Scheduling optimization method and system for mine micro-grid

Abstract

The invention discloses a dispatching optimization method and system for a mine micro-grid, and relates to the technical field of micro-grids. The method comprises the steps of constructing a system model of the mining area comprehensive energy system; a mining area comprehensive energy operation model is constructed based on the system model, and constraint functions of the mining area comprehensive energy operation model comprise a net income maximization function, a renewable energy consumption maximization function and a carbon emission minimization function; the model is converted into a Markov decision process for determining a scheduling strategy for performing day-ahead scheduling control on the mining area integrated energy system, and the scheduling strategy comprises an output power range and an equipment state of each equipment in a day-ahead scheduling period; and under the limitation of the agreed range of the scheduling strategy and the constraint condition of the mining area comprehensive energy operation model, with the purpose of minimizing the operation cost, performing intra-day scheduling control by adopting mixed integer linear programming. Therefore, the problem that a scheduling optimization method in the prior art cannot be applied to the mine micro-grid or has many defects during application is solved.

Description

Technical Field

[0001] This application relates to the field of microgrid technology, and more specifically, to a scheduling optimization method and system for a mining microgrid. Background Technology

[0002] Integrated energy technology enables the coupling and conversion of multiple energy forms, constructing an integrated energy system for mines. This system can meet the diverse energy demands of the load side, achieve the synergistic and efficient utilization of derived energy sources, improve energy efficiency, and reduce environmental pollution.

[0003] With increasing demands for clean energy, integrated mine energy systems combining battery storage, wind, solar, and hydropower resources are being promoted in most countries to facilitate the clean energy transition. However, due to the volatility and uncertainty of renewable energy sources, and the fact that mine loads are production loads with strong rigidity, the complementarity of wind, solar, and hydropower for power balance is unstable. To mitigate these negative impacts, recent research has emphasized the energy transfer within the time domain of battery storage and pumped-storage hydropower to form a hybrid wind-solar-hydropower-storage energy system, also known as a mine microgrid. However, such systems require high-level scheduling; improper scheduling can lead to severe inefficiencies, resulting in significant resource waste and severe power shortages. Current scheduling optimization methods are either unsuitable for the optimal scheduling of integrated energy systems or suffer from numerous shortcomings, such as inability to meet multi-timescale constraints, difficulty in solving problems, and low scheduling accuracy. Summary of the Invention

[0004] In view of this, the embodiments of this application aim to provide a scheduling optimization method, system, microgrid and electronic equipment for mining microgrids, so as to solve the problem that the scheduling optimization methods of the prior art cannot be applied to mining microgrids, or have many defects when applied.

[0005] In a first aspect, the present invention provides a scheduling optimization method for a mining microgrid, comprising: Constructing a system model for an integrated energy system in a mining area; Based on the system model, a comprehensive energy operation model for the mining area is constructed. The constraint functions of the comprehensive energy operation model for the mining area include a net profit maximization function, a renewable energy consumption maximization function, and a carbon emission minimization function. The integrated energy operation model of the mining area is converted into a Markov decision process to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system of the mining area. The scheduling strategy includes the output power range and equipment status of each device of the integrated energy system of the mining area during the day-ahead scheduling cycle. Within the agreed scope of the scheduling strategy and the constraints of the integrated energy operation model of the mining area, with the goal of minimizing operating costs, mixed integer linear programming is used to perform intraday scheduling control of the integrated energy system of the mining area.

[0006] In one possible implementation, the scheduling optimization method for mining microgrids further includes: The operating cost of executing the intraday scheduling control is fed back to the Markov decision process to optimize the scheduling strategy.

[0007] In one possible implementation, the system model for constructing the integrated energy system of the mining area includes: Obtain a system model of the comprehensive energy infrastructure system in the mining area as the initial system model. The initial system model includes a compressed air energy storage device sub-model, a photovoltaic power generation system sub-model, a wind turbine power generation system sub-model, a pumped storage power station system sub-model, a battery system sub-model, and a diesel engine system sub-model. A coal gas reservoir model is constructed, which includes a pressure swing adsorption unit sub-model and a combined power and heat system sub-model. The gas layer model is combined with the initial system model to obtain the system model of the integrated energy system of the mining area.

[0008] In one possible implementation, the constraints include: Power balance constraints, line transmission power constraints, heat pump constraints, heat pipeline constraints, heat load constraints, and heating network constraints.

[0009] In one possible implementation, the step of converting the integrated energy operation model of the mining area into a Markov decision process to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system of the mining area includes: The integrated energy operation model of the mining area is converted into a Markov decision process, and the state space and action space of the agent of the Markov decision process and the reward function are defined respectively. The state space includes the current state of each device of the integrated energy system of the mining area, and the action space includes the output power and state of each device of the integrated energy system of the mining area at a preset date time scale. The agent is trained; Based on the trained agent, a scheduling strategy for day-ahead scheduling control of the integrated energy system in the mining area is determined.

[0010] In one possible implementation, the reward function includes rewards corresponding to maximizing net revenue, maximizing renewable energy consumption, and minimizing carbon emissions of the integrated energy operation model of the mining area, as well as penalties for violating the constraints of the integrated energy operation model of the mining area.

[0011] In one possible implementation, training the agent includes: The agent is trained using a dual-delay deep deterministic policy gradient algorithm.

[0012] In one possible implementation, training the agent further includes: Obtain the first historical state space of the integrated energy system of the mining area and the first historical action space corresponding to the first historical state space within each scheduling cycle of the preset date time scale; A preset perturbation factor is added to the first historical state space to obtain the third historical state space; Obtain a second historical state space and a historical value of the reward function corresponding to the second historical state space. The second historical state space is a state space corresponding to the intraday scheduling control, determined within the agreed range of the first historical action space and under the constraints of the said conditions, with the goal of minimizing operating costs. The third historical state space, the first historical action space, the second historical state space, and the historical values ​​of the establishment function are used as data samples to train the agent.

