Method, device and system for predicting coal consumption of thermal power plant and storage medium
By using sparse attention mechanism and data slicing processing technology in coal consumption prediction in thermal power plants, the problem of difficult to capture nonlinear relationships and high computational complexity of deep learning algorithms is solved, and efficient coal consumption prediction is achieved.
Patent Information
- Application Number
- CN202510260662.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-06
- Publication Date
- 2025-06-06
AI Technical Summary
Traditional coal daily consumption prediction methods are difficult to capture nonlinear relationships. Deep learning time series prediction algorithms significantly increase in computational volume and memory requirements when processing long sequences, resulting in inefficient prediction.
The model is trained by the sparse attention mechanism, and the calculation complexity is reduced through data slicing and normalization processing, and the prediction efficiency is improved.
The sparse attention mechanism reduces the computational complexity, improves the speed and efficiency of the model when processing long sequence data, and significantly improves the prediction efficiency of coal consumption in thermal power plants.
Smart Images
Figure CN120106299A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of energy technology, and in particular to a method, device, system and storage medium for predicting coal consumption in a thermal power plant. Background Art
[0002] In the field of coal daily consumption prediction, traditional methods such as moving average (MA) and exponential smoothing (EMA) are difficult to effectively capture the nonlinear relationship in the fluctuation of coal daily consumption. In the time series prediction algorithm based on deep learning, by combining CNN with other frameworks such as LSTM and Transformer, the prediction effect of time series can be significantly enhanced. However, when processing long sequences, the amount of calculation and memory requirements increase significantly as the sequence grows.
[0003] Therefore, how to provide a method for predicting coal consumption in thermal power plants to improve the prediction efficiency of coal consumption in thermal power plants has become a technical problem that needs to be solved urgently. Summary of the invention
[0004] The present application provides a method, device, system and storage medium for predicting coal consumption of a thermal power plant, so as to improve the prediction efficiency of coal consumption of a thermal power plant.
[0005] The present application provides a method for predicting coal consumption of a thermal power plant, comprising:
[0006] Collecting the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0007] Preprocessing the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0008] Using at least part of the preprocessed data as training data and training the model using a sparse attention mechanism;
[0009] The daily coal consumption of the future target thermal power plant is predicted based on the trained model.
[0010] The beneficial effects of the present application are as follows: the present application collects the daily coal consumption of the target thermal power plant and the preset parameters related to the daily coal consumption, and pre-processes the daily coal consumption of the target thermal power plant and the preset parameters related to the daily coal consumption; then uses at least part of the pre-processed data as training data, and uses the sparse attention mechanism to train the model, and predicts the daily coal consumption of the future target thermal power plant based on the trained model. Since the present application completes model training and prediction based on the sparse attention mechanism, by reducing unnecessary calculations, the complexity is controlled at an approximately linear level, which greatly improves the speed and efficiency of the model in processing long sequence data, thereby improving the prediction efficiency of the coal consumption of thermal power plants.
[0011] In one embodiment, the preprocessing of the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption includes:
[0012] Performing data slicing on the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0013] Normalize the data after data slicing.
[0014] In one embodiment, the data slicing of the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption includes:
[0015] Get the preset time length of data slices and the time length of each prediction;
[0016] The data is sliced according to a preset time length of the data slice and a time length of each prediction.
[0017] In one embodiment, the method further comprises:
[0018] Determine the data sliced according to the preset time length as input data;
[0019] Determine the data sliced according to the time length of each prediction as output data;
[0020] Correlating the input data and the output data accordingly to obtain a plurality of associated data groups;
[0021] The multiple associated data sets are divided into a training set and a test set, wherein the training set is used for model training and the test set is used for model testing.
[0022] In one embodiment, the preset parameters related to the daily coal consumption include daily historical coal consumption, inventory, planned power generation, and meteorological data, and the normalization of the data after data slicing includes:
[0023] Obtain daily historical coal consumption, inventory, and planned power generation, and normalize the daily historical coal consumption, inventory, and planned power generation according to a standard normalization method;
[0024] Acquire meteorological data, and perform data normalization on the meteorological data according to a maximum and minimum normalization method.
