Cascade power station dispatching parameter access and influence factor prediction method

Abstract

The invention discloses a cascade power station scheduling parameter access and influence factor prediction method, and relates to the technical field of power station scheduling, and the method comprises the following steps: obtaining the operation parameter data of each cascade power station, including but not limited to flow, output, storage capacity and water level; multi-dimensional impact factor data related to power station dispatching is collected, and impact factors comprise meteorological data, geographic data and hydrological data. According to the method, the reservoir level is accurately predicted through multi-source data fusion and deep learning, prediction errors are reduced, scheduling is optimized, and extreme meteorology is especially coped with. A feedback adjustment mechanism is introduced to reduce abandoned water and improve water and electricity utilization efficiency and power grid flexibility. By combining an intelligent optimization algorithm and deep reinforcement learning, a scheduling scheme is dynamically adjusted, the utilization of hydroelectric resources is maximized, the safety of a power grid is ensured, overload or insufficient supply is avoided, and the emergency response capability and stability of the power grid are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of power station dispatching, and in particular to a method for accessing dispatching parameters of cascade power stations and predicting influencing factors. Background Art

[0002] The access to dispatching parameters and prediction of influencing factors of cascade power stations refers to the process of joint dispatching of hydropower cascades. The operating parameters of each cascade power station (such as flow, output, storage capacity, water level, etc.) are connected to the dispatching system in real time, and various factors that may affect the dispatching of the power station are predicted to achieve the optimal dispatching decision. Influencing factors include meteorological data (such as rainfall, temperature, wind speed, etc.), hydrological data (such as water level, water flow, snow melt, etc.) and other natural or man-made factors that may interfere with the operation of the power system (such as ice and snow accumulation, environmental changes, etc.). Changes in these factors may have a significant impact on the water inflow, power generation and regulation capacity of the hydropower station, so accurate prediction and analysis are needed.

[0003] In high-altitude cold regions such as Tibet, the prediction of these influencing factors is particularly important due to the complex and volatile climate conditions. Through accurate prediction of these factors, the dispatching system can prepare response plans in advance, such as adjusting the reservoir dispatching method, avoiding power waste (abandoned water) due to forecasting errors, and ensuring the balance of power supply and demand and the stable operation of the power grid. Therefore, the access of cascade power station dispatching parameters and the prediction of influencing factors are key links in realizing the joint optimization dispatching of hydropower stations, which can greatly improve the dispatching accuracy and flexibility of the power system and ensure the safe operation of the power grid.

[0004] The existing technology has the following deficiencies: In the existing technology, in the process of accessing the dispatching parameters of cascade power stations and predicting the influencing factors, inaccurate modeling of the nonlinear relationship between the influencing factors and the key water level may lead to serious consequences. Although the functional relationship between the key water level of the reservoir and the influencing factors can be fitted by data mining or regression methods, these relationships usually present nonlinear characteristics due to the complexity of meteorological, geographical and hydrological factors. Existing regression models often find it difficult to fully capture these complex nonlinear changes, especially when extreme climate events or sudden precipitation occur. At this time, the prediction of the model may have a large deviation, resulting in the dispatching system making wrong decisions. For example, when there is a sudden increase in extreme precipitation or ice and snow melt water flow, failure to adjust the reservoir dispatch in time may result in the water flow not being fully utilized, or conversely, failure to release too much water, resulting in insufficient power supply or excessive load on the power grid, and even in severe cases, it may cause the collapse of the power system. Therefore, accurately modeling the nonlinear relationship between the influencing factors and the key water level is crucial to ensuring the efficiency of the joint dispatching of hydropower stations and the safe and stable operation of the power grid.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a method for accessing the dispatching parameters of cascade power stations and predicting influencing factors, and to accurately predict the water level of reservoirs through multi-source data fusion and deep learning, reduce prediction errors and optimize dispatching, especially in response to extreme weather. The introduction of a feedback adjustment mechanism can reduce water abandonment, improve hydropower utilization efficiency and grid flexibility. Combined with intelligent optimization algorithms and deep reinforcement learning, the dispatching plan is dynamically adjusted to maximize the utilization of hydropower resources, ensure grid safety, avoid overload or insufficient supply, and enhance the emergency response capability and stability of the grid to solve the problems in the above-mentioned background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a method for accessing scheduling parameters and predicting influencing factors of cascade power stations, comprising the following steps:

[0008] Obtain operating parameter data of each cascade power station, including but not limited to flow, output, reservoir capacity and water level;

[0009] Collect multi-dimensional influencing factor data related to power station dispatching, wherein the influencing factors include meteorological data, geographical data, and hydrological data;

[0010] Based on the difficulty of predicting the influencing factors, the correlation analysis between the influencing factors and the critical water level of the reservoir is carried out to identify and determine the most relevant influencing factors;

[0011] The regression analysis method is used to fit the nonlinear functional relationship between the most relevant influencing factors and the critical water level of the reservoir;

[0012] The obtained nonlinear function relationship is integrated into the dispatching system of cascade power stations to predict the changes of key water levels in power station reservoirs in real time;

[0013] Based on the prediction results, the joint dispatching plan of the power station is dynamically adjusted to optimize the output regulation of the hydropower station, reduce water abandonment, and ensure the supply and demand balance of the power system and the stable operation of the power grid.

