Power station water network water consumption prediction method and system based on multi-source data fusion, and storage medium

Abstract

The invention belongs to the technical field of power system water resource scheduling and intelligent prediction, and relates to a power station water network water consumption prediction method and system based on multi-source data fusion, and a storage medium. The method comprises the following steps: acquiring historical multi-source data and historical water consumption data influencing water consumption of a power station water network; preprocessing the historical multi-source data and the historical water consumption data; constructing a multi-source data fusion neural network basic model based on the preprocessed data; training and optimizing the multi-source data fusion neural network basic model based on historical multi-source data and historical water consumption data to obtain a multi-source data fusion neural network model; and collecting real-time multi-source data influencing water consumption of the power station water network, and inputting the real-time multi-source data into the multi-source data fusion neural network model to obtain a prediction result. The precision and timeliness of power station water network water consumption prediction can be improved, the operation cost of a power station is effectively reduced, and the power generation efficiency of the power station is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system water resource scheduling and intelligent prediction, and relates to a method, a system and a storage medium for predicting water consumption of a power station water network by fusing multi-source data. Background Art

[0002] A highly efficient large-scale power station water network is the most important water resource supply system in the power production process, and it is necessary to maintain its operating efficiency. Its water consumption situation will directly affect the operating cost and power generation efficiency of the power station. At the same time, it is also necessary to predict and schedule the future water consumption of the power station water network during production to meet the efficient production scheduling of the power station.

[0003] Traditional methods for predicting water network water consumption mainly predict and schedule based on the historical data of the water network. However, in the actual power production process, the water consumption of the water network is not only affected by the operating load of the power station and the operating conditions of the units, but also affected by factors such as climate change and seasonal changes, as well as the water resource reserves such as the water level of the reservoir near the power station and the river flow. The prediction method based on a single data source of the historical data of the water network has serious problems such as low accuracy and slow response speed. It is necessary to adopt some new prediction methods to efficiently obtain more accurate water consumption predictions and improve the accuracy and timeliness of the water consumption prediction of the power station water network. Summary of the Invention

[0004] In order to solve the problems in the prior art, the purpose of the present invention is to provide a method, a system and a storage medium for predicting water consumption of a power station water network by fusing multi-source data, which can improve the accuracy and timeliness of the water consumption prediction of the power station water network, and thus effectively reduce the operating cost of the power station and improve the power generation efficiency of the power station.

[0005] To achieve the above object, the present invention adopts the following technical solutions: In the first aspect, the present invention provides a method for predicting water consumption of a power station water network by fusing multi-source data, including the following steps: Obtain historical multi-source data and historical water consumption data that affect the water consumption of the power station water network; Preprocess the historical multi-source data and historical water consumption data; Construct a multi-source data fusion neural network basic model based on the preprocessed data; Train and optimize the multi-source data fusion neural network basic model based on the historical multi-source data and historical water consumption data to obtain a multi-source data fusion neural network model; Collect real-time multi-source data that affect the water consumption of the power station water network, and input the real-time multi-source data into the multi-source data fusion neural network model to obtain a prediction result.

[0006] Preferably, the multi-source data includes power station load and power generation data, water resource reserve data, and climate change data.

[0007] Preferably, the specific method for preprocessing the historical multi-source data and historical water usage data is as follows: Analyze the sensitivity of historical power station load and generation data, historical water resource reserve data, and historical climate change data to historical water usage data, and extract the influence weight parameters of historical multi-source data on historical water usage data.

[0008] Preferably, the preprocessed data is standardized before constructing the basic model of the multi-source data fusion neural network.

[0009] Preferably, the specific method for constructing the basic model of the multi-source data fusion neural network based on the preprocessed data is as follows: Use a long short-term memory network model to describe the long-term dependence relationship between historical water usage data and the water consumption of the power station water network based on the weight parameters, and use a multi-layer perceptron fusion model to describe the relationship between power station load and generation data, water resource reserve data, climate change data, and the water consumption of the power station water network; Integrate the long short-term memory network model and the multi-layer perceptron fusion model to form the basic model of the multi-source data fusion neural network.