[0025] In one embodiment, using at least part of the preprocessed data as training data and using a sparse attention mechanism to train the model includes:
[0026] Merge all training data into input sequences according to time;
[0027] Encode the input sequence into an input vector through an encoder;
[0028] Calculate the attention weight for each input vector;
[0029] Input vectors whose attention weights are lower than a preset weight threshold are eliminated to perform sparse processing on the input vectors.
[0030] In one embodiment, the method further comprises:
[0031] When training the model, additional data is generated by adding noise and / or random edits to the preprocessed data;
[0032] The additional data is input into the model as training data.
[0033] The present application also provides a device for predicting coal consumption in a thermal power plant, comprising:
[0034] A collection module, used for collecting the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0035] A preprocessing module, used for preprocessing the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0036] A training module, configured to use at least part of the preprocessed data as training data and train the model using a sparse attention mechanism;
[0037] The prediction module is used to predict the daily coal consumption of the future target thermal power plant based on the trained model.
[0038] In one embodiment, the preprocessing module comprises:
[0039] A slicing submodule, used for slicing data of the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0040] The normalization submodule is used to normalize the data after data slicing.
[0041] In one embodiment, the slicing submodule is further used to:
[0042] Get the preset time length of data slices and the time length of each prediction;
[0043] The data is sliced according to a preset time length of the data slice and a time length of each prediction.
[0044] In one embodiment, the apparatus further comprises:
[0045] A determination module, used to determine data sliced according to a preset time length as input data;
[0046] The determination module is further used to determine the data sliced according to the time length of each prediction as output data;
[0047] An associating module, used to associate the input data with the output data to obtain a plurality of associated data groups;
[0048] The partitioning module is used to partition a plurality of associated data groups into a training set and a test set, wherein the training set is used for model training and the test set is used for model testing.
[0049] In one embodiment, the preset parameters related to the daily coal consumption include daily historical coal consumption, inventory, planned power generation, and meteorological data. The normalization submodule is further used to:
[0050] Obtain daily historical coal consumption, inventory, and planned power generation, and normalize the daily historical coal consumption, inventory, and planned power generation according to a standard normalization method;
[0051] Acquire meteorological data, and perform data normalization on the meteorological data according to a maximum and minimum normalization method.
[0052] In one embodiment, the training module includes:
[0053] The merging submodule is used to merge all training data into input sequences according to time;
[0054] An encoding submodule, used for encoding the input sequence into an input vector through an encoder;
[0055] The calculation submodule is used to calculate the attention weight of each input vector;
[0056] The elimination submodule is used to eliminate input vectors whose attention weights are lower than a preset weight threshold so as to perform sparse processing on the input vectors.
[0057] In one embodiment, the training module includes:
[0058] Add submodules for generating additional data by adding noise and / or random edits to preprocessed data while training the model;
[0059] The input submodule is used to input the additional data into the model as training data.
[0060] The present application also provides a coal consumption prediction system for a thermal power plant, comprising:
[0061] at least one processor; and,
[0062] a memory communicatively connected to the at least one processor; wherein,
[0063] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to implement the method for predicting coal consumption of a thermal power plant recorded in any of the above embodiments.
[0064] The present application also provides a computer-readable storage medium. When the instructions in the storage medium are executed by a processor corresponding to the thermal power plant coal consumption prediction system, the thermal power plant coal consumption prediction system can implement the thermal power plant coal consumption prediction method recorded in any of the above embodiments.
[0065] Other features and advantages of the present application will be described in the following description, and partly become apparent from the description, or be understood by practicing the present application. The purpose and other advantages of the present application can be realized and obtained by the structures specifically pointed out in the written description, claims, and drawings.
[0066] The technical solution of the present application is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] The accompanying drawings are used to provide a further understanding of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings:
[0068] Figure 1 This is a flow chart of a method for predicting coal consumption of a thermal power plant in one embodiment of the present application;
[0069] Figure 2 This is a schematic diagram of the structure of a device for predicting coal consumption in a thermal power plant in one embodiment of the present application;
[0070] Figure 3 This is a schematic diagram of the hardware structure of a coal consumption prediction system for a thermal power plant in one embodiment of the present application. DETAILED DESCRIPTION
[0071] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application and are not used to limit the present application.
