CN-122022271-A - Method and device for dispatching park comprehensive energy system based on time sequence feature extraction and strategy optimization

CN122022271ACN 122022271 ACN122022271 ACN 122022271ACN-122022271-A

Abstract

The application provides a method and a device for dispatching a park comprehensive energy system based on time sequence feature extraction and strategy optimization, wherein the method comprises the steps of acquiring park energy state data of a plurality of historical moments, wherein the park energy state data reflects the output condition of clean energy in a park and the energy load condition in the park, and the energy load comprises electric load and thermal load; the method comprises the steps of inputting a plurality of park energy state data at historical moments into a trained time sequence feature extraction model to obtain time sequence features output by the time sequence feature extraction model, inputting input data into a strategy optimization model, determining an energy system scheduling strategy based on output data of the strategy optimization model, wherein the input data comprises the time sequence features and the park energy state data at the current moment, and the energy system scheduling strategy comprises demand response execution quantity and equipment adjustment quantity in a park energy system. The method and the system can improve the accuracy of the scheduling strategy of the park comprehensive energy system.

Inventors

LIU LISHAN
ZENG YUAN
SONG CHANGPENG
ZHANG LIDONG
ZHAO YUZE
XU XIANDONG

Assignees

天津大学

Dates

Publication Date: 20260512
Application Date: 20251231

Claims (10)

1. A park comprehensive energy system scheduling method based on time sequence feature extraction and strategy optimization is characterized by comprising the following steps: acquiring park energy state data of a plurality of historical moments, wherein the park energy state data reflects the output condition of clean energy in a park and the energy load condition in the park, and the energy load comprises electric load and thermal load; Inputting the park energy state data at the plurality of historical moments into a trained time sequence feature extraction model to obtain time sequence features output by the time sequence feature extraction model; and inputting input data into a strategy optimization model, and determining an energy system scheduling strategy based on output data of the strategy optimization model, wherein the input data comprises the time sequence characteristics and the park energy state data at the current moment, and the energy system scheduling strategy comprises a demand response execution amount and an equipment adjustment amount in the park energy system.
2. The method for scheduling a campus integrated energy system based on time series feature extraction and policy optimization of claim 1, wherein the campus energy status data comprises power, electrical load, thermal load, electricity price, state of charge of the energy storage system, indoor temperature, and outdoor temperature of the clean energy.
3. The method for scheduling a campus integrated energy system based on time sequence feature extraction and policy optimization according to claim 1, wherein the training process of the policy optimization model comprises: inputting the sample time sequence characteristics into the strategy optimization model, and determining a sample energy system scheduling strategy based on sample output data of the strategy optimization model; Determining evaluation scores of the sample energy system scheduling strategy in a plurality of indexes based on a state transition result corresponding to the sample energy system scheduling strategy; Weighting and fusing the evaluation scores of the multiple indexes to obtain an objective function value; updating the policy optimization model based on the objective function value.
4. The method for scheduling a campus integrated energy system based on time series feature extraction and policy optimization of claim 3, wherein the metrics comprise a first metric comprising a campus energy cost, a carbon emission, and a temperature bias.
5. The method for campus integrated energy system scheduling based on time series feature extraction and policy optimization of claim 3, wherein the metrics further comprise a second metric reflecting the degree of violation of the sample energy system scheduling policy with respect to the constraint condition of the energy system; The constraint conditions of the energy system comprise comfort constraint, supply and demand balance constraint of the electric energy source, supply and demand balance constraint of the heat energy source and operation constraint of a plurality of devices in the energy system.
6. The method for dispatching the campus integrated energy system based on time sequence feature extraction and strategy optimization according to claim 1, wherein the time sequence feature extraction model is trained based on a plurality of sets of training data, each set of training data comprises sample campus energy state data and a state prediction label corresponding to the sample campus energy state data, and the training process of the time sequence feature extraction model comprises the following steps: acquiring sample park energy state data, inputting the sample park energy state data into the time sequence feature extraction model, and acquiring sample time sequence features output by the time sequence feature extraction model; inputting the sample time sequence characteristics into a prediction model, and obtaining a state prediction result output by the prediction model; Determining training loss based on the state prediction result and a state prediction label corresponding to the sample park energy state data; Updating the timing feature extraction model based on the training loss; the method comprises the step of dynamically normalizing the sample park energy state data before the sample park energy state data is input into the time sequence feature extraction model.
7. A campus integrated energy system scheduling device based on time sequence feature extraction and policy optimization, the device comprising: The system comprises a historical data acquisition module, a storage module and a control module, wherein the historical data acquisition module is used for acquiring park energy state data at a plurality of historical moments, the park energy state data reflects the output condition of clean energy in a park and the energy load condition in the park, and the energy load comprises an electric load and a thermal load; The time sequence feature extraction module is used for inputting the park energy state data at the historical moments into the trained time sequence feature extraction model to obtain the time sequence features output by the time sequence feature extraction model; The scheduling strategy output module is used for inputting input data into a strategy optimization model, determining an energy system scheduling strategy based on the output data of the strategy optimization model, wherein the input data comprises the time sequence characteristics and the park energy state data at the current moment, and the energy system scheduling strategy comprises the demand response execution amount and the equipment adjustment amount in the park energy system.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the computer program, implements the campus integrated energy system scheduling method based on timing feature extraction and policy optimization of any one of claims 1 to 6.
9. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the campus integrated energy system scheduling method based on timing feature extraction and policy optimization of any one of claims 1 to 6.
10. A computer program product comprising a computer program which when executed by a processor implements a campus integrated energy system scheduling method based on time series feature extraction and policy optimization as claimed in any one of claims 1 to 6.