[0014] Preferably, the operating parameter data of each cascade power station is collected in real time through sensors and monitoring equipment, and transmitted to the dispatching system through wireless communication technology or Internet connection, so as to realize efficient transmission and centralized management of real-time data, ensure the timeliness and accuracy of the data, and ensure that the dispatching system can quickly respond to changes in the operating status of the power station.

[0015] Preferably, meteorological data, geographic data and hydrological data are comprehensively collected through a variety of data sources, including meteorological stations, meteorological satellites, remote sensing systems and hydrological monitoring stations. The data are fused and cleaned to remove outliers and noise to improve prediction accuracy. In addition, by adopting data fusion technology and combining the spatiotemporal characteristics of different data sources, the collected data is more representative, thereby enhancing the reliability and accuracy of subsequent predictions.

[0016] Preferably, the correlation analysis between the influencing factors and the critical water level of the reservoir adopts a weighted correlation analysis method. Different weights are assigned to different factors according to the distribution characteristics of the influencing factors in time and space. The weighted Pearson correlation coefficient is used to analyze the correlation between each influencing factor and the critical water level. By considering the importance between the influencing factors, the accuracy of the correlation analysis is improved. This method is suitable for areas with complex meteorological and hydrological characteristics, and improves the accuracy of the selection of influencing factors and the prediction of the critical water level of the reservoir.

[0017] Preferably, the regression analysis method uses polynomial regression. When dealing with nonlinear relationships, a high-order polynomial regression model is used. By cross-validating and tuning historical data, the hyperparameters of the regression model are optimized to ensure that the fitted nonlinear function can adapt to water level prediction under different seasons and climate changes. In the regression process, the error function is minimized to ensure the high-precision prediction capability of the regression model in various practical scenarios.

[0018] Preferably, the dispatching system integrates artificial intelligence algorithms, including deep learning models, to analyze in real time the complex nonlinear relationship between influencing factors and critical water levels of reservoirs. Through continuous learning and optimization, it adaptively adjusts weights and corrects prediction errors based on feedback from historical data and real-time data to ensure the accuracy of reservoir dispatching. During the adjustment process, it takes into account the actual operating status of each cascade power station, while comprehensively considering various meteorological and hydrological changes to further improve the flexibility and response speed of the joint dispatching of hydropower cascades.

[0019] Preferably, the deep learning model further includes a long short-term memory network, which is used to process the temporal relationship between the influencing factors and the critical water level of the reservoir, and specifically includes the following steps:

[0020] By preprocessing historical meteorological, geographical and hydrological data, it is converted into time series data and input into the LSTM model for training;

[0021] During the training process of the LSTM network, the back-propagation algorithm is used to optimize the model parameters, calculate the error and adjust the network weights to achieve accurate modeling of nonlinear temporal relationships;

[0022] The trained LSTM model is used to predict the critical water level of the reservoir in real time, and the prediction results are input into the dispatching system as feedback, thereby optimizing the joint dispatching strategy of hydropower cascade power stations, reducing water abandonment and improving dispatching flexibility.

[0023] Preferably, the specific steps of dynamically adjusting the joint dispatching scheme of the power station are as follows:

[0024] Calculate the critical water level of each cascade power station reservoir and obtain the predicted water level based on the output of the regression model;

[0025] For each cascade power station, the dispatch deviation is calculated, and the calculation expression is as follows:

[0026] , where W i is the current critical water level of the reservoir of the i-th cascade power station, is the predicted critical water level of the reservoir of the i-th cascade power station calculated according to the regression model, D i is the dispatch deviation of the i-th cascade power station;

[0027] During the scheduling process, the feedback adjustment factor based on the prediction error is calculated, and the calculation expression is as follows:

[0028] ,

[0029] In the formula, F i (t) is the feedback adjustment factor, which is used to dynamically adjust the dispatching plan of the power station according to the prediction error. i (t-1) is the dispatch deviation of the i-th cascade power station at time t-1, α is the weight factor of the dispatch deviation, β is the weight factor of the dispatch deviation of other cascade power stations, and w ij is the influence coefficient of cascade power station j on the dispatch of cascade power station i, and N is the total number of cascade power stations involved in the dispatch

[0030] According to the calculated feedback adjustment factor F i (t) Dynamically modify the dispatching plan to ensure the optimal dispatching of cascade power stations under different meteorological, hydrological and water level conditions, and ensure the regulation capacity and operational stability of the dispatching system.