[0072] Figure 1 FIG. 1 is a flow chart of a method for predicting coal consumption of a thermal power plant in one embodiment of the present application. Figure 1As shown, the method can be implemented as the following steps S101-S104:
[0073] In step S101, the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption are collected;
[0074] In step S102, the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption are preprocessed;
[0075] In step S103, at least part of the preprocessed data is used as training data, and the model is trained using a sparse attention mechanism;
[0076] In step S104, the daily coal consumption of the future target thermal power plant is predicted based on the trained model.
[0077] The daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption are collected.
[0078] The daily consumption data of thermal power plants usually has a strong time dependence, that is, the power consumption on a certain day may be strongly correlated with the consumption on the previous few days and may change with the seasons. In addition, it is related to external weather changes. For example, extreme weather (such as cold waves or heat waves) may lead to a sharp increase in electricity demand. The historical inventory level can serve as a buffer to help power plants cope with demand fluctuations. Therefore, the collected data can include the historical daily coal consumption throughout the year, planned power generation, coal consumption coefficient, inventory, and local meteorological data, such as minimum temperature, wind speed, rainfall, etc.
[0079] The daily coal consumption of the target thermal power plant and the preset parameters related to the daily coal consumption are preprocessed. First, the daily coal consumption of the target thermal power plant and the preset parameters related to the daily coal consumption are data sliced. For example, the preset time length of the data slice is 50 days, and the time length of each prediction is 15 days, which means that 50 days of data are used to predict 15 days of data. When 50 days of data are used to predict 15 days of data, the data can be sliced into: (1) 1-50 days of data, and the corresponding predicted data is 51-65 days of daily consumption data; (2) 2-51 days of input data, and the corresponding output data is 52-66 days of data... The last one is 301-350 days of input data, and the corresponding output data is 351-365 days of data; and so on. Then, the data after data slicing is normalized. For example, the preset parameters related to daily coal consumption include daily historical coal consumption, inventory, and planned power generation, and the daily historical coal consumption, inventory, and planned power generation are normalized according to the standard normalization method; for meteorological data normalization, the meteorological data is normalized according to the maximum and minimum normalization method. Determine the data sliced according to the preset time length as input data; determine the data sliced according to the time length of each prediction as output data; associate the input data and output data accordingly to obtain multiple associated data groups; divide the multiple associated data groups into training sets and test sets, wherein the training set is used for model training and the test set is used for model testing. For example, the input data can be: historical 50-day daily consumption + historical 50-day inventory + historical 50-day meteorological data (daily rainfall, daily average wind speed, daily average air pressure, daily average body temperature, daily total solar radiation) + planned power generation for the next 15 days; the output data can be: daily consumption for the next 15 days. At least part of the preprocessed data is then used as training data, and the model is trained using a sparse attention mechanism.
[0080] During training, all training data are combined into an input sequence in time; training data are often collected at different time points and may exist in discrete form. In order for the model to learn how data changes over time, we need to combine these chronologically arranged training data into a continuous input sequence. Suppose we have a series of time points t1, t2, ..., tn, corresponding to training data samples x1, x2, ..., xn, where the sample xi can be a single value or a feature vector. Combining them in chronological order forms an input sequence S = [x1, x2, ..., xn], which retains the chronological order of the data.
[0081] The input sequence is encoded into an input vector through an encoder. The role of the encoder is to convert the input sequence into a vector representation that is more suitable for subsequent processing by the model. The encoder is usually a complex structure composed of multiple neural network layers, such as a convolutional neural network (CNN) layer, a recurrent neural network (RNN) layer, or an encoder layer in a Transformer. Different types of encoders differ in structure and calculation methods, but the overall goal is to map the input sequence to a low-dimensional and representative vector space so that the model can extract and process the information therein more efficiently. For example, CNN encoders are good at extracting local features, RNN encoders have advantages in processing long-term dependencies in sequences, and Transformer encoders can better capture global information in sequences through self-attention mechanisms.
[0082] Taking the fully connected neural network as an encoder as an example, the input sequence S first enters the first layer of the neural network. The neurons in this layer perform a weighted summation on each element in the input sequence and perform a nonlinear transformation through an activation function (such as the ReLU function). This process can be expressed as:
[0083] h 1 =ReLU(W 1 S+b 1 );
[0084] Among them, W 1 is the weight matrix of the first layer of the neural network, b 1 is the bias vector, h 1 It is the output after being processed by the first layer of neural network.
