] in, Indicates that the parameters are extremely poor. represents the upper quantile, represents the lower quantile;

[0018] When the steam turbine unit heat rate parameter in the steam turbine unit heat rate parameter set is greater than Or the steam turbine unit heat rate parameter in the steam turbine unit heat rate parameter set is less than When , the steam turbine unit heat rate parameter in the steam turbine unit heat rate parameter set is an abnormal value, the abnormal value is deleted, and the steam turbine unit heat rate parameter set after the abnormal value is deleted is obtained;

[0019] S112, filling data with the heat rate parameters of the steam turbine unit in the heat rate parameter set of the steam turbine unit after the outliers are deleted, initializing the heat rate parameter set of the steam turbine unit after the outliers are deleted, constructing a complete value data matrix, calculating the Euclidean distance between the heat rate parameters of the steam turbine unit in the heat rate parameter set of the steam turbine unit after the outliers are deleted and the data in the complete value data matrix, and obtaining a distance set; in the distance set, selecting k distance data with the smallest distance value as k nearest neighbor data, calculating the weights of the k nearest neighbor data, estimating the missing values ​​using the weights of the k nearest neighbor data, and filling the missing values ​​in the heat rate parameter set of the steam turbine unit after the outliers are deleted; repeating S112 until there are no missing values ​​in the heat rate parameter set of the steam turbine unit after the outliers are deleted, and obtaining a processed heat rate parameter set of the steam turbine unit;

[0020] S12, setting the steam turbine unit heat rate parameter to be selected from the processed steam turbine unit heat rate parameter set, the generator power is , the main steam flow is , the main steam enthalpy is , the main water flow rate is , the main feed water enthalpy is , the reheat steam flow rate is , the reheat steam enthalpy is , the high pressure cylinder exhaust flow rate is , the exhaust enthalpy of the high pressure cylinder is , the superheater desuperheating flow rate is , the superheater cooling enthalpy is , the reheater desuperheating flow rate is , the reheater de-temperature enthalpy is , then the heat rate of the steam turbine unit is B The calculation formula is as follows,

[0021] ;

[0022] Set the correction factor for the heat rate of the steam turbine unit to form a correction factor set ,in Indicates the main steam correction factor, Indicates the main feed water correction factor, Reheat steam correction factor, Indicates the high pressure cylinder exhaust correction factor, Indicates the superheater desuperheating correction factor, represents the reheater temperature reduction correction factor; the weight corresponding to the correction factor in the correction factor set is recorded as the weight set, and the comprehensive correction factor is obtained by the weight set and the correction factor set. The calculation formula of the corrected steam turbine unit heat rate is as follows:

[0023] ,

[0024] in, represents the corrected heat rate of the steam turbine unit, Indicates the comprehensive correction factor.

[0025] The invention obtains parameters of the heat rate of the steam turbine unit, divides the parameters of the heat rate of the steam turbine unit, screens out abnormal values ​​by the degree of data discreteness, and then uses the k-nearest neighbor algorithm to supplement the missing values ​​by the data weights near the missing values, calculates the heat rate of the steam turbine unit, and introduces a correction factor to correct the heat rate of the steam turbine unit, so that the corrected heat rate of the steam turbine unit can objectively reflect the energy consumption level of the steam turbine unit.

[0026] Preferably, S2 comprises the following steps:

[0027] S21. Obtain the operating parameters of the steam turbine units in previous years to obtain a set of operating parameters of the steam turbine units to be processed, recorded as ,in Indicates a The steam turbine unit operating parameters to be processed are then divided into a steam turbine unit operating load range, a steam turbine unit medium load range and a steam turbine unit high load range;

[0028] Set the number of cluster centers of the turbine unit operating parameter set to be processed in the low load interval of the turbine unit to be , the weighted index is , the membership degree is ; The number of cluster centers of the set of operating parameters of the steam turbine unit to be processed in the load interval of the steam turbine unit is , the weighted index is , the membership degree is ; The number of cluster centers of the set of steam turbine unit operating parameters to be processed in the high load interval of the steam turbine unit is , the weighted index is , the membership degree is ; Calculate the cluster center of the set of operating parameters of the steam turbine unit to be processed in the low load range of the steam turbine unit , get the first cluster center set; calculate the cluster center of the steam turbine unit operating parameter set to be processed in the load interval of the steam turbine unit , obtain the second cluster center set; calculate the cluster center of the set of steam turbine unit operating parameters to be processed in the high load interval of the steam turbine unit , get the third cluster center set; the calculation formula is as follows,

[0029] , , ,

[0030] in, Indicates that the set of steam turbine unit operating parameters to be processed is the first in the low load range of the steam turbine unit. The membership degree corresponding to the operating parameters of each steam turbine unit is Indicates that the set of steam turbine unit operating parameters to be processed is the first in the load range of the steam turbine unit. The membership degree corresponding to the operating parameters of each steam turbine unit is Indicates that the set of steam turbine unit operating parameters to be processed is in the high load range of the steam turbine unit. The membership degree corresponding to the operating parameters of each steam turbine unit is Indicates the first The operating parameters of the steam turbine units are: Indicates the first The operating parameters of the steam turbine units are: Indicates the first The operating parameters of the steam turbine units are: , , ;

