New energy equipment interconnection AI energy intelligent management monitoring system

Abstract

The invention discloses a new energy equipment interconnection AI energy intelligent management monitoring system, and relates to the technical field of energy scheduling. Preprocessing the weather data to obtain effective weather data; determining operation parameters of a first preset model according to the effective weather data; updating a first preset model according to the operation parameters, and substituting the effective weather data into the updated first preset model to obtain first generating capacity prediction; determining first electricity consumption prediction according to the historical electricity consumption and the current electricity consumption; and if the first power generation prediction is smaller than the first power consumption prediction, performing optimization scheduling on the power of the power grid, the new energy equipment and the energy storage equipment according to the power grid electricity price, so that the power consumption in the target power consumption area is met. New energy power generation and regional power consumption demands are respectively and accurately predicted through power generation amount prediction and power consumption prediction, and finally the power grid, new energy and energy storage equipment power are optimized and dispatched in combination with the power grid electricity price, so that the problem of low efficiency under complex energy management demands is solved, and the energy management efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the field of energy dispatching technology, and specifically relates to an AI energy intelligent management and monitoring system for interconnected new energy equipment. Background Art

[0002] With the growing global demand for clean energy, renewable energy sources, such as wind power and photovoltaics, are increasingly accounting for a larger share of the energy mix. While the booming new energy industry is helping to alleviate pressure on traditional energy supply, renewable energy generation suffers from significant volatility and intermittency. Its output is significantly affected by weather factors such as sunlight intensity, wind speed, and temperature, making it difficult to achieve stable output. For example, photovoltaic power generation plummets on cloudy days and at night, while wind power generation is constrained by seasonal wind speed fluctuations. This poses significant challenges to the stability of power supply. Furthermore, user-side electricity demand presents complex uncertainties. Electricity load is affected by factors such as seasonal variations, weather conditions, and user behavior patterns. Traditional simple statistical or linear regression forecasting methods based on historical data are unable to accurately capture these dynamic changes in electricity load.

[0003] At the energy dispatch level, existing energy management systems often face the dilemma of energy supply and demand imbalances due to their lack of ability to accurately predict renewable energy generation and electricity demand. When renewable energy generation is insufficient, it is difficult to rationally adjust the power supply ratio between the power grid and energy storage equipment, resulting in high electricity costs or reduced power supply reliability. Conversely, when there is excess power generation, energy storage equipment cannot be effectively used to store excess electricity, resulting in energy waste and economic losses. Although some existing systems have attempted to introduce intelligent algorithms for scheduling, due to incomplete data integration and poor model adaptability, they still cannot meet the complex and changing energy management needs.

[0004] Patent CN119134643A discloses an AI-powered intelligent energy management and monitoring system and method for interconnected new energy devices. The system comprises a photovoltaic device connected to a photovoltaic panel, a storage battery, a charging pile, a smart meter, a diesel engine, and a smart charging socket; a charging pile connected to an electric vehicle; a smart meter connected to the power grid; and a smart socket connected to a water heater and common appliances. The photovoltaic device, storage battery, charging pile, smart meter, diesel engine, and smart charging socket are connected to a cloud server via a router. This invention proposes an AI-powered intelligent energy management and monitoring system for interconnected new energy devices. This system can double hardware sales, provide warranty renewal services for users whose devices have expired, ensure user usage, and increase renewal service fees, improving user experience. The system continuously saves electricity for users, and offers users an optimized electricity strategy service through a paid subscription, increasing software revenue beyond hardware.

[0005] However, this solution still cannot meet the complex energy management needs when it comes to energy scheduling, resulting in low energy management efficiency. Summary of the Invention

[0006] The purpose of this invention is to solve the problem that when facing energy scheduling, complex energy management needs cannot be met, resulting in low energy management efficiency, and to propose an AI energy intelligent management and monitoring system for interconnected new energy equipment.