[0010] Preferably, the construction method of the multi-layer perceptron fusion model is as follows: Use the first multi-layer perceptron neural network model to describe the relationship between power station load and generation data and the water consumption of the power station water network based on the weight parameters, use the second multi-layer perceptron neural network model to describe the relationship between water resource reserve data and the water consumption of the power station water network, and use the third multi-layer perceptron neural network model to describe the relationship between climate change data and the water consumption of the power station water network; Integrate the first multi-layer perceptron neural network model, the second multi-layer perceptron neural network model, and the third multi-layer perceptron neural network model to form the multi-layer perceptron fusion model.

[0011] Preferably, the specific method for training and optimizing the basic model of the multi-source data fusion neural network based on historical multi-source data and historical water usage data is as follows: Input the historical multi-source data and historical water usage data into the basic model of the multi-source data fusion neural network for training, and optimize the parameters of the basic model of the multi-source data fusion neural network through cross-validation and grid search.

[0012] Preferably, after obtaining the prediction result, compare the prediction result with the actual water consumption of the power station water network to obtain error feedback, and perform iterative optimization on the multi-source data fusion neural network model based on the error feedback.

[0013] In a second aspect, the present invention provides a power station water network water usage prediction system for multi-source data fusion, including: The first module: Obtain historical multi-source data and historical water usage data that affect the water usage of the power station water network; The second module: Preprocess the historical multi-source data and historical water usage data; The third module: Construct a basic model of a multi-source data fusion neural network based on preprocessed data; The fourth module: Train and optimize the basic model of the multi-source data fusion neural network based on historical multi-source data and historical water usage data to obtain a multi-source data fusion neural network model; The fifth module: Collect real-time multi-source data affecting the water usage of the power station water network, input the real-time multi-source data into the multi-source data fusion neural network model, and obtain a prediction result.

[0014] Thirdly, the present invention provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the method for predicting the water usage of a power station water network with multi-source data fusion are implemented.

[0015] Compared with the prior art, the present invention has the following beneficial effects: By effectively fusing multi-source data and water usage data, the present invention can provide more accurate predictions of key parameters, and then accurately predict the future water usage of the power station water network, providing a scientific basis for the water resource scheduling and management of the power station; through the multi-source data fusion neural network model, short-term and long-term water network water usage predictions are realized, which helps the power station water network system to adjust the water resource scheduling in real time, effectively avoid water resource waste, reduce the operating cost of the power station, and improve the power generation efficiency of the power station.

[0016] Furthermore, the present invention iteratively optimizes the multi-source data fusion neural network model based on the prediction result and the real-time scheduling result of the power station, continuously obtains the optimal power generation load and water resource scheduling; at the same time, it reduces manual intervention and improves the automation and intelligence level of the power station water resource management. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic flowchart of the method for predicting the water usage of a power station water network with multi-source data fusion of the present invention; Figure 2 It is a schematic diagram of the principle for constructing a basic model of a multi-source data fusion neural network. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. The components of the embodiments of the present invention usually described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0020] Therefore, the detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0021] It should be noted that like reference numerals and letters denote like items in the following drawings. Therefore, once an item is defined in one drawing, it does not require further definition and explanation in subsequent drawings.

[0022] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper", "lower", "horizontal", "inner", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, or the orientations or positional relationships in which the inventive product is customarily placed during use, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation of the present invention. In addition, terms such as "first", "second", etc. are only used for descriptive distinction and cannot be understood as indicating or implying relative importance.

[0023] In addition, if the term "horizontal" appears, it does not mean that the component is required to be absolutely horizontal, but it can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that the structure must be completely horizontal, but it can be slightly inclined.

[0024] In the description of the embodiments of the present invention, it should also be noted that unless otherwise clearly specified and limited, if terms such as "set", "installed", "connected", "coupled" are to be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0025] The present invention will be further described in detail below with reference to the accompanying drawings: The present invention provides a method for predicting water consumption in a power station water network with multi-source data fusion, as Figure 1 shown, which includes the following steps: (1) Obtain historical multi-source data and historical water consumption data that affect the water consumption in the power station water network: The multi-source data includes power station load and generation data, water resource reserve data, and climate change data; among them, the power station load and generation data includes the operating load and power generation volume data of the power station and the cooling water usage required by the power station water network; the water resource reserve data includes data on water resource supply status such as reservoir water level and river flow; the climate change data includes meteorological data such as temperature, precipitation, humidity, and wind speed around the power station; the historical water consumption data includes water consumption data such as the historical daily water consumption and seasonal fluctuations of the power station.