[0031] S22, set the minimization objective function of the set of operating parameters of the steam turbine unit to be processed as , respectively obtain the first minimization objective function value, the second minimization objective function value and the third minimization objective function value; calculate the weighted sum of the first cluster center set and the first minimization objective function value, recorded as the first weighted sum, and set the first threshold to , when the first weighted sum is less than , clustering is completed, and the first steam turbine unit operating parameter set is obtained, otherwise the membership degree is reselected and S21 is repeated until the first weighted sum is less than ; Calculate the weighted sum of the second cluster center set and the second minimization objective function value, recorded as the second weighted sum, and set the second threshold to , when the second weighted sum is less than , clustering is completed, and the second steam turbine unit operating parameter set is obtained, otherwise the membership degree is reselected and S21 is repeated until the second weighted sum is less than ; Calculate the weighted sum of the third cluster center set and the third minimization objective function value, recorded as the third weighted sum, and set the third threshold to , when the third weighted sum is less than , clustering is completed, and the third steam turbine unit operating parameter set is obtained, otherwise reselect the membership degree and repeat S21 until the third weighted sum is less than ;

[0032] According to the first steam turbine unit operating parameter set, the second steam turbine unit operating parameter set and the third steam turbine unit operating parameter set, a processed steam turbine unit operating parameter set is formed; the number of input nodes of the BP (backward propagation) neural network is set to , the number of hidden nodes is , the learning rate is , the processed steam turbine unit operating parameter set is divided into a parameter training set and a parameter test set, the maximum number of iterations is set to C, the current number of iterations is set to c, the parameter training set is input into the BP neural network, and iterates continuously. When the current number of iterations is equal to the maximum number of iterations, the iteration is stopped to obtain a trained BP neural network; then the parameter test set is input into the trained BP neural network, and the error threshold is set to , when the error of the trained BP neural network is less than , get the final BP neural network, otherwise adjust the weights until the error of the trained BP neural network is less than ;

[0033] S23, obtaining the operating parameters of the current steam turbine unit, combining them with the corrected steam turbine unit heat rate, obtaining the current steam turbine unit operating parameter set, inputting the current steam turbine unit operating parameter set into the final BP neural network, and the final BP neural network outputs a recognition result; if the BP neural network outputs 10, then there is leakage data in the current steam turbine unit operating parameter set, that is, the steam turbine unit tightness does not meet the requirements; if the BP neural network outputs 01, then there is no leakage data in the current steam turbine unit operating parameter set, that is, the steam turbine unit tightness meets the requirements; the steam turbine unit tightness is recorded as the first steam turbine unit heat rate evaluation parameter.

[0034] The invention obtains the operating parameters of the steam turbine unit, divides it into different load intervals, uses the fuzzy C-means clustering algorithm to perform clustering processing, and clusters similar feature data into one category by continuously iterating the cluster center and membership degree, and trains the BP neural network to realize the recognition of leakage data and judge the tightness of the steam turbine unit; the fuzzy C-means clustering algorithm has good adaptability to fuzzy data sets, and the BP neural network has fast recognition speed and high accuracy.

[0035] Preferably, S3 comprises the following steps:

[0036] S31, obtaining steam quantity parameters during the operation of the steam turbine unit in previous years, the steam quantity parameters including boiler bed pressure, boiler bed temperature, feed water quantity, combustion feed quantity, primary air feed quantity, secondary air feed quantity, furnace temperature and furnace pressure, and generating a steam parameter set, the steam parameters in the steam parameter set include jdata elements; calculate the average value of the steam parameters in the steam parameter set and standard deviation , the data elements of the steam parameters in the steam parameter set are less than The data elements are replaced by , the data elements of the steam parameters in the steam parameter set are greater than The data elements are replaced by , obtain the steam parameter set after cleaning;

[0037] The steam parameter set after cleaning is normalized to obtain a normalized steam parameter set. The calculation formula is as follows:

[0038] ,

[0039] in, Represents the first in the normalized steam parameter set i The steam parameters j data elements, Indicates the first i The steam parameters j data elements, Indicates the first i The minimum data element for a steam parameter, Indicates the first i The largest data element of a steam parameter, i =1,2,3,...,8;

[0040] S32. Set the penalty factor for support vector machine regression to , the kernel function parameters are , the normalized steam parameter set is divided into a normalized training set and a normalized test set, the normalized training set is input into the support vector machine regression, and the accuracy threshold is set to , when the support vector machine regression accuracy is less than When , the iteration is stopped to obtain a trained support vector machine regression; the normalized test set is then input into the trained support vector machine regression, and the penalty factor and kernel function parameters are optimized using the Sky Eagle optimization algorithm until the trained support vector machine regression converges to obtain a support vector machine regression model. The specific steps are as follows:

[0041] S321, set the fitness function to the mean absolute error of the trained support vector machine regression. When the Eagle is identifying prey, set the current number of iterations to e , the maximum number of iterations is E , Tianying Chapter e The position at the iteration is , No.e The prey position at the iteration is , No. e The mean position at iterations is , represents a random number and , Zetian Eagle e +1 iteration position The calculation formula is as follows,