[0031] Preferably, the joint scheduling scheme further comprises the following steps:

[0032] Calculate the power generation of the cascade power station at each time point. In order to maximize the output potential of the hydropower station, calculate the maximum output of each cascade power station. After obtaining the power generation of the cascade power station and the maximum output of the cascade power station, calculate the dispatch constraints of each power station. The calculation expression is as follows:

[0033] C i (t) = P max，i (t)-Pi (t),

[0034] In the formula, C i (t) is the dispatch constraint of the i-th cascade power station, P max , i(t) is the maximum output of the i-th cascade power station at time t, P i (t) is the actual power generation of the i-th cascade power station at time t;

[0035] Through the dynamic programming algorithm, combined with the grid load demand and the dispatch constraints of each power station, the output distribution of each cascade power station is optimized to achieve global optimal dispatch and ensure grid load balance and safe and stable operation of the power system.

[0036] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0037] By introducing multi-source data fusion and deep learning models (such as LSTM), the present invention can accurately predict the key water level of the reservoir. This accurate prediction not only relies on the input of real-time meteorological and hydrological data, but can also dynamically adapt to the timing characteristics of water level changes, especially in the face of extreme meteorological events. Compared with traditional water level prediction methods, the present invention can better capture the nonlinear relationship of influencing factors and reduce prediction errors, especially in complex environments (such as high-cold areas, dry seasons, etc.) to maintain a high degree of accuracy. Real-time prediction and scheduling optimization can reduce the phenomenon of water abandonment in hydropower stations, improve the efficiency of electricity use, ensure that changes in reservoir water levels match the load demand of the power grid, and enhance the overall flexibility and responsiveness of the power grid.

[0038] The present invention introduces an adaptive mechanism based on feedback adjustment in scheduling, which can adjust the scheduling plan in real time according to the error between the reservoir water level and the prediction result. This mechanism effectively avoids the phenomenon of water abandonment due to model error or prediction delay. Especially when the water level is close to the upper limit, the scheduling system can adjust the output of the power station in advance, so that the hydropower resources can be fully utilized and the ineffective water abandonment can be reduced. At the same time, by optimizing the joint operation of the cascade power stations, the water flow between the reservoirs can be effectively regulated and compensated, thereby improving the overall output quality of electric energy. While reducing the amount of abandoned water, the stability and reliability of the hydropower system are enhanced, the power supply is made more stable and continuous, and the power quality of the power grid is ultimately improved.

[0039] The present invention combines the scheduling decision of deep reinforcement learning with intelligent optimization algorithms (such as genetic algorithms and particle swarm optimization). The present invention can dynamically adjust the joint scheduling scheme of hydropower cascade power stations according to real-time power grid load demand, meteorological and hydrological changes. The optimized scheduling scheme not only maximizes the use of hydropower resources, but also avoids the risk of power grid overload or insufficient power supply. By considering multi-dimensional constraints (such as maximum power station output, reservoir capacity limit, power grid load, etc.), the system can make the most appropriate scheduling decisions in a complex operating environment to ensure the safe and stable operation of the power grid. In addition, the system can also adjust the scheduling strategy according to real-time feedback to ensure the rapid response and efficient response of the power grid in the face of emergencies (such as extreme weather, power station failures, etc.), further improving the emergency handling capability and system stability of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0041] Figure 1 The present invention is a method flow chart of the method for accessing dispatching parameters of cascade power stations and predicting influencing factors. DETAILED DESCRIPTION

[0042] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0043] The present invention provides Figure 1 The method for accessing the dispatching parameters of cascade power stations and predicting the influencing factors shown includes the following steps:

[0044] Obtain operating parameter data of each cascade power station, including but not limited to flow, output, reservoir capacity and water level;

[0045] The operating parameter data of each cascade power station is collected in real time through sensors and monitoring equipment, and transmitted to the dispatching system through wireless communication technology or Internet connection, realizing efficient transmission and centralized management of real-time data. The real-time data includes but is not limited to power station operation information such as flow, output, storage capacity and water level, ensuring the timeliness and accuracy of the data and ensuring that the dispatching system can quickly respond to changes in the operating status of the power station.