[0013] Secondly, the present invention provides a scheduling optimization system for a mining microgrid, comprising: The system model building unit is used to construct a system model of the integrated energy system in the mining area. The operation model construction unit is used to construct a comprehensive energy operation model for the mining area based on the system model. The constraint functions of the comprehensive energy operation model for the mining area include a net profit maximization function, a renewable energy consumption maximization function, and a carbon emission minimization function. The day-ahead scheduling control unit is used to convert the integrated energy operation model of the mining area into a Markov decision process, and to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system of the mining area. The scheduling strategy includes the output power range and equipment status of each device of the integrated energy system of the mining area during the day-ahead scheduling cycle. The intraday scheduling control unit is used to perform intraday scheduling control of the integrated energy system of the mining area under the constraints of the agreed scope of the scheduling strategy and the constraints of the integrated energy operation model of the mining area, with the goal of minimizing operating costs, using mixed integer linear programming.

[0014] In one possible implementation, the scheduling optimization system for the mining microgrid further includes: The feedback unit is used to feed back the operating cost of executing the intraday scheduling control to the Markov decision process in order to optimize the scheduling strategy.

[0015] The scheduling optimization method for mine microgrids provided by this invention constructs a system model of the integrated energy system in the mining area. Based on this system model, a comprehensive energy operation model for the mining area is constructed. The constraint functions of this comprehensive energy operation model include a net revenue maximization function, a renewable energy absorption maximization function, and a carbon emission minimization function. The comprehensive energy operation model is converted into a Markov decision process to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system. The scheduling strategy includes the output power range and equipment status of each device in the integrated energy system within the day-ahead scheduling cycle. Under the constraints of the scheduling strategy and the constraints of the comprehensive energy operation model, mixed-integer linear programming is used to perform intraday scheduling control of the integrated energy system, with the goal of minimizing operating costs. Thus, by combining day-ahead and intraday scheduling of the integrated energy system, more precise control and effective coordination are achieved between different operating timelines. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 The diagram shows a flowchart of a scheduling optimization method for a mining microgrid provided in an embodiment of the present invention.

[0018] Figure 2 The figure shown is a system model of the integrated energy infrastructure system for mining areas provided in an embodiment of the present invention.

[0019] Figure 3 The figure shown is a system model of the integrated energy system for mining areas provided in an embodiment of the present invention.

[0020] Figure 4 The diagram shown illustrates the relationship between the policy network and the evaluation network of the dual-delay deep deterministic policy gradient algorithm provided in this embodiment of the invention.

[0021] Figure 5 The diagram shown is a structural diagram of the policy network and evaluation network provided in an embodiment of the present invention.

[0022] Figure 6 The diagram shown is a structural diagram of a scheduling optimization system for a mining microgrid provided in an embodiment of the present invention.

[0023] Figure 7 The diagram shown is a structural diagram of another mine microgrid scheduling optimization system provided in an embodiment of the present invention.

[0024] Figure 8 The diagram shown is a structural schematic of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0025] Unless otherwise defined, the technical or scientific terms used in the embodiments of this specification shall have the ordinary meaning understood by one of ordinary skill in the art to which this specification pertains. The terms "first," "second," and similar terms used in the embodiments of this specification do not indicate any order, quantity, or importance, but are merely used to avoid confusion of constituent elements.

[0026] Unless the context otherwise requires, throughout this specification, "a plurality of" means "at least two," and "including" is interpreted as open-ended or encompassing, that is, "including, but not limited to." In the description of this specification, terms such as "one embodiment," "some embodiments," "exemplary embodiment," "example," "specific example," or "some examples" are intended to indicate that a particular feature, structure, material, or characteristic associated with that embodiment or example is included in at least one embodiment or example of this specification. The illustrative representations of the above terms do not necessarily refer to the same embodiment or example.

[0027] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0028] It should be noted that a wind-solar-hydro-storage hybrid energy system is a system that incorporates both long-term and short-term dispatching elements. Long-term dispatching strategies impose operational boundaries on short-term regulation capabilities, while the dynamics of short-term operation often significantly impact strategic decisions in long-term dispatching. Therefore, neglecting the critical interdependencies between these operational timelines when formulating integrated energy system dispatching strategies can lead to severe inefficiencies, resulting in significant resource waste and severe power shortages. In this regard, it is necessary to study the coordination of long-term and short-term operations to achieve the rational dispatching of multiple energy resources in a mining integrated energy system. However, current dispatching optimization methods are not suitable for the optimal dispatching of mining integrated energy systems.

[0029] For example, scheduling optimization methods using only a single time scale cannot provide effective solutions for the long-term operation of energy systems because they only study a single time scale. Traditional model-driven scheduling optimization methods suffer from excessive computational burden in both long-term and short-term scheduling. Current methods for coordinating long-term and short-term power balance can generally be divided into two categories. The first category can optimize system operation for 8760 hours throughout the year with a 1-hour time resolution. However, due to the introduction of too many variables by various uncertainties, leading to the curse of dimensionality and difficulty in solving, this method is not suitable for integrated mining energy systems. The second category extracts typical scenarios through scenario reduction to represent changes throughout the year. Although this method is effective in reducing model complexity, it weakens the temporal fluctuation characteristics of source-load power throughout the year due to insufficient information provided. Furthermore, the widely used data-driven methods based on deep reinforcement learning cannot ensure that the scheduling scheme for integrated mining energy systems strictly meets operational constraints across multiple time scales; therefore, they cannot be applied to the scheduling of integrated mining energy systems.

[0030] This invention aims to solve the above problems by combining deep reinforcement learning and mixed integer linear programming for collaborative scheduling optimization of long-term and short-term multi-timescale integrated energy systems in mines.