Description

Method and device for dispatching park comprehensive energy system based on time sequence feature extraction and strategy optimization Technical Field The invention relates to the technical field of energy scheduling, in particular to a park comprehensive energy system scheduling method and device based on time sequence feature extraction and strategy optimization. Background The park is a designated area for centralized and unified planning, and enterprises, companies and the like of a certain type of specific industries and forms are arranged in the park for unified management. The park is a core carrier for urban energy consumption and carbon emission, an energy system of the park is undergoing transformation from a traditional single fossil energy function mode to a comprehensive energy service model with 'source-net-charge-storage' multi-element cooperation, and the park electric-thermal comprehensive energy system relies on cascade utilization and complementary scheduling of electric and thermal core energy, so that the park has the potential of improving clean energy consumption rate and reducing system operation cost. However, in practical engineering application, the clean energy output has certain fluctuation, and the park electric load (including business office, industrial production and resident life loads) and the heat load (including winter heating, life hot water and industrial heat load) have obvious time sequence coupling characteristics. The existing energy system scheduling method is difficult to capture the trend and the periodicity characteristic of the fluctuation of the clean energy output, and the strategy of independent optimization of an electric subsystem and a thermal subsystem is adopted. The existing energy system scheduling method has the problems of high wind-abandoning and light-abandoning rate, energy resource waste, power grid voltage fluctuation, frequency deviation and the like, the scheduling reliability is insufficient, simultaneously the user demand response is mostly limited to peak clipping and valley filling of electric loads, the load translation is realized without combining the thermal inertia characteristics of the thermal loads, and the multi-energy complementary potential of the system is not fully excavated. In summary, the accuracy of the scheduling scheme generated by the park energy system scheduling method in the prior art is not high. Disclosure of Invention The invention provides a park comprehensive energy system scheduling method and device suitable for time sequence feature extraction and strategy optimization, which are used for solving the defect of low accuracy of a park energy system scheduling method in the prior art and improving the accuracy of a park energy system scheduling scheme. The invention provides a park comprehensive energy system scheduling method based on time sequence feature extraction and strategy optimization, which comprises the following steps: acquiring park energy state data of a plurality of historical moments, wherein the park energy state data reflects the output condition of clean energy in a park and the energy load condition in the park, and the energy load comprises electric load and thermal load; Inputting the park energy state data at the plurality of historical moments into a trained time sequence feature extraction model to obtain time sequence features output by the time sequence feature extraction model; and inputting input data into a strategy optimization model, and determining an energy system scheduling strategy based on output data of the strategy optimization model, wherein the input data comprises the time sequence characteristics and the park energy state data at the current moment, and the energy system scheduling strategy comprises a demand response execution amount and an equipment adjustment amount in the park energy system. According to the method for dispatching the park comprehensive energy system based on time sequence feature extraction and strategy optimization, the park energy state data comprise power, electric load, heat load, electricity price, charge state of an energy storage system, indoor temperature and outdoor temperature of clean energy. According to the park comprehensive energy system scheduling method based on time sequence feature extraction and strategy optimization, the training process of the strategy optimization model comprises the following steps: inputting the sample time sequence characteristics into the strategy optimization model, and determining a sample energy system scheduling strategy based on sample output data of the strategy optimization model; Determining evaluation scores of the sample energy system scheduling strategy in a plurality of indexes based on a state transition result corresponding to the sample energy system scheduling strategy; Weighting and fusing the evaluation scores of the multiple indexes to obtain an objective function value; updating the