[0007] The present invention proposes an AI energy intelligent management and monitoring system for interconnected new energy equipment, the system comprising: A preprocessing module is used to obtain weather data of the next period at the current moment and preprocess the weather data to obtain valid weather data; an operating parameter determination module, configured to determine operating parameters of a first preset model based on the valid weather data; a power generation prediction and determination module, configured to update the first preset model according to the operating parameters to obtain an updated first preset model, and substitute the valid weather data into the updated first preset model to obtain a first power generation prediction of the new energy device in the next cycle; A first power consumption prediction and determination module is configured to obtain historical power consumption and current power consumption within a target power consumption area, and substitute the historical power consumption and the current power consumption into a second preset model to obtain a first power consumption prediction for the next cycle at the current moment; The power dispatching module is used to optimize the power dispatching of the power grid, new energy equipment and energy storage equipment according to the power grid electricity price if the first power generation forecast is less than the first power consumption forecast, so as to meet the power consumption in the target power consumption area.

[0008] Optionally, the operating parameter determination module includes: A feature data determination module is used to extract feature data of the valid weather data; the feature data includes DNI data, GHI data, temperature data and wind speed data; An operating parameter generation module is used to determine the target type corresponding to the current weather by combining the characteristic data with a clustering algorithm, and substitute the target type and the weather data into a parameter optimization algorithm to obtain the operating parameters of the first preset model.

[0009] Optionally, the first preset model includes a gated recurrent unit, an informer model, and a support vector regression model; and the power generation prediction and determination module includes: A feature extraction module, configured to substitute the weather data into the gated recurrent unit and the informer model to obtain training features and prediction features; The power generation prediction generation module is used to substitute the training features and the prediction features into the support vector regression model to obtain the power generation prediction of the new energy equipment in the next cycle.

[0010] Optionally, the first power consumption prediction and determination module includes: A historical electricity consumption screening module is used to screen the historical electricity consumption data according to the current weather conditions to obtain valid historical electricity consumption data; An effective feature determination module, configured to process the effective historical power consumption data by a backward elimination method to obtain effective features; A target LSTM model generation module is used to substitute the effective features into the LSTM model and optimize the LSTM model through Bayesian optimization to obtain a target LSTM model; The preprocessing module is used to substitute the current power consumption into the target LSTM model to obtain the first power consumption prediction of the next cycle at the current moment.

[0011] Optionally, the power scheduling module includes: a target demand determination module, configured to calculate a difference between the first power consumption forecast and the first power generation forecast to obtain a target demand; The grid electricity price determination module is used to obtain the current grid electricity price and determine the grid electricity price for the next cycle at the current moment based on historical grid electricity price fluctuations; The first power supply module is used to connect to the power grid if the power grid electricity price of the next cycle at the current moment is greater than the current power grid electricity price, and supply power to the target power consumption area through the power grid.

[0012] Optionally, the system further includes: The second power supply module is used to supply power to the target power consumption area through the energy storage device if the power grid electricity price of the next cycle at the current moment is less than or equal to the current power grid electricity price.

[0013] Optionally, the system further includes: a surplus power determination module, configured to calculate the difference between the first power generation prediction and the first power consumption prediction to obtain surplus power if the first power generation prediction is greater than the first power consumption prediction; An energy storage device power supply module, configured to determine whether the energy storage device is fully charged, and if not, to add the excess power to the energy storage device; The electricity selling module is used to sell the excess electricity if there is still excess electricity after adding to the energy storage device.

