CN-121984110-A - Wind-solar-water-storage integrated base capacity configuration mutual-aid scheduling combined optimization method

CN121984110ACN 121984110 ACN121984110 ACN 121984110ACN-121984110-A

Abstract

The invention relates to the technical field of power system optimization, and discloses a wind-light-water-storage integrated base capacity configuration mutual-aid scheduling combined optimization method which comprises the steps of constructing a wind-light-water-storage combined topology and device physical model, constructing an optimal scheduling model based on the wind-light-water-storage combined topology and device physical model and aiming at minimum comprehensive operation cost and minimum pollutant emission in a scheduling period, solving the optimal scheduling model by adopting an improved depth deterministic strategy gradient algorithm, introducing a manual strategy guide term in action exploration to optimize a reward function, setting a manual exploration probability mechanism to balance exploration and utilization, and outputting an energy storage charge-discharge, unit output and cross-region mutual-aid power optimal scheduling instruction through an Actor-Critic network, so as to realize multi-time-scale rolling optimization and online control. The invention has the advantages that the clean energy consumption capability is obviously improved, and the cross-region synergistic benefit and the economic and environmental protection level of the whole operation are improved.

Inventors

ZHAO WEIJIE
JIANG YI
HAN JIANWEI
Pan Jujian
SUN HAO
YANG TAO

Assignees

中国南方电网有限责任公司超高压输电公司昆明局

Dates

Publication Date: 20260505
Application Date: 20251222

Claims (6)

1. The method for the combined optimization of the capacity allocation mutual-aid scheduling of the wind-solar-water-storage integrated base is characterized by comprising the following steps of, Constructing a wind-light-water-storage combined topology and device physical model, wherein the wind-light-water-storage combined topology and device physical model comprises a photovoltaic output model, a wind power output model, a hydroelectric output model, a reservoir model and an energy storage model; based on wind-light-water-storage combined topology and device physical model, constructing an optimized dispatching model with the aim of minimum comprehensive operation cost and minimum pollutant emission in a dispatching cycle; The improved depth deterministic strategy gradient algorithm is adopted to solve an optimized scheduling model, manual strategy guide items are introduced in action exploration to optimize a rewarding function, a manual exploration probability mechanism is set to balance exploration and utilization, and an optimized scheduling instruction of energy storage charge and discharge, unit output and cross-region mutual power is output through training an Actor-Critic network, so that rolling optimization and on-line control of multiple time scales are realized.
2. The wind-solar-water-storage integrated base capacity configuration mutual-aid scheduling joint optimization method according to claim 1, which is characterized by comprising the following steps: The photovoltaic output model is as follows: ; In the formula, The actual output power of the photovoltaic cell; rated output power of the photovoltaic cell; is the actual intensity of solar radiation; The illumination intensity under standard test conditions; Is the temperature coefficient of the photovoltaic module; The actual temperature of the photovoltaic module; is the temperature tested under standard conditions; Is the shading coefficient of the photovoltaic array; The wind power output model is as follows: ; In the formula, The wind power output power is the wind power; Rated output power of the fan; Is the actual wind speed; is the rated wind speed; is the cut-in wind speed; To cut out wind speed; The water power output model is as follows: ; In the formula, The water power station output in the t period; Gravitational acceleration; The efficiency of the unit is; The water flow passing through the generator set for the period t; The power generation water purification head at the time t; The reservoir model is as follows: ; In the formula, Is the reservoir capacity in the period t; And Respectively is The volume and the flow rate of the warehouse entry in the time period; Is that Rainfall in the reservoir area in the period; Is that The water amount consumed by the water motor group in the period; And Respectively is Reservoir water drainage and corresponding water drainage flow in the period; And Respectively is Water supply amount and water supply flow rate of the water reservoir in the period; Is that The evaporation capacity of the water reservoir in the period; the flow of electricity generation of the hydroelectric generating set; Is a time interval; The energy storage model is as follows: ; ; In the formula, The rated capacity of the storage battery under the standard test condition is shown as k, wherein k is a capacity temperature coefficient; The actual working temperature of the battery; the SOC is the charge state of the storage battery; Is the initial state of charge; Is the initial time; Is the current moment; Is a time interval; Charging and discharging current for the storage battery; the reaction current is lost to the battery.
3. The wind, light and water storage integrated base capacity configuration mutual-aid scheduling combined optimization method according to claim 1 is characterized in that an objective function of the optimal scheduling model is as follows: ; ; ; In the formula, A comprehensive operation cost function for the micro-grid; is a pollutant emission function of the micro-grid; for a period of 1 day; the quantity of the cogeneration units, the photovoltaic units and the pure heating units is respectively; Is the first CHP at The cost of time; Is the first At PV of The cost of time; Is the first At the Heat site The cost of time; Active power for interaction with a large power grid; For the moment of time Electricity price of (2); is the CHP pollutant emission coefficient; the thermal power of the ith CHP at the time t; Is the Heat pollutant emission coefficient; the thermal power of the jth Heat at the time t; is the emission of the large power grid power generation process.
4. The wind, light and water storage integrated base capacity configuration mutual-aid scheduling joint optimization method according to claim 3 is characterized in that constraint conditions of the optimal scheduling model comprise: Electric balance constraint: ; In the formula, Is the first The power generation of the platform CHP; And Respectively charging and discharging power of the energy storage unit at the time t; The generated power of the kth PV at the time t; The electric load requirement at the time t; Thermal equilibrium constraint: ; In the formula, Is a thermal load demand; The thermal power for the j-th Heat; thermal power for the ith CHP; Upper and lower limit constraints: ; ; ; ; ; ; In the formula, Maximum generated power for PV; Maximum charging power for the energy storage unit; maximum discharge power of the energy storage unit; Discharge power for CHP; minimum and maximum discharge power of CHP respectively; Heating power of CHP; 、 the minimum and maximum heating power of CHP are respectively; The heating power of Heat at the time t is set; Is the maximum heating power of Heat.
5. The wind, light and water storage integrated base capacity configuration mutual-aid scheduling joint optimization method according to claim 1 is characterized by comprising the following steps of introducing a manual strategy guide item to optimize a reward function: ; ; In the formula, Is a reward function; the running cost of the micro-grid is; The cost is balanced for the micro-grid power; The actual charge and discharge power of the energy storage system in the period; Estimating charge and discharge power for the energy storage system in the period; Is a reward coefficient; The action value output by the strategy network in reinforcement learning is obtained; The action value used for environment update after the manual strategy modification is used; Responding actions for a demand side output by the strategy network; responding actions for the demand side after the manual strategy correction; The energy storage charging and discharging actions are output by the strategy network; and the energy storage charging and discharging actions after the manual strategy correction are performed.
6. The method for the combined optimization of the capacity allocation and mutual-aid scheduling of the wind, solar and water storage integrated base according to claim 5 is characterized in that the artificial exploration probability mechanism adopts a decay strategy and is expressed as follows: ; In the formula, To explore probabilities; Is the number of rounds; is a positive number greater than 1.