[0026] In this step, by obtaining the power station load and generation data, water resource reserve data, climate change data, and historical water consumption data, the core dynamic factors affecting the water consumption in the power station water network are comprehensively covered, effectively breaking through the limitation of traditional prediction methods that only rely on single historical water consumption data. Specifically, the power station load and generation data can directly reflect the real-time demand for water resources by the unit operation; the water resource reserve data quantifies the dynamic constraints of the water network water supply capacity; the climate change data reveals the indirect influence law of the environment on water consumption; and the historical water consumption data provides a long-term trend benchmark for water consumption behavior. By fusing the above multi-source data, the present invention can accurately depict the coupling relationship among power station operation, environmental conditions, and resource supply, thereby significantly improving the coverage dimension and data-driven ability of the prediction model, and laying a scientific and reliable data support for high-precision water consumption prediction.

[0027] (2) Preprocess the historical multi-source data and historical water consumption data: Analyze the sensitivity of historical power station load and generation data, historical water resource reserve data, and historical climate change data to historical water consumption data, and extract the influence weight parameters of historical multi-source data on historical water consumption data.

[0028] In this step, by quantifying the contribution degree of different factors to water consumption (such as the direct influence of power generation load on cooling water demand and the indirect effect of climate change on evaporation loss), key variables with high correlation can be scientifically screened, noise data can be eliminated, and the prediction accuracy of the model can be prevented from being reduced due to interference by irrelevant or weakly correlated features. At the same time, the extracted weight parameters provide an objective basis for data fusion in subsequent model construction, enabling the neural network to allocate learning weights according to the actual influence intensity, which not only enhances the interpretability of the model but also optimizes the efficiency of feature input, improving the robustness of the model and the reliability of the prediction results.

[0029] (3) Build a multi-source data fusion neural network basic model based on the preprocessed data: Standardize historical multi-source data and historical water usage data to ensure the consistency of the dimensions of multi-source data, facilitating the next step of data fusion; Based on the artificial neural network intelligent algorithm of deep learning, perform weighted fusion in sequence according to the characteristics and influence weight parameters of different data sources, and adopt different artificial neural network algorithms according to different data characteristics. Specifically, use the long short-term memory network model to describe the long-term dependence relationship between historical water usage data and the water consumption of the power station water network based on the weight parameters; use the first multi-layer perceptron neural network model to describe the relationship between power station load and power generation data and the water consumption of the power station water network based on the weight parameters, use the second multi-layer perceptron neural network model to describe the relationship between water resource reserve data and the water consumption of the power station water network, and use the third multi-layer perceptron neural network model to describe the relationship between climate change data and the water consumption of the power station water network; as Figure 2 shown, fuse the first multi-layer perceptron neural network model, the second multi-layer perceptron neural network model, and the third multi-layer perceptron neural network model to form a multi-layer perceptron fusion model; fuse the long short-term memory network model and the multi-layer perceptron fusion model to form a multi-source data fusion neural network basic model.

[0030] In this step, the dimension differences of multi-source data are eliminated through standardization processing to ensure the fairness and comparability of data fusion. At the same time, a modular neural network model is constructed based on data characteristics and weight parameters, solving the problem that it is difficult for traditional single models to take into account both time series dependencies and static features. Specifically, use the long short-term memory network (LSTM) to model the time series law of historical water usage data (such as seasonal fluctuations) to accurately capture long-term dynamic associations; for non-time series or static data such as power station load and power generation data, water resource reserve data, and climate change data, design multi-layer perceptron (MLP) sub-models respectively, and extract their non-linear mapping relationships with water consumption through independent modeling to avoid the decline of model performance caused by feature mixing. Finally, by fusing the LSTM and multi-branch MLP models, a composite neural network architecture with both time series dynamic analysis and static multi-factor collaboration capabilities is formed, which not only strengthens the model's long-term memory ability for historical water usage trends but also improves its response sensitivity to real-time operating parameters, environmental conditions, and resource constraints, significantly enhancing the prediction accuracy and scenario adaptability under complex multi-source data.

[0031] (4) Train and optimize the multi-source data fusion neural network basic model based on historical multi-source data and historical water usage data to obtain a multi-source data fusion neural network model: Input historical multi-source data and historical water usage data into the multi-source data fusion neural network basic model for training, and optimize the parameters of the multi-source data fusion neural network basic model through methods such as cross-validation and grid search.