[0014] Beneficial effects of the present invention: The present invention proposes an AI energy intelligent management and monitoring system for interconnected new energy equipment, including a preprocessing module for obtaining weather data of the next period at the current moment, and preprocessing the weather data to obtain valid weather data; an operating parameter determination module for determining the operating parameters of a first preset model according to the valid weather data; a power generation prediction determination module for updating the first preset model according to the operating parameters to obtain an updated first preset model, and substituting the valid weather data into the updated first preset model to obtain a first power generation prediction of the new energy equipment in the next period; a first power consumption prediction determination module for obtaining historical power consumption and current power consumption in a target power consumption area, and substituting the historical power consumption and current power consumption into a second preset model to obtain a first power consumption prediction of the next period at the current moment; and a power scheduling module for optimizing the power scheduling of the power grid, new energy equipment and energy storage equipment according to the power grid electricity price if the first power generation prediction is less than the first power consumption prediction, so as to meet the power consumption in the target power consumption area. The preprocessing module cleans the weather data of the next cycle to obtain valid data, and the operating parameter determination module determines the model parameters accordingly. The power generation and power consumption prediction and determination modules accurately predict the renewable energy power generation and regional power demand respectively. Finally, when power generation is less than power consumption, the power scheduling module optimizes the power scheduling of the power grid, renewable energy and energy storage equipment in combination with the power grid electricity price, thereby solving the problem of low efficiency under complex energy management needs and improving the efficiency of energy management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 A framework diagram of an AI energy intelligent management and monitoring system for interconnected new energy equipment provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0018] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of the present invention.

[0019] The embodiment of the present invention provides an AI energy intelligent management and monitoring system for interconnected new energy equipment. Figure 1 , Figure 1 The present invention provides a framework diagram of an AI-powered intelligent energy management and monitoring system for interconnected new energy devices. The system includes the following steps: The preprocessing module is used to obtain the weather data of the next cycle at the current moment and preprocess the weather data to obtain valid weather data; An operating parameter determination module, configured to determine operating parameters of a first preset model based on valid weather data; A power generation prediction and determination module is configured to update the first preset model according to the operating parameters to obtain an updated first preset model, and substitute valid weather data into the updated first preset model to obtain a first power generation prediction of the new energy equipment in the next cycle; A first power consumption prediction and determination module is used to obtain historical power consumption and current power consumption in the target power consumption area, substitute the historical power consumption and current power consumption into a second preset model to obtain a first power consumption prediction for the next cycle at the current moment; The power dispatching module is used to optimize the power dispatch of the power grid, new energy equipment and energy storage equipment according to the power grid electricity price if the first power generation forecast is less than the first power consumption forecast, so as to meet the power consumption in the target power consumption area.

[0020] Based on an AI energy intelligent management and monitoring system for interconnected new energy equipment provided by an embodiment of the present invention, a preprocessing module cleans the weather data of the next cycle to obtain valid data, an operating parameter determination module determines the model parameters accordingly, and a power generation and power consumption prediction and determination module accurately predicts new energy power generation and regional power demand respectively. Finally, when power generation is less than power consumption, the power scheduling module optimizes the scheduling of power grid, new energy and energy storage equipment power in combination with the power grid electricity price, thereby solving the problem of low efficiency under complex energy management needs and improving the efficiency of energy management.

[0021] In one implementation, the raw weather data is cleaned and preprocessed to remove outliers and missing values, ensuring that the weather data input into the model (such as DNI, GHI, temperature, wind speed, etc.) is authentic and reliable, and avoiding prediction bias caused by "junk data." Preprocessing is used to extract valid weather data (such as solar radiation intensity and wind speed) that is strongly correlated with renewable energy power generation (photovoltaic and wind power), reducing redundant information interference and improving the efficiency of subsequent model training and prediction.

[0022] In one implementation, historical feature data (DNI, GHI, temperature, wind speed) is divided into different weather types (sunny, rainy, and cloudy) through clustering algorithms (such as K-means and DBSCAN). For each weather type, a parameter optimization algorithm (particle swarm optimization PSO) is used to automatically adjust the parameters of the first preset model. Once the real-time weather matches the target type, the model parameters are immediately updated to the optimal solution for that category, avoiding prediction bias caused by static parameters.