[0042] ;

[0043] After the eagle finds its prey, it prepares to attack by spiral flight. At this time, it continuously searches for the best fitness function value and sets the flight distribution function as f , No. e The random position at the iteration is , the spiral flight parameters are and , represents a random number and , Sky Eagle Chapter e +1 update position at iteration time to get the first updated position , the calculation formula is as follows,

[0044] ;

[0045] S322, the eagle is in the process of hunting prey, flying low and descending slowly. e The prey position at iteration That is e The best fitness function value at the iteration is set to adjust the parameter and , the space dimension of the Eagle is , the lower bound of the space dimension of the Eagle is , represents a random number and , Sky Eagle Chapter e +1 iteration updates the position again to get the second updated position , find the e The best fitness function value at +1 iteration is calculated as follows:

[0046] ;

[0047] The Eagle will attack its prey randomly. e The prey position at iteration , at this time, the optimal fitness function value is continuously updated, and the search strategy quality function is set to , represents a random number and , It represents the movement of the eagle as its prey escapes. It represents the flight slope of the eagle from the initial hunting position to the final hunting position during the prey's escape. e The position of the Eagle at the iteration , Tianying Chapter e The third updated position at +1 iteration , the calculation formula is as follows,

[0048] ,

[0049] ;

[0050] The eagle continuously iterates its own position. When the eagle reaches the final hunting position, the hunting is successful and the iteration stops. The coordinate values ​​corresponding to the rectangular coordinates of the final hunting position are the penalty factor and the kernel function parameter respectively.

[0051] S33. Obtain steam quantity parameters during the operation of the steam turbine unit in previous years and record them as a current steam parameter set. After data cleaning and normalization, the current steam parameter set is processed to obtain a current normalized steam parameter set. The current normalized steam parameter set is input into a support vector machine regression model, and the predicted steam consumption of the steam turbine unit is output. The predicted steam consumption of the steam turbine unit is recorded as a heat rate evaluation parameter of the second steam turbine unit.

[0052] The invention uses the steam quantity parameters during the operation of the steam turbine unit, cleans and normalizes the steam quantity parameters during the operation of the steam turbine unit, trains the support vector machine regression to obtain the support vector machine regression model, and optimizes the penalty factor and kernel function parameters of the support vector machine regression model based on the Eagle optimization algorithm, and outputs the predicted steam consumption of the steam turbine unit; the support vector machine regression model can effectively reduce errors and improve prediction accuracy, and the Eagle optimization algorithm has a fast convergence speed and a strong optimization ability by imitating the behavior of the Eagle during the hunting process.

[0053] Preferably, S4 comprises the following steps:

[0054] S41, setting the weight coefficient of the heat rate evaluation parameter of the first steam turbine unit to , the weight coefficient of the heat rate evaluation parameter of the second steam turbine unit is If the tightness of the steam turbine unit meets the requirements, then ,otherwise ; Set the consumption threshold to , if the predicted steam consumption of the steam turbine unit is less than When =1, otherwise ; Set the evaluation threshold to 2, the steam turbine unit heat rate evaluation coefficient When the heat rate evaluation coefficient of the steam turbine unit is equal to 2, the heat rate of the steam turbine unit meets the requirements, otherwise the heat rate of the steam turbine unit does not meet the requirements.

[0055] This embodiment also discloses a system for a method for evaluating the heat rate of a steam turbine unit based on artificial intelligence, which specifically includes: a steam turbine unit heat rate correction module, a steam turbine unit leakage identification module, a steam turbine unit steam consumption prediction module, and a steam turbine unit heat rate evaluation module;

[0056] The steam turbine unit heat rate correction module is used to calculate the steam turbine unit heat rate using the preprocessed steam turbine unit heat rate parameters and perform correction;

[0057] The steam turbine unit leakage identification module is used to train the BP neural network through the operating parameters of the steam turbine unit to realize the identification of leakage data;

[0058] The steam consumption prediction of the steam turbine unit is used to train a support vector machine regression model using steam quantity parameters during the operation of the steam turbine unit, and to optimize the parameters using the Sky Eagle optimization algorithm;

[0059] The steam turbine unit heat rate evaluation module is used to evaluate the steam turbine unit heat rate in combination with the steam turbine unit heat rate evaluation parameters.

[0060] The present invention has the following beneficial effects:

[0061] 1. The invention first obtains the parameters of the heat rate of the steam turbine unit, deletes the abnormal values ​​and fills the data on the parameters of the heat rate of the steam turbine unit, then calculates the heat rate of the steam turbine unit, and introduces a correction factor to correct the heat rate of the steam turbine unit. The corrected heat rate of the steam turbine unit can objectively reflect the energy consumption level of the steam turbine unit.

[0062] 2. The invention obtains the operating parameters of the steam turbine unit, performs fuzzy C-means clustering processing, continuously iterates the membership, classifies the normal state and leakage state data, and trains the BP neural network to obtain the final BP neural network, thereby realizing the recognition of leakage data and judging the tightness of the steam turbine unit. The neural network has a high recognition rate and a fast learning speed.