[0046] Collect multi-dimensional influencing factor data related to power station dispatching, the influencing factors include meteorological data, geographical data, and hydrological data, where meteorological data includes precipitation, temperature, wind speed, etc., and hydrological data includes water level, water flow, snow melt, etc.;

[0047] Meteorological data, geographic data and hydrological data are collected comprehensively through a variety of data sources, including meteorological stations, meteorological satellites, remote sensing systems and hydrological monitoring stations. The data are fused and cleaned to remove outliers and noise to improve prediction accuracy. In addition, by adopting data fusion technology and combining the temporal and spatial characteristics of different data sources, the collected data are more representative, thereby enhancing the reliability and accuracy of subsequent predictions.

[0048] Based on the difficulty of predicting the influencing factors, the correlation analysis between the influencing factors and the critical water level of the reservoir is carried out to identify and determine the most relevant influencing factors;

[0049] The correlation analysis between influencing factors and critical water levels of reservoirs adopts the weighted correlation analysis method. Different factors are assigned different weights according to the distribution characteristics of influencing factors in time and space. The weighted Pearson correlation coefficient is used to analyze the correlation between each influencing factor and the critical water level. By considering the importance of influencing factors, the accuracy of correlation analysis is improved. This method is suitable for areas with complex meteorological and hydrological characteristics, and improves the accuracy of influencing factor selection and reservoir critical water level prediction.

[0050] The regression analysis method is used to fit the nonlinear functional relationship between the most relevant influencing factors and the critical water level of the reservoir;

[0051] The regression analysis method adopts polynomial regression. When dealing with nonlinear relationships, a high-order polynomial regression model is used. By cross-validating and tuning historical data, the hyperparameters of the regression model are optimized to ensure that the fitted nonlinear function can adapt to water level prediction under different seasons and climate changes. In the regression process, the error function is minimized to ensure the high-precision prediction ability of the regression model in various practical scenarios.

[0052] The obtained nonlinear function relationship is integrated into the dispatching system of cascade power stations to predict the changes of key water levels in power station reservoirs in real time;

[0053] The dispatching system integrates artificial intelligence algorithms, including deep learning models, to analyze in real time the complex nonlinear relationship between influencing factors and critical water levels of reservoirs. Through continuous learning and optimization, it adaptively adjusts weights and corrects prediction errors based on feedback from historical data and real-time data to ensure the accuracy of reservoir dispatching. During the adjustment process, it takes into account the actual operating status of each cascade power station, while comprehensively considering various meteorological and hydrological changes to further improve the flexibility and response speed of the joint dispatching of hydropower cascades.

[0054] The deep learning model further includes a long short-term memory network (LSTM), which is used to process the temporal relationship between the influencing factors and the critical water level of the reservoir, and specifically includes the following steps:

[0055] By preprocessing historical meteorological, geographical and hydrological data, it is converted into time series data and input into the LSTM model for training;

[0056] During the training process of the LSTM network, the back-propagation algorithm is used to optimize the model parameters, calculate the error and adjust the network weights to achieve accurate modeling of nonlinear temporal relationships;

[0057] The trained LSTM model is used to predict the critical water level of the reservoir in real time, and the prediction results are input into the dispatching system as feedback, thereby optimizing the joint dispatching strategy of hydropower cascade power stations, reducing water abandonment and improving dispatching flexibility.

[0058] Based on the prediction results, the joint dispatching plan of the power station is dynamically adjusted to optimize the output regulation of the hydropower station, reduce water abandonment, and ensure the supply and demand balance of the power system and the stable operation of the power grid;

[0059] The specific steps of the joint dispatching scheme for dynamic adjustment of power stations are as follows:

[0060] Calculate the critical water level of each cascade power station reservoir and obtain the predicted water level based on the output of the regression model;

[0061] For each cascade power station, the dispatch deviation is calculated, and the calculation expression is as follows:

[0062] , where W i is the current critical water level of the reservoir of the i-th cascade power station, is the predicted critical water level of the reservoir of the i-th cascade power station calculated according to the regression model, D i is the dispatch deviation of the i-th cascade power station;

[0063] During the scheduling process, the feedback adjustment factor based on the prediction error is calculated, and the calculation expression is as follows:

[0064] ,

[0065] In the formula, F i (t) is the feedback adjustment factor, which is used to dynamically adjust the dispatching plan of the power station according to the prediction error. i (t-1) is the dispatch deviation of the i-th cascade power station at time t-1, α is the weight factor of the dispatch deviation, which controls the dispatch deviation D of the i-th power station itself. i (t-1) in the feedback adjustment factor Fi (t), β is the weighting factor of the dispatch deviation of other cascade power stations, which controls the influence of the dispatch deviation of other power stations on the current cascade power station i, w ij is the influence coefficient of cascade power station j on the dispatch of cascade power station i, and N is the total number of cascade power stations involved in the dispatch

[0066] According to the calculated feedback adjustment factor F i (t) Dynamically modify the dispatching plan to ensure the optimal dispatching of cascade power stations under different meteorological, hydrological and water level conditions, and ensure the regulation capacity and operational stability of the dispatching system.

