CN-122026390-A - Zero-carbon-park-oriented source network load storage collaborative scheduling control method and system

CN122026390ACN 122026390 ACN122026390 ACN 122026390ACN-122026390-A

Abstract

The invention discloses a source network charge storage collaborative scheduling control method and a system for a zero-carbon park, and relates to the technical field of intelligent power grid scheduling and control, wherein the method comprises the following steps of S1, acquiring operation data of each unit of the local source network charge storage and constructing a collaborative scheduling model; S2, carrying out multi-time scale collaborative scheduling optimization by adopting a group intelligent optimization algorithm based on a collaborative scheduling model to reduce the operation cost and carbon emission of the park, generating a scheduling control instruction, S3, carrying out coordinated control on an energy storage system, an adjustable load unit and a distributed new energy power generation unit in the park according to the scheduling control instruction, and interacting with an upper power grid, S4, carrying out operation effect evaluation on the actual operation state of the park, and dynamically adjusting parameters of the collaborative scheduling model or optimization target weight of the group intelligent optimization algorithm to form closed loop optimization. The energy optimization configuration, the low-carbon operation and the friendly interaction of the power grid are realized, and technical support is provided for the construction of the zero-carbon park.

Inventors

ZHOU YIZHOU
TANG HUA
DING YANG
Lian Chunlin
YOU CHANGQING
PU CHENG
HE DONGWEI

Assignees

河海大学
盐城电力设计院有限公司
国网江苏省电力有限公司盐城供电分公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. A source network load storage cooperative scheduling control method for a zero-carbon park is characterized by comprising the following steps: S1, acquiring operation data of various units of a park, a network, a load and a storage, and constructing a collaborative scheduling model containing energy storage life attenuation characteristics and load regulation characteristics based on the operation data; S2, based on the collaborative scheduling model, taking the reduction of the operation cost and carbon emission of a park as optimization targets, adopting a group intelligent optimization algorithm to perform multi-time-scale collaborative scheduling optimization to generate scheduling control instructions, wherein the multi-time-scale collaborative scheduling optimization adopts a time scale frame combining daily optimization, daily rolling correction and real-time coordination control; S3, according to the dispatching control instruction, the energy storage system, the adjustable load unit and the distributed new energy power generation unit in the park are coordinated and controlled to operate, and power and information interaction is carried out with the upper power grid; and S4, monitoring the actual operation state of the park, calculating an operation effect evaluation index to obtain an operation effect evaluation result, and dynamically adjusting parameters of the collaborative scheduling model or optimization target weights of the group intelligent optimization algorithm according to the operation effect evaluation result to form closed loop optimization.
2. The method according to claim 1, wherein in step S1, the obtaining operation data of each unit of the campus source, the network, the load and the storage specifically includes: The distributed new energy power generation unit data comprise irradiance, assembly temperature and actual output data of distributed photovoltaic, and wind speed, wind direction and actual output data of wind power; the energy storage system state data comprises a state of charge value, real-time charge and discharge power and cycle times of an energy storage battery; load data comprising real-time power of various loads and start-stop states of the adjustable load units; grid operation data including power, voltage at the campus grid site, on-grid electricity prices from the grid, and marginal carbon emission factors.
3. The method according to claim 1 or 2, wherein the collaborative scheduling model constructed in step S1 includes a new energy output prediction model, an energy storage system model, an adjustable load model, a grid operation constraint model, and a carbon emission calculation model, wherein: The new energy output prediction model is constructed based on a long-short-period memory network, inputs historical environment data and output data, and outputs new energy output prediction values with different time scales; The energy storage system model comprises a charge and discharge efficiency model for representing energy conversion loss and a battery life attenuation model for evaluating and restraining long-term operation economy of a battery; The adjustable load model is used for flexible load classified management and setting the adjustment power range, the single adjustment duration and the accumulated adjustment quantity constraint; the power grid operation constraint model is used for ensuring that the exchange power and the voltage of the grid-connected point of the park meet the power grid safe operation requirement; The carbon emission calculation model takes the net purchase power of the park as a calculation boundary to calculate the carbon emission in the scheduling period.
4. The method according to claim 1, wherein step S2 specifically comprises: performing day-ahead optimization, and making a day-ahead plan for 24 hours in the future by taking an hour as a step length, wherein the day-ahead plan comprises an energy storage charge-discharge plan and a load regulation plan; Performing intra-day rolling correction, namely performing quick correction on the pre-day plan based on updated ultra-short-term prediction data by taking 15 minutes as a step length so as to stabilize prediction errors and control grid-connected point power fluctuation; And carrying out real-time coordination control, and generating a minute-level power adjustment instruction for the energy storage system based on real-time power or voltage deviation of the grid-connected point.