[0085] Next, h 1 It will serve as the input of the second layer of the neural network, repeating the above weighted summation and nonlinear transformation process until it is processed through several layers and finally outputs a vector of fixed dimension. This vector is the input vector E.
[0086] Calculate the attention weight of each input vector; the core idea of the attention mechanism is to let the model automatically pay attention to different parts of the input data and assign different importance weights to each part. Therefore, after obtaining the input vector E, it is necessary to calculate the attention weight of each input vector.
[0087] Assume that the input vector set is {E 1 ,E 2 ,...,E m}, where m is the number of input vectors. In this step, the input vector acts as both a key vector and a value vector, and a learnable query vector Q is also introduced. Taking the dot product attention mechanism as an example, first, the dot product of the query vector Q and each input vector (as a key vector) is calculated to obtain the attention score:
[0088] e i =Q T E i (i=1,2,...,m)
[0089] Among them, e i It is the attention score of the i-th input vector, reflecting the degree of relevance between the query vector and each input vector.
[0090] Then, to convert these scores into attention weights in the form of probability distributions:
[0091]
[0092] Among them, α i It is the attention weight of the i-th input vector, which indicates the degree of attention of the model to the i-th input vector, and the sum of all attention weights is 1.
[0093] The input vectors whose attention weights are lower than the preset weight threshold are removed to perform sparse processing on the input vectors. After obtaining the attention weight of each input vector, we set a preset weight threshold θ. This threshold is a hyperparameter and needs to be adjusted according to the specific task and dataset. For each input vector Ei, if its corresponding attention weight α i is less than the threshold θ, that is, α i <θ, then this input vector is removed from the input vector set.
[0094] After this operation, the remaining set of input vectors is the result of sparse processing. The benefit of sparse processing is that it reduces the amount of data that the model needs to process and reduces the computational complexity. It also forces the model to pay more attention to those input vectors with higher attention weights and more important to model decisions, which helps improve the efficiency and performance of the model.
[0095] In addition, in order to improve the generalization and error correction capabilities of the model, this application performs data augmentation operations, that is, introducing random noise into the data, or randomly editing the data to generate additional data, and these additional data are also input into the model as training data.
[0096] In order to enable the model to capture the order of the data sequence and be able to process input sequences of different lengths, position information is added to the data sequence, that is, a position encoding operation is performed to encode the time series information in the time series into a vector and combine it with the input features to help the model better understand the time sequence.
[0097] Finally, the daily coal consumption of the target thermal power plant in the future is predicted based on the trained model. For example, the daily coal consumption forecast for the next 15 days is generated once instead of autoregression, which can avoid error accumulation and improve the reasoning speed of time series prediction.
[0098] This application reduces the computational complexity by introducing a sparse attention mechanism. Combining sparse attention and numerical features, a coal consumption forecast sequence for the next 15 days is dynamically generated. The multi-head sparse attention mechanism is used to capture the key dependencies of the time series to ensure the accuracy and stability of the forecast results. At the same time, this solution adopts a data enhancement strategy to improve the robustness of the model by adding noise, slicing, etc., thereby optimizing the forecast effect.
[0099] The beneficial effects of the present application are as follows: the present application collects the daily coal consumption of the target thermal power plant and the preset parameters related to the daily coal consumption, and pre-processes the daily coal consumption of the target thermal power plant and the preset parameters related to the daily coal consumption; then uses at least part of the pre-processed data as training data, and uses the sparse attention mechanism to train the model, and predicts the daily coal consumption of the future target thermal power plant based on the trained model. Since the present application completes model training and prediction based on the sparse attention mechanism, by reducing unnecessary calculations, the complexity is controlled at an approximately linear level, which greatly improves the speed and efficiency of the model in processing long sequence data, thereby improving the prediction efficiency of the coal consumption of thermal power plants.
[0100] In one embodiment, the above step S102 may be implemented as the following steps A1-A2:
[0101] In step A1, data slicing is performed on the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0102] In step A2, the data after data slicing is normalized.