[0067] The joint scheduling scheme further includes the following steps:

[0068] The power generation of the cascade power station at each time point is calculated. The relationship between this power and factors such as water level and flow is a nonlinear function. In order to maximize the output potential of the hydropower station, the maximum output of each cascade power station is calculated. After obtaining the power generation of the cascade power station and the maximum output of the cascade power station, the dispatch constraint of each power station is calculated. The calculation expression is as follows:

[0069] C i (t) = P max，i (t)-P i (t),

[0070] In the formula, C i (t) is the dispatch constraint of the i-th cascade power station, reflecting the gap between the power station output and the maximum output, P max，i (t) represents the maximum output of the i-th cascade power station at time t, P i (t) is the actual power generation of the i-th cascade power station at time t;

[0071] Through linear programming or dynamic programming algorithms, combined with the grid load demand and the dispatch constraints of each power station, the output distribution of each cascade power station is optimized to achieve global optimal dispatch and ensure grid load balance and safe and stable operation of the power system.

[0072] Specific implementation method 1: In order to improve the prediction accuracy of key water levels in reservoirs, this implementation method adopts multi-source data fusion technology, combines meteorological data, hydrological data and geographical environmental factors, and accurately predicts influencing factors, thereby effectively solving the error problem of water level prediction in cascade power station scheduling.

[0073] First, the reservoir dispatching system collects the operating parameters of each cascade power station in real time through sensors, including flow, water level, output and storage capacity. These parameters are transmitted to the data processing center and connected to the dispatching system in real time through a wireless communication network. At the same time, multi-source data platforms such as meteorological stations, hydrological stations and satellite remote sensing systems also provide relevant influencing factor data. Meteorological data include precipitation, temperature, wind speed, etc.; hydrological data include water level, water flow, snow melt, etc.; and geographic data involves environmental factors such as altitude, slope, soil type, etc. in the reservoir area.

[0074] In the data preprocessing stage, the collected multi-source data is first cleaned and denoised to remove outliers and missing data. This process is completed through anomaly detection algorithms to ensure the quality of input data. In order to improve the timeliness of the data, data synchronization technology is also used to convert multi-source data into a unified time step so that all data can be analyzed on the same time scale. These preprocessed data will serve as the basis for subsequent analysis.

[0075] The selection of influencing factors is the key to prediction accuracy. To this end, a correlation analysis based on statistical analysis methods is used to select the most critical factors affecting reservoir water level changes. The correlation analysis of influencing factors is carried out by calculating the correlation coefficient between different meteorological and hydrological factors and the key water level of the reservoir. By using the weighted Pearson correlation coefficient method, each factor is weighted and calculated, especially under the influence of different time periods (such as seasonality and interannuality), the correlation of certain factors may change significantly. Therefore, this embodiment also introduces spatiotemporal weighting technology, so that under different seasons and different climate changes, the prediction model can effectively adjust the weights to reflect the influence of factors in different time dimensions.

[0076] Based on the results of correlation analysis, the influencing factors most closely related to the changes in the critical water level of the reservoir were selected. In order to establish a more accurate prediction model, this implementation adopts the long short-term memory network (LSTM) in deep learning to model the changes in the critical water level of the reservoir. The LSTM network is a deep neural network that processes time series data. It can capture the time dependence of reservoir water level changes and effectively avoid the gradient vanishing problem of traditional RNN in long time series data processing.

[0077] During the training process of the LSTM model, historical meteorological and hydrological data and real-time operation data of the reservoir are used as input to train the network to learn the temporal characteristics of water level changes. Since LSTM can capture long-term dependencies, the network can identify potential patterns that affect water level changes during learning, and then predict future water level changes. Cross-validation technology is used during model training to ensure that the trained model has good generalization capabilities.

[0078] Once the LSTM model is trained, it can be connected to the dispatching system of the cascade power station in real time, and the water level changes of the reservoir in the future can be predicted by obtaining meteorological and hydrological data in real time. When the predicted water level change deviates greatly from the actual water level, the system will issue an alarm and automatically adjust the parameters in the prediction model to reduce the error.

[0079] In addition, the system also has an adaptive adjustment mechanism, that is, if there is a large error between real-time data and predicted data, the scheduling system will correct the LSTM model and optimize the model parameters. This process is achieved through a feedback loop to ensure that the prediction results output by the model each time are more in line with the actual situation, thereby improving the accuracy and flexibility of scheduling.