[0063] 3. The invention uses the steam volume parameters of the steam turbine unit during operation to train the support vector machine regression to obtain the support vector machine regression model, and optimizes the penalty factor and kernel function parameters of the support vector machine regression model based on the Eagle optimization algorithm to output the predicted steam consumption of the steam turbine unit; the model can effectively reduce errors and improve prediction accuracy. The algorithm imitates the behavior of the Eagle during hunting, with fast convergence speed and strong optimization ability; finally, the heat consumption rate of the steam turbine unit is evaluated.

[0064] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0066] Figure 1 The present invention provides a schematic flow chart of a steam turbine unit heat rate evaluation system based on artificial intelligence for evaluating the heat rate of a steam turbine unit. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0069] This embodiment discloses a method for evaluating the heat rate of a steam turbine unit based on artificial intelligence, which specifically includes the following contents:

[0070] S1. Acquire parameters of the heat rate of the steam turbine unit, preprocess the parameters of the heat rate of the steam turbine unit, calculate the heat rate of the steam turbine unit, and perform data correction on the heat rate of the steam turbine unit to obtain a corrected heat rate of the steam turbine unit;

[0071] The S1 comprises the following steps:

[0072] S11. Obtaining parameters of the heat rate of the steam turbine unit, wherein the parameters of the heat rate of the steam turbine unit include generator power, main steam, main feed water, reheated steam, high-pressure cylinder exhaust, superheater cooling, and reheater cooling. Generating a generator power set according to the parameter data of the heat rate of the steam turbine unit , Main steam assembly , main water supply assembly , Reheat steam assembly , High pressure cylinder exhaust assembly 、Superheater desuperheating assembly and reheater desuperheater assembly , we get the steam turbine unit heat rate parameter set, recorded as ;

[0073] The steam turbine unit heat rate parameter set is preprocessed, and the specific process is as follows:

[0074] S111, performing outlier deletion processing on the steam turbine unit heat rate parameters in the steam turbine unit heat rate parameter set, dividing the steam turbine unit heat rate parameters in the steam turbine unit heat rate parameter set into 4 equal data parts, respectively recorded as the first divided data set, the second divided data set, the third divided data set and the fourth divided data set, calculating the average value of the heat rate parameter of the rightmost steam turbine unit in the first divided data set and the heat rate parameter of the leftmost steam turbine unit in the second divided data set, and recording it as the upper quantile; calculating the average value of the heat rate parameter of the rightmost steam turbine unit in the third divided data set and the heat rate parameter of the leftmost steam turbine unit in the fourth divided data set, and recording it as the lower quantile, then the parameter range calculation formula is as follows,

[0075] ,

[0076] in, Indicates that the parameters are extremely poor. represents the upper quantile, represents the lower quantile;

[0077] When the steam turbine unit heat rate parameter in the steam turbine unit heat rate parameter set is greater than Or the steam turbine unit heat rate parameter in the steam turbine unit heat rate parameter set is less than When , the steam turbine unit heat rate parameter in the steam turbine unit heat rate parameter set is an abnormal value, the abnormal value is deleted, and the steam turbine unit heat rate parameter set after the abnormal value is deleted is obtained;

[0078] S112, filling data with the heat rate parameters of the steam turbine unit in the heat rate parameter set of the steam turbine unit after the outliers are deleted, initializing the heat rate parameter set of the steam turbine unit after the outliers are deleted, constructing a complete value data matrix, calculating the Euclidean distance between the heat rate parameters of the steam turbine unit in the heat rate parameter set of the steam turbine unit after the outliers are deleted and the data in the complete value data matrix, and obtaining a distance set; in the distance set, selecting k distance data with the smallest distance value as k nearest neighbor data, calculating the weights of the k nearest neighbor data, estimating the missing values ​​using the weights of the k nearest neighbor data, and filling the missing values ​​in the heat rate parameter set of the steam turbine unit after the outliers are deleted; repeating S112 until there are no missing values ​​in the heat rate parameter set of the steam turbine unit after the outliers are deleted, and obtaining a processed heat rate parameter set of the steam turbine unit;

[0079] S12, setting the steam turbine unit heat rate parameter to be selected from the processed steam turbine unit heat rate parameter set, the generator power is , the main steam flow is , the main steam enthalpy is , the main water flow rate is , the main feed water enthalpy is , the reheat steam flow rate is , the reheat steam enthalpy is , the high pressure cylinder exhaust flow rate is , the exhaust enthalpy of the high pressure cylinder is , the superheater desuperheating flow rate is , the superheater cooling enthalpy is , the reheater desuperheating flow rate is , the reheater de-temperature enthalpy is , then the heat rate of the steam turbine unit is B The calculation formula is as follows,

[0080] ;

[0081] Set the correction factor for the heat rate of the steam turbine unit to form a correction factor set ,in Indicates the main steam correction factor, Indicates the main feed water correction factor, Reheat steam correction factor, Indicates the high pressure cylinder exhaust correction factor, Indicates the superheater desuperheating correction factor, represents the reheater temperature reduction correction factor; the weight corresponding to the correction factor in the correction factor set is recorded as the weight set, and the comprehensive correction factor is obtained by the weight set and the correction factor set. The calculation formula of the corrected steam turbine unit heat rate is as follows:

[0082] ,

[0083] in, represents the corrected heat rate of the steam turbine unit, represents the comprehensive correction factor;

[0084] S2. Obtain the operating parameters of the steam turbine unit in previous years, perform clustering processing using the fuzzy C-means clustering algorithm to obtain a processed set of steam turbine unit operating parameters, train a BP neural network in combination with the corrected steam turbine unit heat rate and the processed set of steam turbine unit operating parameters to identify leakage data, and record the tightness of the steam turbine unit as the first steam turbine unit heat rate evaluation parameter;

[0085] The S2 comprises the following steps:

[0086] S21. Obtain the operating parameters of the steam turbine units in previous years to obtain a set of operating parameters of the steam turbine units to be processed, recorded as ,in Indicates a The steam turbine unit operating parameters to be processed are then divided into a steam turbine unit operating load range, a steam turbine unit medium load range and a steam turbine unit high load range;

[0087] Set the number of cluster centers of the turbine unit operating parameter set to be processed in the low load interval of the turbine unit to be , the weighted index is , the membership degree is ; The number of cluster centers of the set of operating parameters of the steam turbine unit to be processed in the load interval of the steam turbine unit is , the weighted index is , the membership degree is ; The number of cluster centers of the set of steam turbine unit operating parameters to be processed in the high load interval of the steam turbine unit is , the weighted index is , the membership degree is ; Calculate the cluster center of the set of operating parameters of the steam turbine unit to be processed in the low load range of the steam turbine unit , get the first cluster center set; calculate the cluster center of the steam turbine unit operating parameter set to be processed in the load interval of the steam turbine unit , obtain the second cluster center set; calculate the cluster center of the set of steam turbine unit operating parameters to be processed in the high load interval of the steam turbine unit , get the third cluster center set; the calculation formula is as follows,

[0088] , , ,

[0089] in, Indicates that the set of steam turbine unit operating parameters to be processed is the first in the low load range of the steam turbine unit. The membership degree corresponding to the operating parameters of each steam turbine unit is Indicates that the set of steam turbine unit operating parameters to be processed is the first in the load range of the steam turbine unit. The membership degree corresponding to the operating parameters of each steam turbine unit is Indicates that the set of steam turbine unit operating parameters to be processed is in the high load range of the steam turbine unit. The membership degree corresponding to the operating parameters of each steam turbine unit is Indicates the first The operating parameters of the steam turbine units are: Indicates the first The operating parameters of the steam turbine units are: Indicates the first The operating parameters of the steam turbine units are: , , ;

[0090] S22, set the minimization objective function of the set of operating parameters of the steam turbine unit to be processed as , respectively obtain the first minimization objective function value, the second minimization objective function value and the third minimization objective function value; calculate the weighted sum of the first cluster center set and the first minimization objective function value, recorded as the first weighted sum, and set the first threshold to , when the first weighted sum is less than , clustering is completed, and the first steam turbine unit operating parameter set is obtained, otherwise the membership degree is reselected and S21 is repeated until the first weighted sum is less than ; Calculate the weighted sum of the second cluster center set and the second minimization objective function value, recorded as the second weighted sum, and set the second threshold to , when the second weighted sum is less than , clustering is completed, and the second steam turbine unit operating parameter set is obtained, otherwise the membership degree is reselected and S21 is repeated until the second weighted sum is less than ; Calculate the weighted sum of the third cluster center set and the third minimization objective function value, recorded as the third weighted sum, and set the third threshold to , when the third weighted sum is less than , clustering is completed, and the third steam turbine unit operating parameter set is obtained, otherwise reselect the membership degree and repeat S21 until the third weighted sum is less than ;

[0091] According to the first steam turbine unit operating parameter set, the second steam turbine unit operating parameter set and the third steam turbine unit operating parameter set, a processed steam turbine unit operating parameter set is formed; the number of BP neural network input nodes is set to , the number of hidden nodes is , the learning rate is , the processed steam turbine unit operating parameter set is divided into a parameter training set and a parameter test set, the maximum number of iterations is set to C, the current number of iterations is set to c, the parameter training set is input into the BP neural network, and iterates continuously. When the current number of iterations is equal to the maximum number of iterations, the iteration is stopped to obtain a trained BP neural network; then the parameter test set is input into the trained BP neural network, and the error threshold is set to , when the error of the trained BP neural network is less than , get the final BP neural network, otherwise adjust the weights until the error of the trained BP neural network is less than ;

[0092] S23, obtaining the operating parameters of the current steam turbine unit, combining the corrected steam turbine unit heat rate, obtaining the current steam turbine unit operating parameter set, inputting the current steam turbine unit operating parameter set into the final BP neural network, and the final BP neural network outputs the recognition result; if the BP neural network outputs 10, then the current steam turbine unit operating parameter set has leakage data, that is, the steam turbine unit tightness does not meet the requirements; if the BP neural network outputs 01, then the current steam turbine unit operating parameter set does not have leakage data, that is, the steam turbine unit tightness meets the requirements; the steam turbine unit tightness is recorded as the first steam turbine unit heat rate evaluation parameter;