[0032] In this step, the basic model of the multi-source data fusion neural network is systematically trained and parameter-optimized through cross-validation and grid search methods, solving the problems of traditional model parameter tuning relying on experience and being prone to overfitting or underfitting. Cross-validation validates through multiple rounds of data segmentation to ensure that the model has stable generalization ability under complex data distributions and avoid prediction distortion caused by local data biases; grid search traverses hyperparameter combinations to accurately locate the optimal parameter configuration (such as learning rate, network depth), significantly improving the model's convergence speed and prediction accuracy. At the same time, the training process based on historical multi-source data fully explores the non-linear relationships between factors such as power station load, water resource reserves, and climate change and water consumption, enabling the model to adaptively capture dynamic coupling relationships, and finally forming a prediction model with both high robustness and strong generalization ability, providing reliable technical support for real-time water consumption prediction and effectively supporting the precise scheduling and resource optimization of the power station water network.

[0033] (5) Collect real-time multi-source data affecting the water use of the power station water network, input the real-time multi-source data into the multi-source data fusion neural network model, and obtain the prediction results: Collect real-time multi-source data affecting the water use of the power station water network, and input it into the multi-source data fusion neural network model to obtain the prediction results of the water use demand of the power station water network for a future period of time (such as daily, weekly, monthly).

[0034] In this step, based on real-time data (such as current power generation load, reservoir water level, weather forecast), the coupled effects of multiple factors are dynamically analyzed, and the water consumption prediction results at different scales such as daily, weekly, and monthly are accurately output, enabling the power station to anticipate the peak water use or shortage risks in advance, and timely adjust strategies such as cooling water supply and reservoir scheduling to avoid resource waste or insufficient supply.

[0035] (6) Iterative update of the multi-source data fusion neural network model: After obtaining the prediction results, compare the prediction results with the actual water use of the power station water network to obtain error feedback, and iteratively optimize the multi-source data fusion neural network model based on the error feedback to obtain a more accurate multi-source data fusion neural network model.

[0036] In this step, by establishing a closed-loop feedback mechanism of "prediction - verification - optimization", using the error between the prediction results and the actual water consumption as the basis for dynamic optimization, it effectively solves the problem of the prediction accuracy decay of traditional static models caused by data distribution shift or environmental mutation. Through iterative optimization, continuously correct the model's ability to depict the correlation relationships of multi-source data, avoid cumulative errors caused by historical data limitations or sudden events (such as extreme weather, unit failures), further improve the stability and prediction accuracy of the model's long-term operation, and endow it with the ability of self-evolution to ensure that the prediction system continues to optimize with the evolution of business needs, providing sustainable and highly reliable water consumption prediction services for the power station water network.

[0037] The second object of the present invention is to provide a power station water network water consumption prediction system for multi-source data fusion, including: The first module: obtaining historical multi-source data and historical water consumption data affecting the water consumption of the power station water network; The second module: preprocessing the historical multi-source data and historical water consumption data; The third module: constructing a basic multi-source data fusion neural network model based on the preprocessed data; The fourth module: training and optimizing the basic multi-source data fusion neural network model based on the historical multi-source data and historical water consumption data to obtain a multi-source data fusion neural network model; The fifth module: collecting real-time multi-source data affecting the water consumption of the power station water network, inputting the real-time multi-source data into the multi-source data fusion neural network model, and obtaining a prediction result.

[0038] By effectively fusing multi-source data and water consumption data, the present invention takes into account various influencing factors and can achieve more accurate prediction of the water consumption of the power station water network. The present invention can improve the accuracy and timeliness of the prediction of the water consumption of the power station water network, provide a scientific basis for the water resource scheduling and management of the power station, and effectively reduce the operating cost of the power station. In addition, the multi-source data fusion neural network model of the present invention realizes the provision of short-term and long-term water network water consumption predictions, which helps the power station water network system to adjust the water resource scheduling in real time, effectively avoid water resource waste, reduce the operating cost of the power station, and improve the power generation efficiency of the power station.

[0039] The third object of the present invention is to provide a storage medium, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is a memory device in a computer device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and, of course, the extended storage medium supported by the computer device. The computer-readable storage medium provides a storage space, and this storage space stores the operating system of the terminal. And, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the method for predicting the water consumption of the power station water network for multi-source data fusion in the above embodiments.