[0080] Through real-time predicted water level data, the dispatching system can dynamically adjust the dispatching plan of the power station. For example, when it is predicted that the water level will reach the upper limit in the future, the system will automatically calculate and start the dispatching strategy to reduce the risk of reservoir overflow and avoid the waste of electricity caused by excessive water storage. On the contrary, if the system predicts that the water level is low and the future precipitation is not enough to replenish the water volume of the reservoir, the dispatching system will start the adjustment plan to increase the output of the power station.

[0081] Through the influencing factor prediction method of multi-source data fusion, power station scheduling can be flexibly adjusted under the conditions of changes in multiple factors such as meteorology and hydrology, effectively solving the problem of prediction error in the reservoir scheduling process in high-altitude and cold areas and improving the utilization efficiency of water resources.

[0082] Specific implementation method 2: In this implementation method, a cascade power station joint dispatch optimization method based on adaptive feedback adjustment is proposed for the nonlinear relationship modeling problem between influencing factors and critical water levels of reservoirs. This method optimizes the prediction of critical water levels of reservoirs through a real-time feedback system and dynamically adjusts the dispatch strategy to achieve more accurate dispatch.

[0083] First, the nonlinear relationship between the influencing factors and the critical water level of the reservoir is modeled using the regression analysis method through the existing meteorological, hydrological and power station operation data. Traditional regression models such as linear regression and support vector regression (SVR) often cannot capture the complex nonlinear relationship between factors. Therefore, this embodiment uses polynomial regression or support vector regression methods for modeling, which can perform more accurate nonlinear fitting under complex data.

[0084] During the regression modeling process, the system will use historical data for training to build a prediction model that can predict future reservoir water levels based on input data such as meteorological factors, geographical factors, and hydrological factors. It should be noted that during the model training process, the weights of influencing factors will be adjusted for different seasons and different climate conditions to ensure its accuracy in different time periods.

[0085] In actual operation, the system will obtain the operating data of each cascade power station and external meteorological and hydrological data in real time, and predict the critical water level of the reservoir based on the existing regression model. When there is a large deviation between the predicted water level and the actual water level, the dispatching system will trigger the feedback adjustment mechanism. Specifically, the system will calculate the prediction error and feed the error information back to the regression model, and optimize the prediction accuracy of the model by adjusting the parameters in the regression model.

[0086] The core of the feedback mechanism lies in its adaptability: the system not only adjusts the model based on the real-time prediction error, but also adjusts it based on the effect of historical scheduling data. If the historical error is large, the system will increase the learning rate of the model to make the prediction adjustment more sensitive; otherwise, it will appropriately reduce the adjustment range to ensure that the system operates in a stable state.

[0087] Based on the real-time water level prediction and feedback adjustment mechanism, the dispatching system can intelligently optimize the operation plan of each cascade power station. For example, when the water level is predicted to rise to the dangerous range, the system will dispatch the upstream power station in advance and allocate the output of the reservoir to avoid the risk of overflow caused by excessive water level in the downstream power station. If it is predicted that the water level of a cascade power station will drop too low and there is not enough precipitation to replenish it, the system will start the backup dispatch of other power stations to ensure the stability of power supply.

[0088] This scheduling method not only reduces water abandonment, but also improves the system's responsiveness to extreme weather events such as heavy rains or droughts. Through the adaptive feedback mechanism, the scheduling system can maintain efficient and stable operation in a changing environment.

[0089] Specific implementation method three: This implementation method proposes a joint scheduling method for cascade power stations combined with an intelligent optimization algorithm. This method calculates the optimal scheduling plan based on the real-time load demand, meteorological data and reservoir water level of the power station through a genetic algorithm (GA) or a particle swarm optimization (PSO) algorithm.

[0090] First, the system needs to build a multi-objective optimization model, which aims to maximize the utilization of hydropower resources, minimize the amount of abandoned water, and ensure the load balance of the power grid. The constraints of the optimization model include the maximum output and minimum output of the power station, the safe load range of the power grid, and the reservoir capacity limit of each power station. By setting these objective functions and constraints, the model can automatically find the optimal scheduling plan under various constraints.

[0091] Next, the system uses a genetic algorithm (GA) or a particle swarm optimization (PSO) algorithm to search for the optimal scheduling solution. The genetic algorithm simulates the natural selection process, selects excellent scheduling solutions for crossover and mutation, and gradually approaches the optimal solution. The particle swarm optimization algorithm simulates the foraging behavior of bird flocks to search for the optimal point in the optimal solution space. Both algorithms have global search capabilities, which can avoid falling into local optimal solutions and find the most suitable solution for the current power grid operation.