Description

Wind-solar-water-storage integrated base capacity configuration mutual-aid scheduling combined optimization method Technical Field The invention relates to the technical field of power system optimization, in particular to a mutual-aid dispatching combined optimization method for capacity allocation of a wind-light-water-storage integrated base. Background In a high-proportion renewable energy accessed power system, along with rapid development of high-proportion wind power and photovoltaic in a cross-region delivery scene, clean energy output has obvious space-time fluctuation and uncertainty, a receiving end load is mismatched with a source side output, waste wind and waste light are easily caused, the delivery utilization rate is reduced, meanwhile, capacity allocation and operation scheduling are long-term split, cross-region cooperation and mutual economy lack of executable triggering, allocation and settlement rules, and an end-to-end integrated optimization and closed loop check mechanism is difficult to form. In engineering modeling, a multi-energy system such as wind, light and water storage is generally based on a device physical mechanism to establish an operation model, wherein a reservoir side couples inflow, rainfall, evaporation, water discarding, water supply, power generation and water discharging and other processes in a volume conservation relation, and the reservoir area evaporation and reservoir capacity dynamics are drawn through a non-linear function such as reservoir capacity, water level and evaporation area, and the energy storage side represents a constraint boundary of charge state evolving along with charge and discharge current and loss through a capacity temperature correction and SOC integral model, and the model provides a basic energy conservation and operation boundary for subsequent optimization. In the operation optimization level, the existing dispatching is expanded around conventional constraints such as electric/thermal balance, unit-energy storage power/energy upper and lower limits and the like, equipment such as CHP/Heat and the like is usually limited by minimum/maximum output boundaries, local economy can be realized under static rules, unified description of tie line margin and standby cooperation is insufficient, and inter-regional mutual aid is mostly dependent on fixed threshold values or post coordination. The closest existing scheme of the invention has the commonality that firstly, a split capacity planning-operation scheduling route is adopted, the capacity side is independently optimized on the basis of historical weather and load statistics to install and store energy configuration, the operation side is used for independently solving economic/emission targets on the daily front/daily inner scale, the two are lack of shared indexes and regular linkage, and linkage triggering of mutual aid and forming of executable settlement caliber are difficult under uncertain conditions. Secondly, although the conventional model solution based on the equipment boundary can meet the requirements of the unit/energy storage safety boundary and the electric heat balance, the unified measurement and collaborative optimization of the net load, standby and tie line margin still have insufficient under the multi-time scale, so that the operability and the foresight of the inter-regional mutual aid are limited. Thirdly, part of work introduces DDPG and other continuous action reinforcement learning to acquire a fine-grained scheduling strategy, but under the traditional framework, an agent usually performs unconstrained greedy or blind exploration, so that the problems of low convergence speed, easy sinking into local optimization and the like exist, and the injection of a man-machine cooperative strategy is insufficient. To sum up, in the prior art, although a certain effect is obtained on the capacity side or the operation side, it is still difficult to simultaneously meet the systematic requirements of "capacity-scheduling integrated optimization under uncertainty", "mutual-aid triggering and distribution with the payload, standby and link margin as common indexes", "having executable settlement and closed-loop check", and it is difficult to support the unified targets of safety, economy and low carbon under the inter-regional collaboration scene. Disclosure of Invention The invention aims to overcome the defects of the prior art and provides a combined optimization method for mutual-aid scheduling of capacity allocation of a wind-solar-water-storage integrated base. The invention aims at realizing the technical scheme that the wind-solar-water-storage integrated base capacity configuration mutual-aid scheduling combined optimization method comprises the following steps of, Constructing a wind-light-water-storage combined topology and device physical model, wherein the wind-light-water-storage combined topology and device physical model comprises a photovoltaic output mod