[0031] The present invention provides a scheduling optimization method for a mining microgrid, which is executed on an electronic device. The electronic device can be a control module of the mining microgrid, or a smart terminal device such as a laptop, mobile phone, or tablet connected to the control module, or a server remotely connected to the control module.

[0032] See Figure 1 , Figure 1 This is a flowchart of a scheduling optimization method for a mining microgrid provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the process of this method mainly includes the following steps: S100, Construct a system model for the integrated energy system of the mining area.

[0033] Specifically, by providing the system structure of the integrated energy system in the mining area, and then analyzing the characteristics of each distributed unit based on the system's operating mechanism, a model of each unit can be established, thereby obtaining the system model of the integrated energy system in the mining area.

[0034] In a preferred embodiment, a system model for a comprehensive energy system in a mining area is constructed, including: Obtain a system model of the comprehensive energy infrastructure system in the mining area as the initial system model. The initial system model includes a compressed air energy storage device sub-model, a photovoltaic power generation system sub-model, a wind turbine power generation system sub-model, a pumped storage power station system sub-model, a battery system sub-model, and a diesel engine system sub-model. A coal gas reservoir model is constructed, which includes a pressure swing adsorption unit sub-model and a combined power and heat system sub-model. By incorporating a composite gas layer model into the initial system model, a system model of the integrated energy system of the mining area is obtained.

[0035] Specifically, the integrated energy infrastructure system of the mining area is as follows: Figure 2 The integrated energy system structure shown, consisting of an electric-heat network coupling, includes a power grid, a heating network, and a compressed air energy storage system. Among these, for example... Figure 2 As shown, the energy supply side consists of the power grid, wind power, photovoltaic power, and heating network. The energy conversion and storage equipment includes compressed air energy storage devices, battery energy storage systems, and pumped storage devices. The mine's electrical load is provided by the power grid, wind power generation, photovoltaic power generation, battery discharge, pumped storage power generation, and compressed air energy storage power generation. The user's heat load is supplied by the heating network and the heat storage equipment in the compressed air energy storage system.

[0036] The compressed air energy storage device includes a compressor, an air storage tank, a turbine, and a heat regeneration system. During charging, the compressor uses off-peak electricity and reduced wind power to compress atmospheric pressure air and stores it in the air storage tank. When electricity is needed, the high-pressure air stored in the air storage tank can be released through stored heat energy and preheated by a heating device to drive the turbine to generate electricity.

[0037] Based on this, it can be understood that the compressed air energy storage device sub-model includes a compressor model, an oil cooling device model, an air storage tank model, a turbine model, a high and low temperature oil heat storage tank model, and a heating device model.

[0038] Specifically, the compressor model represents the operating power consumption of the r-th stage compressor at time t in the compressed air energy storage device. As shown in equation (1): (1) in, The air insulation coefficient, for t The air mass flow rate that constantly flows into the r-th stage compressor. For the corresponding gas constant, for t Time of the first r The inlet temperature of the stage compressor, For the first r The compression ratio of a stage compressor.

[0039] The power consumption limit of the compressor is shown in equation (2): (2) in, for t The upper limit of the power consumption of the compressor at all times. for t The lower limit of the power consumption of the compressor at any given time.

[0040] The upper and lower limits of the mass flow rate of the air entering the compressor are shown in equation (3): (3) in, for t The upper limit of the power consumption of the compressor at all times. for t The lower limit of the electrical power consumed by the compressor at any given time.

[0041] t Time of the first r Inlet and outlet air temperatures of the stage compressor The relationship is: (4) The oil cooling device model is represented by equations (5) and (6): (5) (6) in, The heat absorbed by the oil cooling unit after the first-stage compressor. For the first The heat absorbed by the oil cooling unit after the compressor stage. The specific heat at constant pressure of air. and The first r and Stage compressor outlet temperature, For the first r +1 stage compressor inlet temperature, This is the standard temperature at the compressor outlet.

[0042] The gas storage model is represented by equation (7): (7) in, for t The pressure of the gas storage facility at all times. For the temperature of the gas storage facility, V For the gas storage capacity, This represents the initial gas outflow rate from the storage tank.

[0043] The pressure constraint of the gas storage facility is: (8) in, This is the lower limit of the gas storage pressure. This is the upper limit of the gas storage pressure.

[0044] The turbine model is represented by equation (9): (9) in, for t The turbine's power generation at all times for t The constant air mass flow rate entering the turbine. for t The inlet temperature of the time-limited turbine. This refers to the turbine expansion ratio.

[0045] The upper and lower limits of the power output of the turbine are constrained as shown in equation (10): (10) in, for t The lower limit of the power output of the constant turbine. for t The upper limit of the power output of a turbine at any given time.

[0046] The upper and lower limits of the air mass flow rate entering the turbine are: (11) in, This is the lower limit of the mass flow rate of air entering the turbine. This is the upper limit of the mass flow rate of air entering the turbine.

[0047] Turbine inlet temperature and outlet temperature The relationship is: (12) The high and low temperature oil thermal storage tank model is represented by equations (13) and (14): (13) (14) in, for t The heat stored in the thermal storage tank at all times for t External heating at all times This represents the lower limit of the heat storage capacity of the thermal storage tank. This is the upper limit of the heat storage tank's heat capacity.

[0048] The heating device model is represented by equation (15): (15) in, This refers to the heating amount of the heating device.