[0103] In one embodiment, the above step A1 may be implemented as the following steps A11-A12:
[0104] In step A11, the preset time length of the data slice and the time length of each prediction are obtained;
[0105] In step A12, the data is sliced according to the preset time length of the data slice and the time length of each prediction.
[0106] In this embodiment, the preset time length of data slicing and the time length of each prediction are obtained; wherein the preset time length of data slicing is the length of data input into the model at one time, and the time length of prediction is the length of predicted data output by the model at one time. For example, if the preset time length of data slicing is 50 days and the time length of each prediction is 15 days, it means that 50 days of data are used to predict 15 days of data. The preset time length of data slicing and the time length of each prediction can be set in advance, and the optimal preset time length of data slicing and the time length of each prediction can also be determined by the model prediction effect after multiple slicing.
[0107] Then, the data is sliced according to the preset time length of the data slice and the time length of each prediction. When cutting, a set of preset time lengths and predicted time lengths are determined as a set of slice lengths. According to the slice lengths, multiple sets of slice data are obtained in sequence starting from the first day. For example, if 50 days of data are used to predict 15 days of data for a year, the data can be sliced into: (1) 1-50 days of data, the corresponding predicted data is 51-65 days of daily consumption data; (2) 2-51 days of input data, the corresponding output data is 52-66 days of data... The last one is 301-350 days of input data, and the corresponding output data is 351-365 days of data; and so on.
[0108] In one embodiment, the method can also be implemented as the following steps B1-B4:
[0109] In step B1, data sliced according to a preset time length is determined as input data;
[0110] In step B2, data sliced according to the time length of each prediction is determined as output data;
[0111] In step B3, the input data and the output data are associated with each other to obtain a plurality of associated data groups;
[0112] In step B4, the plurality of associated data sets are divided into a training set and a test set, wherein the training set is used for model training and the test set is used for model testing.
[0113] In this embodiment, the data sliced according to the preset time length is determined as input data; the data sliced according to the time length of each prediction is determined as output data; the preset time length is 50 days, and the time length of each prediction is 15 days. Then 50 days of data are used as input data, and 15 days of data are used as output data. The input data and output data are correspondingly associated to obtain multiple associated data groups; the multiple associated data groups are divided into training sets and test sets, wherein the training set is used for model training, and the test set is used for model testing.
[0114] In one embodiment, the preset parameters related to the daily coal consumption include daily historical coal consumption, inventory, planned power generation, and meteorological data. The above step A2 can be implemented as the following steps A21-A22:
[0115] In step A21, the daily historical coal consumption, inventory and planned power generation are obtained, and the daily historical coal consumption, inventory and planned power generation are normalized according to a standard normalization method;
[0116] In step A22, meteorological data is acquired, and the meteorological data is normalized according to the maximum and minimum normalization method.
[0117] In one embodiment, the above step S103 may be implemented as the following steps C1-C2:
[0118] In step C1, all training data are combined into input sequences according to time;
[0119] In step C2, the input sequence is encoded into an input vector by an encoder;
[0120] In step C3, the attention weight of each input vector is calculated;
[0121] In step C4, input vectors whose attention weights are lower than a preset weight threshold are eliminated to perform sparse processing on the input vectors.
[0122] In one embodiment, the method can also be implemented as the following steps D1-D2:
[0123] In step D1, additional data is generated by adding noise and / or random editing to the preprocessed data while training the model;
[0124] In step D2, the additional data is input into the model as training data.
[0125] Figure 2 FIG. 1 is a schematic diagram of a device for predicting coal consumption in a thermal power plant according to an embodiment of the present application. Figure 2 As shown, the device comprises:
[0126] A collection module 201 is used to collect the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0127] A preprocessing module 202, configured to preprocess the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0128] A training module 203, configured to use at least part of the preprocessed data as training data and train the model using a sparse attention mechanism;
[0129] The prediction module 204 is used to predict the daily coal consumption of the future target thermal power plant according to the trained model.
[0130] In one embodiment, the preprocessing module comprises:
[0131] A slicing submodule, used for slicing data of the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption;
[0132] The normalization submodule is used to normalize the data after data slicing.
