A power plant integrated operation control system

Abstract

The application discloses a power plant comprehensive operation control system, comprising: a data acquisition module, which is used for collecting real-time operation data and historical normal operation data of various data categories of the power plant; a preprocessing module, which is used for preprocessing the real-time operation data and the historical normal operation data to obtain a prediction data set and a training data set respectively; a training module, which is used for receiving the training data set and performing iterative training through a CNN-LSTM model; a prediction analysis module, which is used for predicting the prediction data set by using the trained CNN-LSTM model, and obtaining prediction values of various data categories by subtracting real-time operation data from the prediction values, obtaining operation residuals of various data categories, comparing the operation residuals with corresponding preset residual thresholds, and outputting prediction results; and a strategy control module, which is used for formulating corresponding control strategies according to the prediction results, and controlling the power plant to operate according to the corresponding control strategies; automatic operation control of the power plant is realized, future operation states can be predicted, and pre-control can be performed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent operation, and particularly relates to a power plant comprehensive operation control system. BACKGROUND

[0002] With the continuous expansion of the power grid scale and the continuous increase of power system operation equipment, the complexity of power system operation also increases, and the safe and stable operation of the power system becomes more and more important. Among them, the reliable operation of the power plant has a significant impact on the stability and reliability of the power grid.

[0003] At present, the operation of the power plant mainly relies on real-time regulation and control by the operator. The operator judges whether the power plant is in a normal operation state according to the real-time data of each load and each data category of the power plant, and then manually adjusts the corresponding strategy. This operation control method requires a large amount of manpower, has poor reliability, low automation degree, and cannot predict the operation trend of the power plant in the future, the data has a lag, and cannot be pre-controlled. SUMMARY

[0004] The present application aims to at least solve one of the technical problems existing in the prior art. To this end, the purpose of the present application is to provide a power plant comprehensive operation control system to solve the problems of non-automatic and intelligent operation of the power plant, lag and the like in the prior art.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] The present application provides a power plant comprehensive operation control system, comprising:

[0007] a data acquisition module for acquiring real-time operation data and historical normal operation data of each data category of the power plant;

[0008] a preprocessing module for preprocessing the real-time operation data and the historical normal operation data to obtain a prediction data set and a training data set, respectively;

[0009] a training module for receiving the training data set and iteratively training through a CNN-LSTM model;

[0010] a prediction analysis module for predicting the prediction data set using the trained CNN-LSTM model, and obtaining the prediction value of each data category by differencing the real-time operation data, obtaining the operation residual of each data category, comparing the operation residual with the corresponding preset residual threshold, and outputting the prediction result;

[0011] a strategy control module for formulating a corresponding control strategy according to the prediction result and controlling the power plant to operate according to the corresponding control strategy.

[0012] In some embodiments, the data collection module collects data categories including, but not limited to, one or more of coal consumption, ambient temperature, main steam pressure, reheat steam pressure, flue gas temperature, main and reheat temperature, feedwater temperature, and oxygen content.

[0013] In some embodiments, the preprocessing module is configured to delete null values and error data in the real-time operation data and historical normal operation data, extract valid values, and build respective valid values into a prediction data set and a training data set.

[0014] In some embodiments, the prediction analysis module includes an input layer, a CNN layer, an LSTM network layer, a self-attention layer, and an output layer.

[0015] The input layer is configured to convert the prediction data set into a matrix vector.

[0016] The CNN layer is configured to extract feature information of the prediction data set to obtain global feedback of the feature information.

[0017] The LSTM network layer is configured to learn the feature information.

[0018] The self-attention layer is configured to deeply mine information learned by the LSTM layer and optimize the matrix vector.

[0019] The output layer is configured to continue outputting after performing feature dimension reduction on the optimized matrix vector.

[0020] In some embodiments, the system further includes a data storage module configured to store the prediction data set and the training data set processed by the preprocessing module through local storage and / or cloud server storage.

[0021] In some embodiments, the system further includes an alarm module configured to display real-time operation data of each data category of the power plant and perform corresponding sound and light signal transmission according to a control strategy.