[0049] It is understandable that, considering that the actual output power of photovoltaic cells is mainly affected by solar irradiance and ambient temperature, the photovoltaic power generation system sub-model is represented by equations (16) to (18): (16) (17) (18) in, This refers to the actual output power of the photovoltaic cell. This represents the maximum output power of the photovoltaic cell. This refers to the rated output power of the photovoltaic cell under standard operating conditions (e.g., 25℃, 1.0MPa). These are real-time solar irradiance and solar irradiance under standard test conditions, respectively. These represent the surface temperature of the photovoltaic cell and the temperature of the photovoltaic cell under standard test conditions, respectively, in °C; k is the temperature coefficient. This refers to the ambient temperature, expressed in °C. For photovoltaic cells, the ambient temperature is 20℃ and the solar irradiance is 800W / m 2 The temperature reached when the wind speed is 1 m / s and under open-circuit conditions.

[0050] Since the output power of wind power is affected by wind speed, the output of wind turbine units is intermittent and fluctuating. Therefore, the sub-model of the wind turbine power generation system is represented by equation (19): (19) in, This refers to the active power output of the wind turbine. This refers to the rated power of the fan. For wind speed, The cut-in wind speed of the fan. This refers to the cut-off velocity of the fan. This refers to the cut-in wind speed of the fan.

[0051] Considering that pumped storage units are affected by factors such as head loss, pumping head loss, and turbine generator efficiency in actual operating conditions, and that there are losses during pumping and power generation, the subsystem model of the pumped storage power station is represented by equations (21) and (22): (20) (twenty one) in, and These refer to the pump efficiency and turbine efficiency of the pumped storage unit, respectively. For pumping head, For power generation head, and These represent the initial and final values ​​of the reservoir's capacity, respectively. This refers to the pumping power of the pumped storage unit. This refers to the power generation capacity of the pumped storage unit. Unit conversion factor.

[0052] The reservoir capacity constraint for pumped storage power stations is: (twenty two) in, scheduling period The innermost part represents the initial water volume of the upper reservoir of the pumped storage power station. Scheduling period The maximum and minimum water volume of the upper reservoir of the pumped storage power station. This is the average water / electricity conversion factor during pumping or power generation.

[0053] The power output constraint of a pumped storage power station is: (twenty three) in, This is the minimum power output of a pumped storage power station. This represents the maximum power output of the pumped storage power station.

[0054] The number of start-stop cycles for pumped storage units is limited to: (twenty four) in, and These represent the unit status when the pumped-storage unit is generating electricity and pumping water, respectively, with 1 indicating startup and 0 indicating shutdown. Limit the number of times each unit can be started and stopped per day.

[0055] The battery system sub-model is represented by equation (25): (25) in, This refers to the charging and discharging power of the battery; a value greater than 0 indicates discharging, and a value less than 0 indicates charging. This indicates the battery's charge status. This refers to a single scheduling cycle for the battery.

[0056] The battery's charge limit during operation is as follows: (26) in, These represent the maximum and minimum capacity of a sodium-ion battery.

[0057] The battery charging and discharging power constraints are as follows: (27) in, This represents the maximum power output for discharging and charging sodium-ion batteries.

[0058] The diesel engine generator electronic system model is represented by equations (28) and (29): (28) (29) in, and These are the unit's coal consumption cost and emission cost, respectively. For diesel engines in time Active power within, This represents the generator cost coefficient for diesel engines. It is the quantity of pollutant types. It is a pollutant j The unit quality control cost, It is a diesel engine m Pollutants j The emission coefficient.

[0059] The output constraint of the diesel generator set is: (30) in, This is the maximum output of the diesel generator set.

[0060] The diesel generator set output ramping constraint is: (31) in, For diesel generator set power ramp-up constraints.

[0061] Furthermore, coalbed methane is an important derivative resource in mining areas. By constructing a coalbed methane layer model and then combining it with the initial system model, a system model of the integrated energy system for the mining area can be obtained. In this system model, the combined power and heat (CHP) air compression energy storage can effectively utilize renewable energy, thereby reducing the output of diesel generator sets. Moreover, connecting renewable energy power generation and CHP power generation units to the power-heat network in a multi-energy complementary manner can effectively improve economic efficiency.

[0062] Specifically, the system model is as follows Figure 3As shown, the pressure swing adsorption (PSA) process mainly covers four stages: compression, filtration, purification, and recovery. In the initial stage, the coalbed methane enters the compressor through the regulating valve, and after being pressurized by the compressor, it flows into the gas storage tank. Subsequently, impurities are removed and humidity is reduced in the pretreatment stage. The pretreated gas enters the pressure swing adsorption tower, where gas components are separated. The separated N2 is pressurized by the secondary compressor of the RPU-CAES (Residual Pressure Utilization Compressed Air Energy Storage System) and stored in the gas storage tank, while CH4 is pressurized by the secondary compressor of the PSA and stored in the underground coalbed methane storage tank. The pressure swing adsorption (PSA) unit sub-model includes the adsorption tower model shown in equation (32): (32) in, The concentration of coalbed methane entering the adsorption tower. For reference conditions, the inlet gas concentration of coalbed methane This represents the methane recovery concentration. The methane recovery concentration is under reference conditions. For correction factors, For coalbed methane factors, The factor is methane.

[0063] Underground coalbed methane storage facilities need to meet constraints such as storage balance, storage volume, coalbed methane injection rate, coalbed methane extraction rate, equal storage volume at the beginning and end of the operation, and operating status during operation. Therefore, the relevant constraints for coalbed methane storage facilities are as follows: (33) in, for t Real-time gas storage capacity of underground coalbed methane storage facilities for t Real-time gas intake of underground coalbed methane storage facility for t Real-time gas output from underground coalbed methane storage facilities This is the minimum gas intake for an underground coalbed methane storage facility. This represents the maximum gas intake for the underground coalbed methane storage facility. This represents the minimum gas output of an underground coalbed methane storage facility. This represents the maximum gas output of the underground coalbed methane storage facility. This is the minimum gas storage capacity for underground coalbed methane storage facilities. This represents the maximum gas storage capacity of the underground coalbed methane storage facility.