5. The method of claim 4, wherein the optimization objective of the day-ahead optimization is to minimize a composite objective function F: ; Wherein, the To include the cost of purchasing electricity, the cost of equipment loss and the cost of battery depreciation, For carbon emission costs calculated based on net on-hand power for the campus, And For the weight coefficient dynamically adjusted according to the deviation of the actual operation index from the target value, For a time step index optimized for the day before.
6. The method of claim 5, wherein the population intelligent optimization algorithm employed in step S2 is a clean fish optimization algorithm, the execution of which comprises: Initializing a population by adopting chaotic mapping, updating the positions of individuals in the population by adopting a two-generation circulation strategy comprising a target pursuit mode and a sex conversion mode, and processing the individuals in the population by adopting a quasi-reflection reverse learning strategy based on solution space symmetrical mapping after each iteration so as to maintain diversity.
7. The method according to claim 4, wherein the performing real-time coordination control specifically includes: The method comprises the steps of adopting a mixed control strategy which takes proportional integral control as a main part and model predictive control as a standby part, generating a power adjustment instruction for an energy storage system through proportional integral control based on the power deviation of a grid-connected point under a normal working condition, and automatically switching to the model predictive control strategy for adjustment when the adjustment performance of the proportional integral control exceeds a preset stability threshold.
8. The method of claim 1, wherein in step S3, the operation of the energy storage system, the adjustable load unit, and the distributed new energy power generation unit in the coordination control park is implemented by a hierarchical coordination control strategy, the hierarchical coordination control strategy comprising: the local control layer executes control instructions from the park energy management system by a local controller connected with each energy unit, so that the rapid closed-loop control of the equipment is realized; The park coordination layer is used for executing the multi-time-scale collaborative scheduling optimization of the step S2 by the park energy management system, issuing a control instruction to the local control layer, and simultaneously carrying out information and instruction interaction with the power grid interaction layer; and the power grid interaction layer is used for realizing data uploading and regulation instruction receiving between the park energy management system and the upper power grid dispatching system and supporting power grid interaction requirements.
9. A method according to claim 3, wherein step S4 comprises: Calculating new energy consumption rate, carbon emission intensity and operation cost reduction rate obtained based on actual operation data, and comparing the new energy consumption rate, the carbon emission intensity and the operation cost reduction rate with preset target values of various indexes to obtain index deviation as an operation effect evaluation result; And dynamically adjusting parameters in the collaborative scheduling model or weight coefficients of the optimization targets based on deviation of at least one evaluation index of new energy consumption rate, carbon emission intensity, operation cost and battery residual life from target values thereof, wherein the adjustment comprises at least one of increasing the weight of the carbon emission item in the optimization targets when the carbon emission intensity is higher than the target values, increasing the weight of the operation cost item in the optimization targets when the operation cost is higher than the target values, adjusting the adjustment range constraint of the adjustable load model or the charge state operation upper and lower limits of the energy storage system model when the new energy consumption rate is lower than the target values, and adjusting the charge and discharge power constraint of the energy storage system model when the attenuation speed of the battery residual life exceeds a preset threshold value.
10. Zero-carbon park-oriented source network load storage collaborative scheduling control system is characterized by comprising: the data acquisition and modeling module is used for acquiring the operation data of each unit of the internal source, the network, the load and the storage of the park, and constructing a collaborative scheduling model containing the energy storage life attenuation characteristic and the load regulation characteristic based on the operation data; the multi-time scale optimization module is used for carrying out multi-time scale collaborative scheduling optimization by adopting a group intelligent optimization algorithm based on the collaborative scheduling model and taking the reduction of the operation cost and the carbon emission of a park as optimization targets to generate scheduling control instructions, wherein the multi-time scale collaborative scheduling optimization adopts a time scale frame combining daily optimization, daily rolling correction and real-time coordination control; The hierarchical coordination control module is used for coordinating and controlling the operation of an energy storage system, an adjustable load unit and a distributed new energy power generation unit in the park according to the scheduling control instruction, and carrying out power and information interaction with an upper power grid; The evaluation and closed-loop optimization module is used for monitoring the actual operation state of the park, calculating an operation effect evaluation index to obtain an operation effect evaluation result, and dynamically adjusting parameters of the collaborative scheduling model or optimization target weights of the group intelligent optimization algorithm according to the operation effect evaluation result to form closed-loop optimization.