[0133] In one embodiment, the slicing submodule is further used to:
[0134] Get the preset time length of data slices and the time length of each prediction;
[0135] The data is sliced according to a preset time length of the data slice and a time length of each prediction.
[0136] In one embodiment, the apparatus further comprises:
[0137] A determination module, used to determine data sliced according to a preset time length as input data;
[0138] The determination module is further used to determine the data sliced according to the time length of each prediction as output data;
[0139] An associating module, used to associate the input data with the output data to obtain a plurality of associated data groups;
[0140] The partitioning module is used to partition a plurality of associated data groups into a training set and a test set, wherein the training set is used for model training and the test set is used for model testing.
[0141] In one embodiment, the preset parameters related to the daily coal consumption include daily historical coal consumption, inventory, planned power generation, and meteorological data. The normalization submodule is further used to:
[0142] Obtain daily historical coal consumption, inventory, and planned power generation, and normalize the daily historical coal consumption, inventory, and planned power generation according to a standard normalization method;
[0143] Acquire meteorological data, and perform data normalization on the meteorological data according to a maximum and minimum normalization method.
[0144] In one embodiment, the training module includes:
[0145] The merging submodule is used to merge all training data into input sequences according to time;
[0146] An encoding submodule, used for encoding the input sequence into an input vector through an encoder;
[0147] The calculation submodule is used to calculate the attention weight of each input vector;
[0148] The elimination submodule is used to eliminate input vectors whose attention weights are lower than a preset weight threshold so as to perform sparse processing on the input vectors.
[0149] In one embodiment, the training module includes:
[0150] Add submodules for generating additional data by adding noise and / or random edits to preprocessed data while training the model;
[0151] The input submodule is used to input the additional data into the model as training data.
[0152] Figure 3 FIG. 1 is a schematic diagram of the hardware structure of a coal consumption prediction system for a thermal power plant in one embodiment of the present application. Figure 3 As shown, the coal consumption prediction system of the thermal power plant includes:
[0153] at least one processor 320; and,
[0154] A memory 304 in communication with the at least one processor 320; wherein,
[0155] The memory 304 stores instructions that can be executed by the at least one processor 320, and the instructions are executed by the at least one processor 320 to implement the method for predicting coal consumption of a thermal power plant recorded in any of the above embodiments.
[0156] Reference Figure 3The thermal power plant coal consumption prediction system 300 may include one or more of the following components: a processing component 302, a memory 304, a power supply component 306, a multimedia component 308, an audio component 310, an input / output (I / O) interface 312, a sensor component 314, and a communication component 316.
[0157] The processing component 302 generally controls the overall operation of the thermal power plant coal consumption prediction system 300. The processing component 302 may include one or more processors 320 to execute instructions to complete all or part of the steps of the above method. In addition, the processing component 302 may include one or more modules to facilitate the interaction between the processing component 302 and other components. For example, the processing component 302 may include a multimedia module to facilitate the interaction between the multimedia component 308 and the processing component 302.
[0158] The memory 304 is configured to store various types of data to support the operation of the thermal power plant coal consumption prediction system 300. Examples of these data include instructions for any application or method operating on the thermal power plant coal consumption prediction system 300, such as text, pictures, videos, etc. The memory 304 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.
[0159] The power supply component 306 provides power to various components of the thermal power plant coal consumption prediction system 300. The power supply component 306 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the thermal power plant coal consumption prediction system 300.
[0160] The multimedia component 308 includes a screen that provides an output interface between the thermal power plant coal consumption prediction system 300 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor can not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 308 may also include a front camera and / or a rear camera. When the thermal power plant coal consumption prediction system 300 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have a focal length and optical zoom capability.
[0161] The audio component 310 is configured to output and / or input audio signals. For example, the audio component 310 includes a microphone (MIC), and when the thermal power plant coal consumption prediction system 300 is in an operation mode, such as an alarm mode, a recording mode, a voice recognition mode, and a voice output mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 304 or sent via the communication component 316. In some embodiments, the audio component 310 also includes a speaker for outputting an audio signal.
[0162] I / O interface 312 provides an interface between processing component 302 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.