[0064] Furthermore, by Figure 3As shown, the combined heat and power (CHP) unit consists of a CGT (cogeneration gas turbine). The CGT unit is a crucial energy-producing and multi-energy coupling link in the integrated energy system of the mine, not only tightly connecting the power and heating networks but also coupling it with an air energy storage system through underground coalbed methane storage and thermal storage tanks. Therefore, the CGT sub-model of the CHP system is the CGT model: (34) in, The CGT gas-to-electric reaction coefficient, The gas-to-heat reaction coefficient of CGT, Natural gas consumed by CGT To generate heat for CGT, Electricity generated for CGT This represents the maximum power generation of CGT.

[0065] S110. Based on the system model, construct a comprehensive energy operation model for the mining area. The constraint functions of the comprehensive energy operation model for the mining area include a net profit maximization function, a renewable energy consumption maximization function, and a carbon emission minimization function.

[0066] Specifically, the constraint functions of the integrated energy operation model for the mining area include a net profit maximization function, a renewable energy absorption maximization function, and a carbon emission minimization function. That is, a multi-objective function is established for joint operation, and the integrated energy operation model for the mining area is optimized from three aspects: economic optimization, renewable energy absorption maximization, and carbon emission minimization.

[0067] Among them, the economically optimal target of the integrated energy system in the mining area refers to the net benefit, that is, the difference between the interactive benefits and the various operating costs: (35) in, For the net revenue of the integrated energy system of the mining area, For PSA operating costs, To account for the operating costs of compressed air energy storage systems, For the operating costs of the power-heat network, This is for the interactive benefits between the integrated energy system and the electricity-heat network in the mining area.

[0068] Furthermore, the operating costs of PSA mainly consist of the operating costs of pressure swing adsorption (PSA) and underground coalbed methane storage. Therefore, the operating cost of PSA is: (36) (37) in, To reduce the operating costs of pressure swing adsorption, The operating cost of underground coalbed methane storage facilities, The mass flow rate of methane entering the coalbed methane storage facility. The density of coalbed methane, The unit maintenance cost of coalbed methane storage facilities. For PSA operating costs, T For a single scheduling cycle of PSA.

[0069] The operating cost of compressed air energy storage includes the costs of the compressor, underground compressed air storage tank, and thermal storage tank. Since the compressor cost is already included in the PSA cost, the operating cost of compressed air energy storage is: (38) (39) in, For the operating costs of thermal storage tanks, The operating costs of underground compressed air storage facilities, The unit maintenance cost of the thermal storage tank. The unit maintenance cost of underground compressed air storage facilities. This refers to the mass flow rate of air entering the compressed air storage tank.

[0070] Grid operating costs include diesel generator fuel costs: (40) The interaction between the integrated energy system in the mining area and the power-heat network is mainly reflected in grid transactions and heat network transactions, and the target of its interaction benefits is: (41) (42) in, For the benefit of grid interaction, For the interactive revenue of the hot network, For the electricity purchase cost of heat pumps, For grid interaction power, For the interaction power of the heating network, For electricity sales price, For electricity purchase price, For the price of heat, This refers to the electricity that the heat pump purchases from the power grid.

[0071] The renewable energy consumption index can be represented by the penalty cost for wind and solar curtailment: (43) in, This is the wind curtailment penalty coefficient. This is the penalty coefficient for discarded light. For wind curtailment power, This refers to the power of abandoned light.

[0072] There are two ways to reduce carbon emissions in a mining area's integrated energy system. One is to maximize the utilization of renewable energy through RPU-CAES, thereby reducing diesel generator output and directly reducing fossil fuel combustion. The other is to recover and utilize coalbed methane that was originally directly discharged into the atmosphere, thereby reducing carbon emissions caused by CH4 discharge. (44) in, The coalbed methane conversion coefficient, The reduction in carbon emissions from reduced diesel generator output due to the adoption of renewable energy.

[0073] S120. The integrated energy operation model of the mining area is converted into a Markov decision process to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system of the mining area. The scheduling strategy includes the output power range and equipment status of each device of the integrated energy system of the mining area during the day-ahead scheduling cycle.

[0074] Specifically, the day-ahead scheduling of the integrated energy system in a mining area actually refers to the action response of each piece of equipment in the integrated energy system and the scheduling of the mine's coal production plan. The decision variables are the output power range and equipment status of each piece of equipment in the integrated energy system within the day-ahead scheduling cycle. Among them, based on Figure 3 The system model of the integrated energy system in the mining area shows that the equipment status includes: the range of energy storage battery capacity, the range of pumped water storage capacity, and the range of gas storage tank capacity.

[0075] Based on this, day-ahead dispatch control aims to generate power curves for equipment, water volume curves for pumped-storage reservoirs, charge curves for energy storage batteries, and gas volume curves for air storage for the integrated energy system of the mining area within the day-ahead dispatch cycle. That is, it needs to determine the state control requirements of the aforementioned equipment at each control point within the control cycle, which is a sequential decision problem. Therefore, day-ahead dispatch control is defined as having a 5-tuple... The Markov Decision Process (MDP) is used. Here, S and A represent the agent's state space and action space, respectively. It is the transition probability function. R It is a reward function. It is a discount factor used to balance the contributions of immediate rewards and future rewards.

[0076] In a preferred embodiment, the integrated energy operation model of the mining area is converted into a Markov decision process to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system of the mining area, including: The integrated energy operation model of the mining area is transformed into a Markov decision process, and the state space and action space of the agent in the Markov decision process, as well as the reward function, are defined respectively. The state space includes the current state of each device in the integrated energy system of the mining area, and the action space includes the output power and state of each device in the integrated energy system of the mining area at a preset date time scale. Train the agent; Based on the trained agent, a scheduling strategy for day-ahead scheduling control of the integrated energy system in the mining area is determined.