[0022] In some embodiments, the system further includes a visualization module configured to perform visualization processing on the prediction data set and the training data set, draw data curves, and display the data curves through the alarm module.

[0023] In some embodiments, the strategy control module outputs the control strategy according to a dynamic adjustment period.

[0024] Compared with the prior art, the present application has at least the following beneficial effects:

[0025] The power plant comprehensive operation control system provided by the embodiment of the present application carries out training and prediction based on a CNN-LSTM model for each data category of the power plant, calculates the operation residual between the predicted value and the real-time operation data, compares the operation residual with the corresponding preset residual threshold, judges the operation state of each data category, dynamically adjusts the control strategy, realizes automatic operation control of the power plant, and can predict the future operation state and pre-control in advance.

[0026] The present application will be further described below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0027] The present application will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0028] Figure 1 It is a schematic diagram of the framework of another embodiment of the power plant comprehensive operation control system.

[0029] Figure 2 It is a schematic diagram of the framework of another embodiment of the power plant comprehensive operation control system. DETAILED DESCRIPTION

[0030] The technical solutions of the present application will be described below in conjunction with the accompanying drawings, obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by the ordinary skilled in the art without making creative efforts are within the protection scope of the present application.

[0031] In the description of the present application, it should be noted that the orientations or positional relationships indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", "third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0032] In the description of the present application, when a particular device is described to be located between a first device and a second device, there can be an intervening device between the particular device and the first device or the second device, or there can be no intervening device. When a particular device is described to be connected to other devices, the particular device can be directly connected to the other devices without an intervening device, or can not be directly connected to the other devices with an intervening device.

[0033] Techniques, methods, and devices known to those of ordinary skill in the relevant art can not be discussed in detail herein, but should be considered part of the specification.

[0034] With reference to Figure 1 The embodiment provides a power plant comprehensive operation control system, comprising:

[0035] A data acquisition module is configured to acquire real-time operation data and historical normal operation data of each data category of the power plant; each different data category has corresponding real-time operation data and historical normal operation data;

[0036] A preprocessing module is configured to preprocess the real-time operation data and the historical normal operation data to obtain a prediction data set and a training data set, respectively; in the prediction data set, the real-time operation data of each data category after preprocessing is collected, and in the training data set, the historical normal operation data of each data category after preprocessing is collected;

[0037] A training module is configured to receive the training data set and perform iterative training through a CNN-LSTM model; in order to predict values of N time steps, a traditional CNN-LSTM model is based on a predicted value of a previous step for each prediction. In the embodiment, the actual value is used as an input of a next step in each prediction at a time t∈(0,T) in a training stage, and a neuron state is updated, a gradient propagation of an error is reduced, and the remaining prediction is recycled;

[0038] A prediction analysis module is configured to predict the prediction data set through the trained CNN-LSTM model, predict a predicted value of each data category, and obtain a residual error of each data category by subtracting the real-time operation data from the predicted value of each data category; the residual error is compared with a corresponding preset residual error threshold, and a prediction result is output;

[0039] A strategy control module is configured to formulate a corresponding control strategy according to the prediction result, and control the power plant to operate according to the corresponding control strategy.