[0023] In one implementation, the first preset model (photovoltaic and wind power prediction model) is dynamically updated based on real-time weather data to ensure that the model is always trained based on the latest environmental conditions, avoiding model inaccuracies caused by seasonal changes; the updated model is used to predict the renewable energy power generation in the next cycle, providing a reliable basis for power dispatch; and the decline in power generation of photovoltaic power stations in rainy weather is predicted in advance to avoid power supply gaps caused by over-reliance on renewable energy.

[0024] In one implementation, based on the historical electricity consumption and current electricity consumption data of the target area, the electricity consumption pattern is captured through a second preset model to achieve a refined prediction of the electricity demand for the next cycle; the next cycle is determined by technical personnel.

[0025] In one implementation, when the renewable energy power generation (first power generation forecast) is less than the electricity demand (first electricity consumption forecast), based on the fluctuation of grid electricity prices, priority is given to using grid power during low-price periods or low-price electricity stored in energy storage devices to reduce electricity costs; if the grid electricity price is higher in the next cycle, the power of the energy storage device is released in advance to reduce the demand for high-priced electricity; if the grid electricity price is lower, electricity is purchased directly from the grid, while retaining energy storage capacity for use during high-price periods.

[0026] In one embodiment, the operating parameter determination module includes: A feature data determination module is used to extract feature data of valid weather data; the feature data includes DNI data, GHI data, temperature data and wind speed data; The operating parameter generation module is used to determine the target type corresponding to the current weather by combining the characteristic data with the clustering algorithm, and substitute the target type and weather data into the parameter optimization algorithm to obtain the operating parameters of the first preset model.

[0027] In one implementation, four key features are extracted from valid weather data: DNI (direct normal irradiance): solar radiation directly irradiating vertical surfaces, which determines the actual power generation efficiency of photovoltaic modules; GHI (global horizontal irradiance): total solar radiation received by horizontal surfaces, used to calculate the theoretical output of photovoltaic systems; temperature data: affects the efficiency of photovoltaic panels and the wind turbine power curve; wind speed data: the core input for wind power prediction, which has a nonlinear relationship with wind turbine output. The selected features are all strongly correlated with the renewable energy power generation process to avoid the introduction of redundant data.

[0028] In one embodiment, the first preset model includes a gated recurrent unit, an informer model, and a support vector regression model; the power generation prediction and determination module includes: The feature extraction module is used to substitute weather data into the access control recurrent unit and the informer model to obtain training features and prediction features; The power generation prediction generation module is used to substitute the training features and prediction features into the support vector regression model to obtain the power generation prediction of the new energy equipment in the next cycle.

[0029] In one implementation, the gated recurrent unit is a variant of the recurrent neural network (RNN) and is good at handling short-term dependency problems; the informer model is a long-sequence prediction model based on the Transformer architecture, designed to address the performance bottleneck of traditional models in long-term modeling; support vector regression is a nonlinear regression model used to fit the complex mapping relationship between weather characteristics and power generation; GRU processing is used to obtain short-term training features, and informer processing is used to obtain long-term prediction features. Short-term features reflect immediate impacts, while long-term features reflect cumulative effects. The combination of the two can more comprehensively describe the mechanism of weather effects on power generation.

[0030] In one implementation, the training features output by the GRU and the prediction features output by the Informer are concatenated into a comprehensive feature vector, which is then input into the SVR model for regression modeling to output the final power generation forecast. The SVR can fit the non-monotonic relationship between weather characteristics and power generation. The GRU and Informer process data in parallel, allowing the SVR to quickly complete feature fusion. The overall prediction delay can be controlled at the second level, meeting real-time scheduling requirements.

[0031] In one embodiment, the first power consumption prediction and determination module includes: The historical electricity consumption screening module is used to screen the historical electricity consumption data according to the current weather conditions to obtain valid historical electricity consumption data; An effective feature determination module is used to process the effective historical electricity consumption data through a backward elimination method to obtain effective features; The target LSTM model generation module is used to substitute valid features into the LSTM model and optimize the LSTM model through Bayesian optimization to obtain the target LSTM model; The preprocessing module is used to substitute the current power consumption into the target LSTM model to obtain the first power consumption forecast for the next cycle at the current moment.