[0077] Specifically, combined Figure 3 The system model of the integrated energy system in the mining area shown includes the state space of the intelligent agents, which represents the current state of each device in the integrated energy system. Specifically, this includes the states of the pumped storage system, battery system, compressed air energy storage system, wind power system, diesel generator system, photovoltaic system, electrical load, and thermal load. (45) in, This refers to the status of the pumped storage system. The status of the battery system. This is the status of the compressed air energy storage power generation system. This refers to the status of the wind turbine power generation system. Status of the photovoltaic power generation system. This refers to the status of the diesel generator system. Under electrical load conditions, Under heat load conditions, d This refers to the current moment of the integrated energy system.

[0078] The action space of the intelligent agent refers to the output power and status of each device in the integrated energy system of the mining area at a preset date-ahead time scale, including: changes in the water volume of the pumped storage reservoir, the gas storage capacity of the compressed air energy storage tank, and the state of charge of the battery at the preset date-ahead time scale; and the power of wind power, diesel engines, photovoltaics, electrical load, and thermal load at the preset date-ahead time scale. (46) in, This refers to the change in water volume in pumped storage reservoirs. This refers to the change in the state of charge of the battery. This represents the change in the amount of compressed air stored in the compressed air energy storage tank. For fan power, For photovoltaic power, For diesel engine power, For electrical load power, This represents the heat load power.

[0079] At the same time, because in Markov decision-making, Ornstein-Uhlenbeck noise needs to be added to the action space to represent the randomness of the action space, that is: (47) in, For the Markov decision process d The amount of noise disturbance change during the step. Noise figure This is the noise frequency. Therefore, the actual operation... Represented as: (48) In a preferred embodiment, the reward function includes rewards corresponding to maximizing net revenue, maximizing renewable energy consumption, and minimizing carbon emissions of the integrated energy operation model of the mining area, as well as penalties for violating the constraints of the integrated energy operation model of the mining area.

[0080] Specifically, the agent continuously learns to maximize its cumulative reward within the scheduling cycle. To effectively guide the agent's learning, the reward function is set to include rewards corresponding to maximizing net revenue, maximizing renewable energy consumption, and minimizing carbon emissions according to the integrated energy operation model of the mining area, as well as penalties for violating the constraints of the integrated energy operation model of the mining area: (49) in, These are the weighting coefficients of the reward function. The penalty function for violating the constraint is: (50) in, For the ideal value of the equality constraint, To define the maximum and minimum values ​​of the inequality constraints, This is the actual value. The number of equality constraints and inequality constraints.

[0081] S130. Within the agreed scope of the scheduling strategy and the constraints of the integrated energy operation model of the mining area, with the goal of minimizing operating costs, mixed integer linear programming is used to carry out intraday scheduling control of the integrated energy system of the mining area.

[0082] Specifically, intraday scheduling control aims to achieve the operation of the integrated energy system in the mining area within a single day, i.e., to make real-time action responses based on equipment status. After determining the day-ahead scheduling strategy for the integrated energy system based on Markov decision processes—that is, the output power range and equipment status of each device within the day-ahead scheduling cycle—the control requirements for each device can be derived. Therefore, the operation of intraday scheduling control is limited by the output power and status range of the equipment under day-ahead scheduling control. Simultaneously, intraday scheduling control must also meet the constraints of the integrated energy system operation model. That is, through the scheduling optimization method for the mine microgrid provided in this embodiment, short-term intraday operation scheduling, i.e., intraday scheduling control, can be implemented under the premise of satisfying day-ahead scheduling control (i.e., long-term power range, battery capacity range, pumped storage water capacity range, and air storage gas capacity range), and the constraints of the integrated energy system operation model. For example, using one hour as the scheduling control time scale, the decision variable is the output power of each device in the integrated energy system.

[0083] In a preferred embodiment, the constraints of the integrated energy operation model for the mining area include: Power balance constraints, line transmission power constraints, heat pump constraints, heat pipeline constraints, heat load constraints, and heating network constraints.

[0084] Specifically, combined Figure 3 The system model shown has the following power balance constraints: (51) in, For the power generation of the diesel generator set, For the turbine's power generation capacity, The power generation capacity of the pumped storage power station. The charging and discharging power of the energy storage battery, This represents the actual grid-connected power of the wind turbine. This represents the actual grid-connected power of photovoltaic power. For line transmission power, For the power consumed by the compressor, For the power consumed by the heat pump, For grid load power, It is a diesel generator set. For turbine units, It is a pumped storage unit. For battery systems, For wind turbine units, For photovoltaic power generation systems, This is a collection of power transmission lines.

[0085] The line transmission power constraint is: (52) in, For the line l Minimum transmission power For the line l Maximum transmission power.

[0086] The heat pump constraint is: (53) in, for t Heat pump heat generation during specific time periods for t Heat generated by the time-of-use air energy storage system The specific heat at constant pressure of circulating water, Let i be the mass flow rate of the circulating water at node i. For nodes i The temperature of the water supply system For nodes i The temperature of the return water system For nodes i The lower limit of water supply temperature. For nodes i The upper limit of water supply temperature.

[0087] The constraints on the heat pipes are: (54) in, For supply and return pipelines b temperature, For supply and return pipelines b Traffic, This is the temperature loss coefficient of the pipeline. For pipelines b Length, The ambient temperature.

[0088] (55) in, For pipelines b Upper limit of mass flow rate of circulating water.

[0089] The heat load constraint is: (56) in, For heat load, for t Time period nodes i circulating water mass flow rate, The lowest temperature for return water. This is the highest temperature of the return water.