[0040] It should be noted that in order to improve the intelligentization and automation of power plant operation control, in addition to realizing real-time regulation and control, it is also necessary to predict the predicted value after a certain time step, predict the future operation trend of the power plant, and realize early control. More specifically, the data acquisition module is arranged on multiple loads of the power plant, and different data categories are collected, such as one or more of coal consumption, ambient temperature, main steam pressure, reheat steam pressure, exhaust gas temperature, main and reheat temperature, feed water temperature, and oxygen content. Each data category has real-time running data under the current running state and historical normal running data of normal running at past time. After preprocessing, a prediction data set and a training data set are respectively established, wherein the expression of the prediction data set is X = [x1, x2,..., xn], and x represents different data categories, and there are n categories in total. Since the historical normal running data is a kind of time series data, there are multiple training subsets in the training data set. The expression of each training subset is similar to the above. Multiple training subsets form a space-time feature matrix. The CNN-LSTM model is trained using multiple training subsets for iterative fitting processing. Further, the trained multiple model parameters are input into the genetic algorithm as initial population, and the genetic algorithm is run to select the optimal model parameters. Then, the prediction data set is input into the trained CNN-LSTM model for prediction of each data category, and the predicted value is compared with the real-time running data to obtain the running residual of each data category. The running residual is compared with the corresponding preset residual threshold, and according to the difference value of the comparison, the prediction result is output. If the difference value is large and exceeds the preset range, it is determined that the operation of the power plant deviates from the normal range, and the prediction result is an abnormal situation. Conversely, if the difference value is small and within the preset range, it is determined that the operation of the power plant is within the normal range, and the prediction result is a normal situation. If it is determined to be an abnormal situation, the specific running residual exceeding range data category is located, and a corresponding control strategy is developed according to this data category to adjust the data category back to the normal range and control the power plant to operate according to the corresponding control strategy. n ] T n The expression of the prediction data set is X = [x1, x2,..., xn], and x represents different data categories, and there are n categories in total. Since the historical normal running data is a kind of time series data, there are multiple training subsets in the training data set. The expression of each training subset is similar to the above. Multiple training subsets form a space-time feature matrix. The CNN-LSTM model is trained using multiple training subsets for iterative fitting processing. Further, the trained multiple model parameters are input into the genetic algorithm as initial population, and the genetic algorithm is run to select the optimal model parameters. Then, the prediction data set is input into the trained CNN-LSTM model for prediction of each data category, and the predicted value is compared with the real-time running data to obtain the running residual of each data category. The running residual is compared with the corresponding preset residual threshold, and according to the difference value of the comparison, the prediction result is output. If the difference value is large and exceeds the preset range, it is determined that the operation of the power plant deviates from the normal range, and the prediction result is an abnormal situation. Conversely, if the difference value is small and within the preset range, it is determined that the operation of the power plant is within the normal range, and the prediction result is a normal situation. If it is determined to be an abnormal situation, the specific running residual exceeding range data category is located, and a corresponding control strategy is developed according to this data category to adjust the data category back to the normal range and control the power plant to operate according to the corresponding control strategy.

[0041] As an embodiment, when the preprocessing module preprocesses, the null values and error data in the real-time running data and the historical normal running data are deleted, and the valid values are extracted. The valid values are respectively built into a prediction data set and a training data set. The Pearson correlation coefficient between the historical normal running data and the real-time running data of the power plant can be calculated, and the observation features with larger correlation degree of each data category of the power plant can be determined according to the size of the correlation coefficient.

[0042] ​Further, in the plurality of data categories, different levels of priority can be formulated, for example, a total of 10 data categories, divided into three levels of priority, the first, second and third data categories are the first priority, the fourth, fifth and sixth data categories are the second priority, and the seventh, eighth, ninth and tenth data categories are the third priority. When calculating the running residual of each data category, the data category of the first priority is determined first, and the control strategy is formulated. After all data categories of the first priority are adjusted to the normal range, the control strategy of the data category of the second priority is formulated, and the adjustment is gradually downward.

[0043] In the embodiment, the preprocessing module is used to delete the null values and error data in the real-time running data and historical normal running data, extract valid values, and build the respective valid values into a prediction data set and a training data set.

[0044] In some embodiments, the prediction analysis module includes an input layer, a CNN layer, an LSTM network layer, a self-attention layer, and an output layer.

[0045] The input layer is used to convert the prediction data set into a matrix vector.

[0046] The CNN layer is used to extract feature information of the prediction data set to obtain global feedback of the feature information. The CNN layer includes a convolution layer and a max-pooling layer. The convolution layer can extract features of different sizes of input data, and input the extracted data to the pooling layer. After the pooling processing, the data is spliced into a new vector, thereby extracting the maximum data features in the prediction data set. Finally, the CNN layer outputs a feature vector.

[0047] The LSTM network layer is used to learn the feature information. The LSTM network layer includes an input gate, an output gate, a forget gate, and a memory cell unit. The combination memory unit allows the network to learn and forget the hidden state value at the previous time, and update the hidden state information.