[0032] In one implementation, based on the current weather environment, valid data with similar weather conditions are filtered out from historical electricity consumption data; for example, if the current weather is sunny, the season is summer, and the date is Friday, electricity consumption data under the same conditions in the historical data will be obtained.

[0033] In one implementation, the historical electricity consumption is historical data for the past five years.

[0034] In one implementation method, the backward elimination method is used to perform feature screening on the effective historical electricity consumption data, thereby further screening and selecting data that better meets the conditions. The streamlined feature set makes the model more stable in prediction under certain scenarios.

[0035] In one implementation, a long short-term memory network (LSTM) is used to capture the time series characteristics of electricity consumption data, and the LSTM model hyperparameters (such as the number of layers, number of neurons, learning rate, dropout rate, etc.) are automatically searched to maximize the prediction accuracy through the probabilistic model, thereby improving the accuracy of the prediction.

[0036] In one embodiment, the power scheduling module includes: a target demand determination module, configured to calculate the difference between the first power consumption forecast and the first power generation forecast to obtain the target demand; The grid electricity price determination module is used to obtain the current grid electricity price and determine the grid electricity price for the next cycle at the current moment based on historical grid electricity price fluctuations; The first power supply module is used to connect to the power grid if the power grid electricity price of the next cycle at the current moment is greater than the current power grid electricity price, and supply power to the target power consumption area through the power grid.

[0037] In one implementation, the difference between electricity demand and renewable energy power generation is calculated (target demand = first electricity demand forecast - first power generation forecast). If the result is positive, it indicates that electricity needs to be purchased from the grid to fill the gap; if it is negative, it indicates that there is excess electricity that can be stored or sold. Through the difference calculation, renewable energy and energy storage are prioritized, and electricity is purchased only when necessary, reducing dependence on the grid.

[0038] In one implementation method, when it is predicted that the grid electricity price in the next cycle will be greater than the current electricity price (such as when the electricity price shifts from a low-peak period to a peak period), the grid is connected in advance to purchase electricity, and the current low-priced electricity is used to meet part or all of the target demand, avoiding high-priced electricity purchases in the next cycle.

[0039] In one embodiment, the system further comprises: The second power supply module is used to supply power to the target power consumption area through the energy storage device if the power grid electricity price of the next cycle at the current moment is less than or equal to the current power grid electricity price.

[0040] In one implementation, when the grid electricity price in the next cycle is ≤ the current grid electricity price, energy storage equipment is used to supply power to the target area first. Before the next cycle's electricity price reaches a low point, energy storage is used in advance to supply power (consuming stored energy) to avoid consuming stored energy during the current high-price period. When the electricity price reaches a low point in the next cycle, electricity is purchased from the grid at a low price (either to supplement energy storage or directly supply power), thereby expanding the peak-valley profit space by using storage at a high price and purchasing electricity at a low price.

[0041] In one embodiment, the system further comprises: a surplus power determination module, configured to calculate the difference between the first power generation prediction and the first power consumption prediction to obtain surplus power if the first power generation prediction is greater than the first power consumption prediction; The energy storage device power supply module is used to determine whether the energy storage device is fully charged. If not, the excess power is added to the energy storage device; The electricity sales module is used to sell the excess electricity if there is still excess electricity after adding it to the energy storage device.

[0042] In one implementation, when the predicted power generation is less than the power consumption, the difference between the two (excess power) is calculated to provide a data basis for subsequent energy storage charging and power sales. The available excess renewable energy power is clearly quantified, and when the energy storage is not fully charged, the excess power is preferentially charged into the energy storage to improve the energy storage utilization rate; after it is fully charged, the power sales process is triggered.