[0090] The constraints of the heating network are: (57) in, for t The heat generated by the RPU-CAES during the time period is the sum of the heat generated by the compressor and the heat generated by the CGT.

[0091] Specifically, intraday scheduling control is modeled as a mixed-integer linear programming (MILP) model, with the objective function being: (58) in, These are the weighting coefficients.

[0092] In this embodiment, by decomposing the long-term and short-term optimization problems into two distinct sub-problems—a Markov decision process for day-ahead long-term scheduling optimization and a mixed-integer linear programming approach for intraday short-term scheduling optimization—more precise control and effective coordination can be achieved across different operating timelines. By combining deep reinforcement learning with mixed-integer linear programming, the complexity of integrated mine energy systems can be addressed without the usual surge in variables due to multiple time scales and uncertainties. This method strictly adheres to operational constraints across multiple time scales, significantly improving decision-making accuracy and system reliability. The scheduling optimization method for mine microgrids provided in the above embodiment can optimize long-term and short-term power balance, effectively managing and transferring energy across different time scales by utilizing the dynamic capabilities of battery storage, pumped hydro storage, and compressed air storage, thereby improving the overall efficiency and responsiveness of the integrated mine energy system.

[0093] In a preferred embodiment, the scheduling optimization method for a mining microgrid further includes: The operating costs of executing intraday scheduling control are fed back into the Markov decision process to optimize the scheduling strategy.

[0094] In this embodiment, by feeding back the operating cost of intraday scheduling control to the Markov decision process, intraday scheduling control can guide the optimized operation of day-ahead scheduling control.

[0095] Specifically, in the scheduling optimization method for mining microgrids provided in the above embodiments, day-ahead scheduling control can set operational boundaries for intraday scheduling control, while intraday scheduling control, in turn, influences the optimal day-ahead scheduling strategy. That is, in the scheduling optimization method for mining microgrids provided in this embodiment, the mixed-integer linear programming of intraday scheduling control is embedded as an environmental model into the Markov decision process of long-term scheduling control. This allows the formation of the Markov decision process to consider the interaction between long-term and short-term operations. Through this effective interaction, day-ahead scheduling and intraday scheduling can be effectively coordinated.

[0096] In some possible embodiments, based on the agent's actions in step d of the Markov decision process—namely, the device output power and device state—the device state control requirements at the beginning and end of the d-th control cycle can be represented as part of the environmental state. On the other hand, the training data of the d-th control cycle is also considered part of the environmental state and used for intraday scheduling control. Furthermore, to reduce computational burden, the original training data (e.g., historical data of the operation of a mining area's integrated energy system) is processed into hourly data for W typical days using clustering methods such as k-means. Finally, based on operational constraints and data, a short-term scheduling problem can be solved to minimize operational costs, which are fed back into the reward of step d, thus enabling the Markov decision process to proceed to the next step. By iteratively executing the Markov decision process, experience in the interaction between the agent and the environment is effectively accumulated, which is beneficial for updates.

[0097] In a preferred embodiment, training the agent includes: The agent is trained using a dual-delay deep deterministic policy gradient algorithm.

[0098] In this embodiment, considering that both the state space and action space are continuous spaces, a dual-delay deep deterministic policy gradient algorithm is used to train the agent. The dual-delay deep deterministic policy gradient algorithm is a policy evaluation network method that includes two policy networks and four evaluation networks. The relationships between the networks are as follows: Figure 4 As shown, the policy network and evaluation network structures are as follows: Figure 5 As shown, its network model uses a dual-deep Q-network algorithm, and the policy learning process is as follows: 1. Initialize historical data cache B.

[0099] 2. Initialize the evaluation network using random parameters. and policy network .

[0100] 3. Parameters of the target evaluation network Initialize to the same value , target policy network parameters Initialize to the same value .

[0101] 4. For each scheduling cycle on the preset day-ahead timescale, initialize the state and operational noise of the integrated energy system in the mining area. x 5. Randomly select actions and perform actions .

[0102] 6. Solve the optimal scheduling model (58) on a short time scale to obtain the state of the next integrated energy system in the mining area. and reward function value 7. Store in B, and randomly draw M empirical samples from B.

[0103] 8. Input M empirical samples into the target policy network after noise smoothing to obtain the empirical action policy. And input the smoothed target value evaluation network.

[0104] 9. By minimizing the loss function L Come to Figure 3 The parameters of the two smoothed target evaluation networks are updated, and then updated to... Figure 4 Evaluation network in China: (59) (60) in, The parameters are for the policy network and the evaluation network.

[0105] 10. The Q-value of the evaluation result obtained through the evaluation network. Using value function J Update policy network: (61) (62) 11. Update target network parameters: (63) in, This is the network smoothing coefficient.

[0106] In a preferred embodiment, training the agent further includes: Obtain the first historical state space of the integrated energy system of the mining area and the first historical action space corresponding to the first historical state space within each scheduling cycle of the preset date time scale; Add a preset perturbation factor to the first historical state space to obtain the third historical state space; Obtain the second historical state space and the historical value of the reward function corresponding to the second historical state space. The second historical state space is the state space corresponding to intraday scheduling control determined under the constraints of the agreed range and conditions of the first historical action space, with the goal of minimizing operating costs. The third historical state space, the first historical action space, the second historical state space, and the historical values ​​of the establishment function are used as data samples to train the agent.

[0107] Specifically, considering the multiple uncertainties in the integrated energy system of the mining area, such as the uncertainty of renewable energy sources like solar and wind power, the uncertainty of electrical load, and the uncertainty of heat load, by adding a preset perturbation factor to the first historical state space, the influence of uncertainty perturbation can be applied to the initial training data, thereby improving the agent's ability to explore the environment and enabling the agent to adaptively learn the probability distribution of uncertainty sources.