[0092] In order to improve the optimization efficiency, this implementation also combines the deep reinforcement learning (DQN) algorithm. DQN automatically adjusts the scheduling parameters and gradually improves the scheduling strategy by learning from historical scheduling results. The system stability, abandoned hydropower, load balance and other effects after each scheduling adjustment will serve as a reward signal for reinforcement learning to drive the system to learn the best scheduling strategy.

[0093] The optimal dispatching plan calculated by the optimization algorithm will be fed back to the control system of the cascade power station in real time, automatically adjusting the output of the power station and sending dispatching signals to the power grid dispatching center in a timely manner. The system also has a real-time monitoring function, which can flexibly adjust the dispatching plan according to changes in power grid load and emergencies (such as weather changes, equipment failures, etc.) to ensure the safety and stability of the power grid.

[0094] By introducing multi-source data fusion and deep learning models (such as LSTM), the present invention can accurately predict the key water level of the reservoir. This accurate prediction not only relies on the input of real-time meteorological and hydrological data, but can also dynamically adapt to the timing characteristics of water level changes, especially in the face of extreme meteorological events. Compared with traditional water level prediction methods, the present invention can better capture the nonlinear relationship of influencing factors and reduce prediction errors, especially in complex environments (such as high-cold areas, dry seasons, etc.) to maintain a high degree of accuracy. Real-time prediction and scheduling optimization can reduce the phenomenon of water abandonment in hydropower stations, improve the efficiency of electricity use, ensure that changes in reservoir water levels match the load demand of the power grid, and enhance the overall flexibility and responsiveness of the power grid.

[0095] The present invention introduces an adaptive mechanism based on feedback adjustment in scheduling, which can adjust the scheduling plan in real time according to the error between the reservoir water level and the prediction result. This mechanism effectively avoids the phenomenon of water abandonment due to model error or prediction delay. Especially when the water level is close to the upper limit, the scheduling system can adjust the output of the power station in advance, so that the hydropower resources can be fully utilized and the ineffective water abandonment can be reduced. At the same time, by optimizing the joint operation of the cascade power stations, the water flow between the reservoirs can be effectively regulated and compensated, thereby improving the overall output quality of electric energy. While reducing the amount of abandoned water, the stability and reliability of the hydropower system are enhanced, the power supply is made more stable and continuous, and the power quality of the power grid is ultimately improved.

[0096] The present invention combines the scheduling decision of deep reinforcement learning with intelligent optimization algorithms (such as genetic algorithms and particle swarm optimization). The present invention can dynamically adjust the joint scheduling scheme of hydropower cascade power stations according to real-time power grid load demand, meteorological and hydrological changes. The optimized scheduling scheme not only maximizes the use of hydropower resources, but also avoids the risk of power grid overload or insufficient power supply. By considering multi-dimensional constraints (such as maximum power station output, reservoir capacity limit, power grid load, etc.), the system can make the most appropriate scheduling decisions in a complex operating environment to ensure the safe and stable operation of the power grid. In addition, the system can also adjust the scheduling strategy according to real-time feedback to ensure the rapid response and efficient response of the power grid in the face of emergencies (such as extreme weather, power station failures, etc.), further improving the emergency handling capability and system stability of the power grid.

[0097] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0098] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0099] It should be noted that, in this article, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0100] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0101] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0102] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0103] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0104] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0105] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0106] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for accessing dispatching parameters and predicting influencing factors of cascade power stations, characterized in that: The following steps are involved: Obtain operating parameter data of each cascade power station, including but not limited to flow, output, reservoir capacity and water level; Collect multi-dimensional influencing factor data related to power station dispatching, wherein the influencing factors include meteorological data, geographical data, and hydrological data; Based on the difficulty of predicting the influencing factors, the correlation analysis between the influencing factors and the critical water level of the reservoir is carried out to identify and determine the most relevant influencing factors; The regression analysis method is used to fit the nonlinear functional relationship between the most relevant influencing factors and the critical water level of the reservoir; The obtained nonlinear function relationship is integrated into the dispatching system of cascade power stations to predict the changes of key water levels in power station reservoirs in real time; Based on the prediction results, the joint dispatching plan of the power station is dynamically adjusted to optimize the output regulation of the hydropower station, reduce water abandonment, and ensure the supply and demand balance of the power system and the stable operation of the power grid.

2. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1 is characterized by: The operating parameter data of each cascade power station is collected in real time through sensors and monitoring equipment, and transmitted to the dispatching system through wireless communication technology or Internet connection, realizing efficient transmission and centralized management of real-time data, ensuring the timeliness and accuracy of the data, and ensuring that the dispatching system can quickly respond to changes in the operating status of the power station.

3. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1 is characterized in that: Meteorological data, geographic data and hydrological data are collected comprehensively through a variety of data sources, including meteorological stations, meteorological satellites, remote sensing systems and hydrological monitoring stations. The data are fused and cleaned to remove outliers and noise to improve prediction accuracy. In addition, by adopting data fusion technology and combining the temporal and spatial characteristics of different data sources, the collected data are more representative, thereby enhancing the reliability and accuracy of subsequent predictions.

4. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1, characterized in that: The correlation analysis between influencing factors and critical water levels of reservoirs adopts the weighted correlation analysis method. Different factors are assigned different weights according to the distribution characteristics of influencing factors in time and space. The weighted Pearson correlation coefficient is used to analyze the correlation between each influencing factor and the critical water level. By considering the importance of influencing factors, the accuracy of correlation analysis is improved. This method is suitable for areas with complex meteorological and hydrological characteristics, and improves the accuracy of influencing factor selection and reservoir critical water level prediction.

5. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1 is characterized by: The regression analysis method adopts polynomial regression. When dealing with nonlinear relationships, a high-order polynomial regression model is used. By cross-validating and tuning historical data, the hyperparameters of the regression model are optimized to ensure that the fitted nonlinear function can adapt to water level prediction under different seasons and climate changes. In the regression process, the error function is minimized to ensure the high-precision prediction ability of the regression model in various practical scenarios.

6. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1, characterized in that: The dispatching system integrates artificial intelligence algorithms, including deep learning models, to analyze in real time the complex nonlinear relationship between influencing factors and critical water levels of reservoirs. Through continuous learning and optimization, it adaptively adjusts weights and corrects prediction errors based on feedback from historical data and real-time data to ensure the accuracy of reservoir dispatching. During the adjustment process, it takes into account the actual operating status of each cascade power station, while comprehensively considering various meteorological and hydrological changes to further improve the flexibility and response speed of the joint dispatching of hydropower cascades.

7. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1, characterized in that: The deep learning model further includes a long short-term memory network, which is used to process the temporal relationship between the influencing factors and the critical water level of the reservoir, and specifically includes the following steps: By preprocessing historical meteorological, geographical and hydrological data, it is converted into time series data and input into the LSTM model for training; During the training process of the LSTM network, the back-propagation algorithm is used to optimize the model parameters, calculate the error and adjust the network weights to achieve accurate modeling of nonlinear temporal relationships; The trained LSTM model is used to predict the critical water level of the reservoir in real time, and the prediction results are input into the dispatching system as feedback, thereby optimizing the joint dispatching strategy of hydropower cascade power stations, reducing water abandonment and improving dispatching flexibility.

8. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1, characterized in that: The specific steps of the joint dispatching scheme for dynamic adjustment of power stations are as follows: Calculate the critical water level of each cascade power station reservoir and obtain the predicted water level based on the output of the regression model; For each cascade power station, the dispatch deviation is calculated, and the calculation expression is as follows: ， Where W i is the current critical water level of the reservoir of the i-th cascade power station, is the predicted critical water level of the reservoir of the i-th cascade power station calculated according to the regression model, D i is the dispatch deviation of the i-th cascade power station; During the scheduling process, the feedback adjustment factor based on the prediction error is calculated, and the calculation expression is as follows: ， In the formula, F i (t) is the feedback adjustment factor, which is used to dynamically adjust the dispatching plan of the power station according to the prediction error. i (t-1) is the dispatch deviation of the i-th cascade power station at time t-1, α is the weight factor of the dispatch deviation, β is the weight factor of the dispatch deviation of other cascade power stations, and w ij is the influence coefficient of cascade power station j on the dispatch of cascade power station i, and N is the total number of cascade power stations involved in the dispatch According to the calculated feedback adjustment factor F i (t) Dynamically modify the dispatching plan to ensure the optimal dispatching of cascade power stations under different meteorological, hydrological and water level conditions, and ensure the regulation capacity and operational stability of the dispatching system.

9. The method for accessing dispatching parameters and predicting influencing factors of cascade power stations according to claim 1, characterized in that: The joint scheduling scheme further includes the following steps: Calculate the power generation of the cascade power station at each time point. In order to maximize the output potential of the hydropower station, calculate the maximum output of each cascade power station. After obtaining the power generation of the cascade power station and the maximum output of the cascade power station, calculate the dispatch constraints of each power station. The calculation expression is as follows: C i (t)=P max ,i(t)-P i (t), In the formula, C i (t) is the dispatch constraint of the i-th cascade power station, P max，i (t) represents the maximum output of the i-th cascade power station at time t, P i (t) is the actual power generation of the i-th cascade power station at time t; Through the dynamic programming algorithm, combined with the grid load demand and the dispatch constraints of each power station, the output distribution of each cascade power station is optimized to achieve global optimal dispatch and ensure grid load balance and safe and stable operation of the power system.

Images