[0163] The sensor assembly 314 includes one or more sensors for providing various aspects of status assessment for the thermal power plant coal consumption prediction system 300. For example, the sensor assembly 314 may include a sound sensor. In addition, the sensor assembly 314 may detect the open / closed state of the thermal power plant coal consumption prediction system 300, the relative positioning of the components, such as the display and keypad of the thermal power plant coal consumption prediction system 300, and the sensor assembly 314 may also detect the operating state of the thermal power plant coal consumption prediction system 300 or a component of the thermal power plant coal consumption prediction system 300, such as the operating state of the air distribution plate, the structural state, the operating state of the discharge scraper, etc., the orientation or acceleration / deceleration of the thermal power plant coal consumption prediction system 300 and the temperature change of the thermal power plant coal consumption prediction system 300. The sensor assembly 314 may include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 314 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 314 may also include a pressure sensor or a temperature sensor.
[0164] The communication component 316 is configured to enable the thermal power plant coal consumption prediction system 300 to provide the ability to communicate with other devices and cloud platforms in a wired or wireless manner. The thermal power plant coal consumption prediction system 300 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 316 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 316 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
[0165] In an exemplary embodiment, the coal consumption prediction system 300 for a thermal power plant can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to execute the coal consumption prediction method for a thermal power plant described in any of the above embodiments.
[0166] The present application also provides a computer-readable storage medium. When the instructions in the storage medium are executed by a processor corresponding to the thermal power plant coal consumption prediction system, the thermal power plant coal consumption prediction system can implement the thermal power plant coal consumption prediction method recorded in any of the above embodiments.
[0167] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.
[0168] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0169] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0170] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0171] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.
Claims
1. A method for predicting coal consumption in a thermal power plant, characterized in that: include: Collecting the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption; Preprocessing the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption; Using at least part of the preprocessed data as training data and training the model using a sparse attention mechanism; The daily coal consumption of the future target thermal power plant is predicted based on the trained model.
2. The method according to claim 1, characterized in that The preprocessing of the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption includes: Performing data slicing on the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption; Normalize the data after data slicing.
3. The method according to claim 2, characterized in that The data slicing of the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption includes: Get the preset time length of data slices and the time length of each prediction; The data is sliced according to a preset time length of the data slice and a time length of each prediction.
4. The method according to claim 3, characterized in that The method further comprises: Determine the data sliced according to the preset time length as input data; Determine the data sliced according to the time length of each prediction as output data; Correlating the input data and the output data accordingly to obtain a plurality of associated data groups; The multiple associated data sets are divided into a training set and a test set, wherein the training set is used for model training and the test set is used for model testing.
5. The method according to claim 2, characterized in that The preset parameters related to the daily coal consumption include daily historical coal consumption, inventory, planned power generation, and meteorological data. The normalization of the data after data slicing includes: Obtain daily historical coal consumption, inventory, and planned power generation, and normalize the daily historical coal consumption, inventory, and planned power generation according to a standard normalization method; Acquire meteorological data, and perform data normalization on the meteorological data according to a maximum and minimum normalization method.
6. The method according to claim 1, characterized in that The method of using at least part of the preprocessed data as training data and using a sparse attention mechanism to train the model includes: Merge all training data into input sequences according to time; Encode the input sequence into an input vector through an encoder; Calculate the attention weight for each input vector; Input vectors whose attention weights are lower than a preset weight threshold are eliminated to perform sparse processing on the input vectors.
7. The method of claim 1, further comprising: When training the model, additional data is generated by adding noise and / or random edits to the preprocessed data; The additional data is input into the model as training data.
8. A device for predicting coal consumption in a thermal power plant, characterized in that: include: A collection module, used for collecting the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption; A preprocessing module, used for preprocessing the daily coal consumption of the target thermal power plant and preset parameters related to the daily coal consumption; A training module, configured to use at least part of the preprocessed data as training data and train the model using a sparse attention mechanism; The prediction module is used to predict the daily coal consumption of the future target thermal power plant based on the trained model.
9. A coal consumption prediction system for a thermal power plant, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to implement the method for predicting coal consumption of a thermal power plant as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor corresponding to the thermal power plant coal consumption prediction system, the thermal power plant coal consumption prediction system can implement the thermal power plant coal consumption prediction method as described in any one of claims 1-7.