[0043] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A new energy equipment interconnected AI energy intelligent management and monitoring system, characterized by: The system comprises: A preprocessing module is used to obtain weather data of the next period at the current moment and preprocess the weather data to obtain valid weather data; an operating parameter determination module, configured to determine operating parameters of a first preset model based on the valid weather data; a power generation prediction and determination module, configured to update the first preset model according to the operating parameters to obtain an updated first preset model, and substitute the valid weather data into the updated first preset model to obtain a first power generation prediction of the new energy device in the next cycle; A first power consumption prediction and determination module is configured to obtain historical power consumption and current power consumption within a target power consumption area, and substitute the historical power consumption and the current power consumption into a second preset model to obtain a first power consumption prediction for the next cycle at the current moment; The power dispatching module is used to optimize the power dispatching of the power grid, new energy equipment and energy storage equipment according to the power grid electricity price if the first power generation forecast is less than the first power consumption forecast, so as to meet the power consumption in the target power consumption area.

2. The AI ​​energy intelligent management and monitoring system for interconnected new energy equipment according to claim 1 is characterized in that: The operating parameter determination module includes: A feature data determination module is used to extract feature data of the valid weather data; the feature data includes DNI data, GHI data, temperature data and wind speed data; An operating parameter generation module is used to determine the target type corresponding to the current weather by combining the characteristic data with a clustering algorithm, and substitute the target type and the weather data into a parameter optimization algorithm to obtain the operating parameters of the first preset model.

3. The AI ​​energy intelligent management and monitoring system for interconnected new energy equipment according to claim 1 is characterized in that: The first preset model includes a gated recurrent unit, an informer model, and a support vector regression model; the power generation prediction and determination module includes: A feature extraction module, configured to substitute the weather data into the gated recurrent unit and the informer model to obtain training features and prediction features; The power generation prediction generation module is used to substitute the training features and the prediction features into the support vector regression model to obtain the power generation prediction of the new energy equipment in the next cycle.

4. The AI ​​energy intelligent management and monitoring system for interconnected new energy equipment according to claim 1 is characterized in that: The first power consumption prediction and determination module includes: A historical electricity consumption screening module is used to screen the historical electricity consumption data according to the current weather conditions to obtain valid historical electricity consumption data; An effective feature determination module, configured to process the effective historical power consumption data by a backward elimination method to obtain effective features; A target LSTM model generation module is used to substitute the effective features into the LSTM model and optimize the LSTM model through Bayesian optimization to obtain a target LSTM model; The preprocessing module is used to substitute the current power consumption into the target LSTM model to obtain the first power consumption prediction of the next cycle at the current moment.

5. The AI ​​energy intelligent management and monitoring system for interconnected new energy equipment according to claim 1 is characterized in that: The power scheduling module includes: a target demand determination module, configured to calculate a difference between the first power consumption forecast and the first power generation forecast to obtain a target demand; The grid electricity price determination module is used to obtain the current grid electricity price and determine the grid electricity price for the next cycle at the current moment based on historical grid electricity price fluctuations; The first power supply module is used to connect to the power grid if the power grid electricity price of the next cycle at the current moment is greater than the current power grid electricity price, and supply power to the target power consumption area through the power grid.

6. The AI ​​energy intelligent management and monitoring system for interconnected new energy equipment according to claim 1 is characterized in that: The system further comprises: The second power supply module is used to supply power to the target power consumption area through the energy storage device if the power grid electricity price of the next cycle at the current moment is less than or equal to the current power grid electricity price.

7. The AI ​​energy intelligent management and monitoring system for interconnected new energy equipment according to claim 1 is characterized in that: The system further comprises: a surplus power determination module, configured to calculate the difference between the first power generation prediction and the first power consumption prediction to obtain surplus power if the first power generation prediction is greater than the first power consumption prediction; An energy storage device power supply module, configured to determine whether the energy storage device is fully charged, and if not, to add the excess power to the energy storage device; The electricity selling module is used to sell the excess electricity if there is still excess electricity after adding to the energy storage device.

Images