[0093] S3. Use the steam volume parameters during the operation of the steam turbine unit in previous years to train the support vector machine regression to obtain a support vector machine regression model, and optimize the support vector machine regression model parameters based on the Tianying optimization algorithm to obtain the predicted steam consumption of the steam turbine unit, and record the predicted steam consumption of the steam turbine unit as the heat rate evaluation parameter of the second steam turbine unit;

[0094] The S3 comprises the following steps:

[0095] S31, obtaining steam quantity parameters during the operation of the steam turbine unit in previous years, the steam quantity parameters including boiler bed pressure, boiler bed temperature, feed water quantity, combustion feed quantity, primary air feed quantity, secondary air feed quantity, furnace temperature and furnace pressure, and generating a steam parameter set, the steam parameters in the steam parameter set include j data elements; calculate the average value of the steam parameters in the steam parameter set and standard deviation , the data elements of the steam parameters in the steam parameter set are less than The data elements are replaced by , the data elements of the steam parameters in the steam parameter set are greater than The data elements are replaced by , obtain the steam parameter set after cleaning;

[0096] The cleaned steam parameter set is normalized to obtain a normalized steam parameter set. The calculation formula is as follows:

[0097] ,

[0098] in, Represents the first in the normalized steam parameter set i The steam parameters j data elements, Indicates the first i The steam parameters j data elements, Indicates the first i The minimum data element for a steam parameter, Indicates the first i The largest data element of a steam parameter, i =1,2,3,...,8;

[0099] S32. Set the penalty factor for support vector machine regression to , the kernel function parameters are , the normalized steam parameter set is divided into a normalized training set and a normalized test set, the normalized training set is input into the support vector machine regression, and the accuracy threshold is set to , when the support vector machine regression accuracy is less than When , the iteration is stopped to obtain a trained support vector machine regression; the normalized test set is then input into the trained support vector machine regression, and the penalty factor and kernel function parameters are optimized using the Sky Eagle optimization algorithm until the trained support vector machine regression converges to obtain a support vector machine regression model. The specific steps are as follows:

[0100] S321, set the fitness function to the mean absolute error of the trained support vector machine regression. When the Eagle is identifying prey, set the current number of iterations to e , the maximum number of iterations is E , Sky Eagle Chapter e The position at the iteration is , No. e The prey position at the iteration is , No. e The mean position at iterations is , represents a random number and , Zetian Eagle e +1 iteration position The calculation formula is as follows,

[0101] ;

[0102] After the eagle finds its prey, it prepares to attack by spiral flight. At this time, it continuously searches for the best fitness function value and sets the flight distribution function as f , No. e The random position at the iteration is , the spiral flight parameters are and , represents a random number and , Tianying Chapter e +1 update position at iteration time to get the first updated position , the calculation formula is as follows,

[0103] ;

[0104] S322, the eagle is in the process of hunting prey, flying low and descending slowly. e The prey position at iteration That is e The best fitness function value at the iteration is set to adjust the parameter and , the space dimension of the Eagle is , the lower bound of the space dimension of the Eagle is , represents a random number and , Tianying Chapter e +1 iteration updates the position again to get the second updated position , find the e The best fitness function value at +1 iteration is calculated as follows:

[0105] ;

[0106] The Eagle will attack its prey randomly. e The prey position at iteration , at this time, the optimal fitness function value is continuously updated, and the search strategy quality function is set to , represents a random number and , It represents the movement of the eagle as its prey escapes. It represents the flight slope of the eagle from the initial hunting position to the final hunting position during the prey's escape. e The position of the Eagle at the iteration , Tianying Chapter e The third updated position at +1 iteration , the calculation formula is as follows,

[0107] ,

[0108] ;

[0109] The eagle continuously iterates its own position. When the eagle reaches the final hunting position, the hunting is successful and the iteration stops. The coordinate values ​​corresponding to the rectangular coordinates of the final hunting position are the penalty factor and the kernel function parameter respectively.

[0110] S33, obtaining steam quantity parameters during the operation of the steam turbine unit in previous years, recording them as a current steam parameter set, obtaining a current normalized steam parameter set after data cleaning and normalization processing, inputting the current normalized steam parameter set into a support vector machine regression model, outputting the predicted steam consumption of the steam turbine unit, and recording the predicted steam consumption of the steam turbine unit as a heat rate evaluation parameter of the second steam turbine unit;

[0111] S4, evaluating the heat rate of the steam turbine unit by combining the first steam turbine unit heat rate evaluation parameter and the second steam turbine unit heat rate evaluation parameter;

[0112] The S4 comprises the following steps:

[0113] S41, setting the weight coefficient of the heat rate evaluation parameter of the first steam turbine unit to , the weight coefficient of the heat rate evaluation parameter of the second steam turbine unit is If the tightness of the steam turbine unit meets the requirements, then ,otherwise ; Set the consumption threshold to , if the predicted steam consumption of the steam turbine unit is less than When =1, otherwise ; Set the evaluation threshold to 2, the steam turbine unit heat rate evaluation coefficient When the heat rate evaluation coefficient of the steam turbine unit is equal to 2, the heat rate of the steam turbine unit meets the requirements, otherwise the heat rate of the steam turbine unit does not meet the requirements.