Description

Zero-carbon-park-oriented source network load storage collaborative scheduling control method and system Technical Field The embodiment of the invention relates to the technical field of intelligent power grid dispatching and control, in particular to a source network load storage collaborative dispatching control method and system for a zero-carbon park. Background Along with the deep propulsion of a double-carbon target and the construction of a novel power system, the zero-carbon park becomes a core carrier for realizing low-carbon transformation, high-proportion new energy consumption, safe and efficient power supply in the user side scenes such as an industrial park, an industrial and commercial complex, an industrial cluster and the like, and is a key ground state of the novel power system on the user side. The energy storage system and the diversified flexible load are widely deployed, the operation of the energy storage system and the diversified flexible load is characterized by strong randomness and volatility of new energy output, high source load space-time matching difficulty, severe grid-connected safety constraint of a power grid, and coupling of dual indexes of a zero-carbon target and economic operation, and clear engineering and scene landing requirements are provided for the multi-link collaborative scheduling control technology of the charge storage of the endogenous network of the park. The scheme builds an environment state space by collecting the storage and transportation data of the regional endogenous network charges in real time, trains a control strategy network based on the SAC depth reinforcement learning algorithm of entropy regularization, generates real-time scheduling instructions by using new energy utilization, low-carbon electricity purchasing and load regulation as optimization targets, and realizes the coordination control of the source network charges in the zero-carbon regional scene. The prior art (comprising the scheme) still has the core defect of serious disjoint with the actual floor operation scene of the zero-carbon park in the power industry, and is characterized in that (1) the suitability of the dispatching time scale and the power grid engineering operation specification is insufficient, the multi-focus real-time single time scale closed-loop control of the prior scheme does not construct a multi-time scale cooperative dispatching frame for adapting to the dispatching management rules of the power industry, the prediction error caused by the ultra-short-term output fluctuation of new energy and the dynamic change of load cannot be stabilized, and the rigid assessment requirement of the upper power grid on the power fluctuation and the voltage deviation of the grid-connected point of the park is difficult to meet. (2) The method for constructing the practical operation characteristics of the non-attached power equipment by the scheduling model and the national standard requirements of industry has the advantages that the existing scheme has an ideal and simplified problem on modeling of each unit of the source network charge storage, only basic charge and discharge power constraint is considered on an energy storage system, and the influence of the life attenuation characteristics of the battery on the whole life cycle operation economy of a park is not included. (3) The optimization algorithm has insufficient suitability, namely the intelligent algorithms such as deep reinforcement learning adopted by the existing scheme have the problems of high dependence of training samples, weak field generalization capability and poor interpretability, and cannot be adapted to non-convex optimization scenes of multivariable, multi-constraint and nonlinear strong coupling in a zero-carbon park. Therefore, the development of the source network load storage collaborative scheduling control method which is fit with the actual running scene of the zero-carbon park in the power industry, the multi-time scale collaborative adaptation, the full-element refined modeling, the strong non-convex scene convergence and the full-flow closed-loop optimization becomes the core engineering technical problem to be solved in the field. Disclosure of Invention The invention aims to solve the technical problems that in the prior art, the fluctuation of the output of a distributed new energy source (photovoltaic and wind power) in a zero-carbon park is strong, the cooperativity of each unit of a source network charge storage is poor, and the carbon emission and the operation cost are difficult to consider, and provides a source network charge storage cooperated scheduling control method and a system for the zero-carbon park, which are used for realizing multi-objective optimization of park energy source consistent capability promotion, superior power grid friendly support and full life cycle carbon emission minimization, are suitable for park-level comprehensive energy systems comprising di