[0040] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0041] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices produce a means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0042] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing devices to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction means that implements the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0043] These computer program instructions can also be loaded onto a computer or other programmable data processing devices, such that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: still can modify the specific implementation manners of the present invention or make equivalent replacements, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for predicting water consumption of a power station water network by fusion of multi-source data, characterized in that: The following steps are involved: Obtain historical multi-source data and historical water use data that affect the water network water use of power stations; Preprocess historical multi-source data and historical water use data; Construct a multi-source data fusion neural network basic model based on preprocessed data; Based on historical multi-source data and historical water use data, the basic model of multi-source data fusion neural network is trained and optimized to obtain a multi-source data fusion neural network model; Collect real-time multi-source data that affects the water consumption of the power station water network, input the real-time multi-source data into the multi-source data fusion neural network model, and obtain the prediction results.

2. The method for predicting water consumption of a power station water network by fusion of multi-source data according to claim 1 is characterized in that: The multi-source data include power station load and power generation data, water resource reserve data and climate change data.

3. The method for predicting water consumption of a power station water network by fusion of multi-source data according to claim 2 is characterized in that: The specific method for preprocessing the historical multi-source data and historical water use data is: analyzing the sensitivity of historical power station load and power generation data, historical water resource reserve data and historical climate change data to historical water use data, and extracting the impact weight parameters of historical multi-source data on historical water use data.

4. The method for predicting water consumption of a power station water network by fusion of multi-source data according to claim 1 is characterized in that: The preprocessed data is standardized before constructing the multi-source data fusion neural network basic model.

5. The method for predicting water consumption of a power station water network by fusion of multi-source data according to claim 1 is characterized in that: The specific method for constructing a multi-source data fusion neural network basic model based on preprocessed data is as follows: using a long short-term memory network model based on weight parameters to describe the long-term dependence between historical water use data and water consumption of power station water networks, and using a multi-layer perceptron fusion model to describe the relationship between power station load and power generation data, water resource reserve data, and climate change data and water consumption of power station water networks; and fusing the long short-term memory network model and the multi-layer perceptron fusion model to form a multi-source data fusion neural network basic model.

6. The method for predicting water consumption of a power station water network by fusion of multi-source data according to claim 1 is characterized in that: The method for constructing the multi-layer perceptron fusion model is as follows: based on weight parameters, a first multi-layer perceptron neural network model is used to describe the relationship between power station load and power generation data and the water consumption of the power station water network, a second multi-layer perceptron neural network model is used to describe the relationship between water resource reserve data and the water consumption of the power station water network, and a third multi-layer perceptron neural network model is used to describe the relationship between climate change data and the water consumption of the power station water network; and the first multi-layer perceptron neural network model, the second multi-layer perceptron neural network model and the third multi-layer perceptron neural network model are fused to form a multi-layer perceptron fusion model.

7. The method for predicting water consumption of a power station water network by fusion of multi-source data according to claim 1 is characterized in that: The specific method for training and optimizing the multi-source data fusion neural network basic model based on historical multi-source data and historical water use data is: inputting historical multi-source data and historical water use data into the multi-source data fusion neural network basic model for training, and optimizing the parameters of the multi-source data fusion neural network basic model through cross-validation and grid search.

8. The method for predicting water consumption of a power station water network by fusion of multi-source data according to claim 1 is characterized in that: After the prediction result is obtained, the prediction result is compared with the actual water consumption of the power station water network to obtain error feedback, and the multi-source data fusion neural network model is iteratively optimized based on the error feedback.

9. A multi-source data fusion power station water network water consumption prediction system, applied to the method described in any one of claims 1 to 8, characterized in that: include: Module 1: Obtain historical multi-source data and historical water use data that affect the water network water use of power stations; Module 2: Preprocessing of historical multi-source data and historical water use data; Module 3: Construct a basic model of multi-source data fusion neural network based on preprocessed data; Module 4: Based on historical multi-source data and historical water use data, the basic model of the multi-source data fusion neural network is trained and optimized to obtain a multi-source data fusion neural network model; The fifth module: Collect real-time multi-source data that affects the water consumption of the power station water network, input the real-time multi-source data into the multi-source data fusion neural network model, and obtain the prediction results.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images