[0114] This embodiment also discloses a steam turbine unit heat rate evaluation system based on artificial intelligence, which specifically includes: a steam turbine unit heat rate correction module, a steam turbine unit leakage identification module, a steam turbine unit steam consumption prediction module and a steam turbine unit heat rate evaluation module;

[0115] The steam turbine unit heat rate correction module is used to calculate the steam turbine unit heat rate using the preprocessed steam turbine unit heat rate parameters and perform correction;

[0116] The steam turbine unit leakage identification module is used to train the BP neural network through the operating parameters of the steam turbine unit to realize the identification of leakage data;

[0117] The steam consumption prediction of the steam turbine unit is used to train a support vector machine regression model using steam quantity parameters during the operation of the steam turbine unit, and to optimize the parameters using the Sky Eagle optimization algorithm;

[0118] The steam turbine unit heat rate evaluation module is used to evaluate the steam turbine unit heat rate in combination with the steam turbine unit heat rate evaluation parameters.

[0119] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0120] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A method for evaluating heat rate of a steam turbine unit based on artificial intelligence, characterized in that: The steps include: S1. Acquire parameters of the heat rate of the steam turbine unit, preprocess the parameters of the heat rate of the steam turbine unit, calculate the heat rate of the steam turbine unit, and perform data correction on the heat rate of the steam turbine unit to obtain a corrected heat rate of the steam turbine unit; S2. Obtain the operating parameters of the steam turbine unit in previous years, perform clustering processing using the fuzzy C-means clustering algorithm to obtain a processed set of steam turbine unit operating parameters, train a BP neural network in combination with the corrected steam turbine unit heat rate and the processed set of steam turbine unit operating parameters to identify leakage data, and record the tightness of the steam turbine unit as the first steam turbine unit heat rate evaluation parameter; S3, using the steam volume parameters during the operation of the steam turbine unit in previous years to train the support vector machine regression to obtain a support vector machine regression model, and optimizing the support vector machine regression model parameters to obtain the predicted steam consumption of the steam turbine unit, and recording the predicted steam consumption of the steam turbine unit as the heat rate evaluation parameter of the second steam turbine unit; S4, evaluating the heat rate of the steam turbine unit by combining the first steam turbine unit heat rate evaluation parameter and the second steam turbine unit heat rate evaluation parameter; S41, setting the weight coefficient of the heat rate evaluation parameter of the first steam turbine unit to , the weight coefficient of the heat rate evaluation parameter of the second steam turbine unit is If the tightness of the steam turbine unit meets the requirements, then ,otherwise ; Set the consumption threshold to , if the predicted steam consumption of the steam turbine unit is less than When =1, otherwise ; Set the evaluation threshold to 2, the steam turbine unit heat rate evaluation coefficient When the heat rate evaluation coefficient of the steam turbine unit is equal to 2, the heat rate of the steam turbine unit meets the requirements, otherwise the heat rate of the steam turbine unit does not meet the requirements.

2. The method for evaluating heat rate of a steam turbine unit based on artificial intelligence according to claim 1, characterized in that: The S1 comprises the following steps: S11, obtaining a heat rate parameter of a steam turbine unit, obtaining a heat rate parameter set of the steam turbine unit, and preprocessing the heat rate parameter set of the steam turbine unit to obtain a processed heat rate parameter set of the steam turbine unit; S12. Calculate the heat rate of the steam turbine unit according to the processed heat rate parameter set of the steam turbine unit, and correct the heat rate of the steam turbine unit to obtain a corrected heat rate of the steam turbine unit.

3. The method for evaluating heat rate of a steam turbine unit based on artificial intelligence according to claim 2, characterized in that: The S11 comprises the following steps: S111, performing outlier deletion processing on the steam turbine unit heat rate parameters in the steam turbine unit heat rate parameter set to obtain a steam turbine unit heat rate parameter set after the outlier values ​​are deleted; S112, performing data filling on the steam turbine unit heat rate parameters in the steam turbine unit heat rate parameter set after the abnormal values ​​are deleted, to obtain a processed steam turbine unit heat rate parameter set.

4. The method for evaluating heat rate of a steam turbine unit based on artificial intelligence according to claim 3 is characterized in that: The S2 comprises the following steps: S21, obtaining the operating parameters of the steam turbine unit in previous years, generating a set of steam turbine unit operating parameters to be processed, and performing clustering processing on the set of steam turbine unit operating parameters to be processed using a fuzzy C-means clustering algorithm to obtain a processed set of steam turbine unit operating parameters; S22, set the number of BP neural network input nodes to , the number of hidden nodes is , the learning rate is , dividing the processed steam turbine unit operating parameter set into a parameter training set and a parameter test set, setting the maximum number of iterations to C and the current number of iterations to c, inputting the parameter training set into the BP neural network, iterating continuously, and stopping the iteration when the current number of iterations is equal to the maximum number of iterations, to obtain a trained BP neural network; Then the parameter test set is input into the trained BP neural network, and the error threshold is set to , when the error of the trained BP neural network is less than , get the final BP neural network, otherwise adjust the weights until the error of the trained BP neural network is less than .