[0048] The self-attention layer is used to deeply mine the information learned by the LSTM layer and optimize the matrix vector. The Self-Attention has good effect in natural language processing, and is integrated into the load prediction model to optimize the matrix vector.

[0049] The output layer is used to continue output after feature dimension reduction of the optimized matrix vector. The output Y = [y1, y2,..., yn] of the prediction step n is calculated. n ] T .

[0050] As an implementation form, the control system further comprises a data storage module configured to store the prediction data set and the training data set processed by the preprocessing module through local storage and / or cloud server storage. The local storage facilitates on-site inspection of the data by the patrol personnel, and the control strategy can be adjusted in real time through a mobile terminal, and data storage is realized in the case of communication interruption. The cloud server storage can store the data of each power plant in the power grid in the cloud to realize the same operation control.

[0051] With reference to Figure 2 As an implementation form, the control system further comprises a warning module configured to display real-time operation data of each data category of the power plant, directly display the real-time operation data to an operator, and perform corresponding sound and light signal transmission according to the control strategy, for example, display warning information on the screen and emit a warning sound in the case of an abnormal situation.

[0052] Further, a visualization module is further included, configured to perform visualization processing on the prediction data set and the training data set, draw a data curve, and display the data curve through the warning module. The warning module comprises a display screen.

[0053] As an implementation form, the strategy control module outputs the control strategy according to a dynamic adjustment period. The dynamic adjustment period can be set according to actual needs. Preferably, the data acquisition period of the data acquisition module is 1 s, and the dynamic adjustment period is 30 s. After 30 groups of real-time operation data are accumulated, an effective average value is obtained through preprocessing, and a prediction data set is established. After prediction and residual error comparison, a corresponding control strategy is formulated. Each load of the power plant is adjusted every 30 s, so that the operation state of the entire power plant gradually approaches the normal range.

[0054] Compared with the prior art, the above embodiment provides a power plant comprehensive operation control system. For each data category of the power plant, training and prediction based on a CNN-LSTM model are performed, the operation residual error between the prediction value and the real-time operation data is calculated, the operation state of each data category is judged by comparing the operation residual error with a corresponding preset residual error threshold, the control strategy is dynamically adjusted, automatic operation control of the power plant is realized, and future operation states can be predicted for early pre-control.

[0055] The technical features of the above embodiments can be combined in any manner. To make the description concise, all possible combinations of the technical features in the above embodiments are not described, but as long as the combinations of the technical features do not exist, they should be considered as the scope of the present disclosure.

[0056] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the application. It should be noted that for ordinary skilled persons in the art, several modifications and improvements can be made without departing from the concept of the present application, which are all within the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A power plant integrated operational control system, characterized by, The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The data categories collected by the data collection module include but are not limited to one or more of coal consumption, ambient temperature, main steam pressure, reheat steam pressure, flue gas temperature, main and reheat temperature, feed water temperature, and oxygen content; each data category has real-time running data under the current running state and historical normal running data at past time points, and after preprocessing, a prediction data set and a training data set are respectively established, wherein the expression of the prediction data set is X=[x1, x2,..., x n ] T n are different data categories, and there are n categories in total. Since the historical normal running data is a kind of time series data, there are multiple training subsets in the training data set, multiple training subsets form a space-time feature matrix, a CNN-LSTM model is trained by using the multiple training subsets, iterative fitting processing is performed, multiple model parameters trained are input into a genetic algorithm as an initial population, the genetic algorithm is run, and model parameters with the optimal effect are selected; then the prediction data set is input into the trained CNN-LSTM model, prediction of each data category is performed, a prediction value is compared with real-time running data, running residuals of each data category are obtained, the running residuals are compared with corresponding preset residual thresholds, a prediction result is output according to the size of the difference value of the comparison;​ The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control.

2. A power plant integrated operational control system as claimed in claim 1, characterized in that, The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control.

3. A power plant integrated operational control system as claimed in claim 2, characterized in that, The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control.

4. A power plant integrated operational control system as claimed in claim 3, characterized in that, The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility model relates to a power plant operation state prediction and control method and system, and relates to the field of power plant operation state prediction and control. The utility

Images