5. The method for evaluating heat rate of a steam turbine unit based on artificial intelligence according to claim 4, characterized in that: The operating parameters of the current steam turbine unit are obtained, and combined with the corrected steam turbine unit heat rate, the current steam turbine unit operating parameter set is obtained, and the current steam turbine unit operating parameter set is input into the final BP neural network, and the final BP neural network outputs a recognition result; if the BP neural network outputs 10, then there is leakage data in the current steam turbine unit operating parameter set, that is, the tightness of the steam turbine unit does not meet the requirements; if the BP neural network outputs 01, there is no leakage data in the current steam turbine unit operating parameter set, that is, the tightness of the steam turbine unit meets the requirements; the steam turbine unit tightness is recorded as the first steam turbine unit heat rate evaluation parameter.

6. The method for evaluating heat rate of a steam turbine unit based on artificial intelligence according to claim 5, characterized in that: The S3 comprises the following steps: S31, obtaining steam quantity parameters during the operation of the steam turbine unit in previous years, generating a steam parameter set, and performing data cleaning and normalization processing on the steam parameter set to obtain a normalized steam parameter set; S32. Set the penalty factor for support vector machine regression to , the kernel function parameters are , the normalized steam parameter set is divided into a normalized training set and a normalized test set, the normalized training set is input into the support vector machine regression, and the accuracy threshold is set to , when the support vector machine regression accuracy is less than When , the iteration is stopped to obtain a trained support vector machine regression; the normalized test set is then input into the trained support vector machine regression, and the penalty factor and kernel function parameters are optimized using the Sky Eagle optimization algorithm until the trained support vector machine regression converges to obtain a support vector machine regression model; S33. Obtain steam quantity parameters during the operation of the steam turbine unit in previous years and record them as a current steam parameter set. After data cleaning and normalization, the current steam parameter set is processed to obtain a current normalized steam parameter set. The current normalized steam parameter set is input into a support vector machine regression model, and the predicted steam consumption of the steam turbine unit is output. The predicted steam consumption of the steam turbine unit is recorded as a heat rate evaluation parameter of the second steam turbine unit.

7. The method for evaluating heat rate of a steam turbine unit based on artificial intelligence according to claim 6, characterized in that: The S32 comprises the following steps: S321, set the fitness function to the mean absolute error of the trained support vector machine regression. When the Eagle is identifying prey, set the current number of iterations to e , the maximum number of iterations is E , Sky Eagle Chapter e The position at the iteration is , No. e The prey position at the iteration is , No. e The mean position at iterations is , represents a random number and , Zetian Eagle e +1 iteration position The calculation formula is as follows, ； After the eagle finds its prey, it prepares to attack by spiral flight. At this time, it continuously searches for the best fitness function value and sets the flight distribution function as f , No. e The random position at the iteration is , the spiral flight parameters are and , represents a random number and , Sky Eagle Chapter e +1 update position at iteration time to get the first updated position , the calculation formula is as follows, ； S322, the eagle is in the process of hunting prey, flying low and descending slowly. e The prey position at iteration That is e The best fitness function value at the iteration is set to adjust the parameter and , the space dimension of the Eagle is , the lower bound of the space dimension of the Eagle is , represents a random number and , Sky Eagle Chapter e +1 iteration updates the position again to get the second updated position , find the e The best fitness function value at +1 iteration is calculated as follows: ； The Eagle will attack its prey randomly. e The prey position at iteration , at this time, the optimal fitness function value is continuously updated, and the search strategy quality function is set to , represents a random number and , It represents the movement of the eagle as its prey escapes. It represents the flight slope of the eagle from the initial hunting position to the final hunting position during the prey's escape. e The position of the Eagle at the iteration , Sky Eagle Chapter e The third updated position at +1 iteration , the calculation formula is as follows, ， ； The eagle continuously iterates its own position. When the eagle reaches the final hunting position, the hunting is successful and the iteration stops. The coordinate values ​​corresponding to the rectangular coordinates of the final hunting position are the penalty factor and the kernel function parameter respectively.

8. A system for implementing the method for evaluating heat rate of a steam turbine unit based on artificial intelligence as claimed in any one of claims 1 to 7, characterized in that: Specifically include: Steam turbine unit heat rate correction module, steam turbine unit leakage identification module, steam turbine unit steam consumption prediction and steam turbine unit heat rate assessment module; The steam turbine unit heat rate correction module is used to calculate the steam turbine unit heat rate using the preprocessed steam turbine unit heat rate parameters and perform correction; The steam turbine unit leakage identification module is used to train the BP neural network through the operating parameters of the steam turbine unit to realize the identification of leakage data; The steam consumption prediction of the steam turbine unit is used to train a support vector machine regression model using steam quantity parameters during the operation of the steam turbine unit, and to optimize the parameters using the Sky Eagle optimization algorithm; The steam turbine unit heat rate evaluation module is used to evaluate the steam turbine unit heat rate in combination with the steam turbine unit heat rate evaluation parameters.

Images