[0108] In some possible embodiments, the Gaussian distribution in equation (64) is used to generate uncertainty disturbances in photovoltaic systems, wind power systems, electrical load systems, and pumped storage systems. : (64) in, The average value of the random disturbance. Let be the standard deviation of the random disturbance.

[0109] The scheduling optimization method for mine microgrids provided in the above embodiments combines deep reinforcement learning and mixed-integer linear programming for coordinated scheduling optimization of long-term and short-term multi-timescale integrated energy systems in mining areas. For example, the day-ahead scheduling cycle is 1-2 weeks, and the intraday scheduling cycle is 1-2 hours. First, a mining area electricity-heat integrated energy system architecture based on surplus pressure utilization compressed air energy storage is proposed, i.e., the system model of the mining area integrated energy system. Then, based on the unit models of the system model, and considering factors such as renewable resources and low carbon emissions, a mining area integrated energy system operation model is established. Subsequently, a coordinated optimization control framework for the mining area integrated energy system is proposed, which decouples the mining area integrated energy system operation model into a long-term day-ahead optimization operation problem and a short-term intraday rolling optimization operation problem. The long-term day-ahead optimization is based on deep reinforcement learning and uses a Markov decision process, while the intraday short-term scheduling optimization uses mixed-integer linear programming to make real-time action responses based on equipment status, thereby achieving more precise control and effective coordination between different operating timelines.

[0110] The following describes a scheduling optimization system for a mining microgrid provided by an embodiment of the present invention. The scheduling optimization system for a mining microgrid described below can be considered as a modular architecture for implementing the scheduling optimization method for a mining microgrid provided by the embodiment of the present invention; the following description can be referred to in conjunction with the above.

[0111] Optional, see Figure 6 , Figure 6 This is a structural block diagram of a scheduling optimization system for a mining microgrid provided in an embodiment of the present invention, as shown below. Figure 6 As shown, the system may include: System model building unit 10 is used to build a system model of the integrated energy system in the mining area; The operation model construction unit 20 is used to construct a comprehensive energy operation model for the mining area based on the system model. The constraint functions of the comprehensive energy operation model for the mining area include a net profit maximization function, a renewable energy consumption maximization function, and a carbon emission minimization function. The day-ahead dispatch control unit 30 is used to convert the mine area integrated energy operation model into a Markov decision process, and to determine the dispatch strategy for day-ahead dispatch control of the mine area integrated energy system. The dispatch strategy includes the output power range and equipment status of each device in the mine area integrated energy system during the day-ahead dispatch cycle. The intraday scheduling control unit 40 is used to perform intraday scheduling control of the integrated energy system of the mining area with the goal of minimizing operating costs, within the agreed scope of the scheduling strategy and the constraints of the integrated energy operation model of the mining area.

[0112] Optional, such as Figure 7 As shown, it also includes: Feedback unit 50 is used to feed back the operating costs of executing intraday scheduling control to the Markov decision process in order to optimize the scheduling strategy.

[0113] Optionally, system model building unit 10 is specifically used for: Obtain a system model of the comprehensive energy infrastructure system in the mining area as the initial system model. The initial system model includes a compressed air energy storage device sub-model, a photovoltaic power generation system sub-model, a wind turbine power generation system sub-model, a pumped storage power station system sub-model, a battery system sub-model, and a diesel engine system sub-model. A coal gas reservoir model is constructed, which includes a pressure swing adsorption unit sub-model and a combined power and heat system sub-model. By incorporating a composite gas layer model into the initial system model, a system model of the integrated energy system of the mining area is obtained.

[0114] Optional constraints include: Power balance constraints, line transmission power constraints, heat pump constraints, heat pipeline constraints, heat load constraints, and heating network constraints.

[0115] Optionally, the day-ahead dispatch control unit 30 is specifically used for: The integrated energy operation model of the mining area is converted into a Markov decision process to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system of the mining area, including: The integrated energy operation model of the mining area is transformed into a Markov decision process, and the state space and action space of the agent in the Markov decision process, as well as the reward function, are defined respectively. The state space includes the current state of each device in the integrated energy system of the mining area, and the action space includes the output power and state of each device in the integrated energy system of the mining area at a preset date time scale. Train the agent; Based on the trained agent, a scheduling strategy for day-ahead scheduling control of the integrated energy system in the mining area is determined.

[0116] Optionally, the reward function includes rewards corresponding to maximizing net revenue, maximizing renewable energy consumption, and minimizing carbon emissions for the integrated energy operation model of the mining area, as well as penalties for violating the constraints of the integrated energy operation model of the mining area.

[0117] Optionally, the day-ahead dispatch control unit 30 is also specifically used for: The agent is trained using a dual-delay deep deterministic policy gradient algorithm.

[0118] Optionally, the day-ahead dispatch control unit 30 is also specifically used for: Obtain the first historical state space of the integrated energy system of the mining area and the first historical action space corresponding to the first historical state space within each scheduling cycle of the preset date time scale; Add a preset perturbation factor to the first historical state space to obtain the third historical state space; Obtain the second historical state space and the historical value of the reward function corresponding to the second historical state space. The second historical state space is the state space corresponding to intraday scheduling control determined under the constraints of the agreed range and conditions of the first historical action space, with the goal of minimizing operating costs. The third historical state space, the first historical action space, the second historical state space, and the historical values ​​of the establishment function are used as data samples to train the agent.

[0119] Optionally, embodiments of the present invention also provide a mining microgrid, including a mining microgrid body and a scheduling optimization system for the mining microgrid as provided in any of the above embodiments.

[0120] Below, for reference Figure 8 The electronic device provided in the embodiments of this application can be described as follows: at least one processor 100, at least one communication interface 200, at least one memory 300 and at least one communication bus 400; In this embodiment of the invention, the number of processor 100, communication interface 200, memory 300, and communication bus 400 is at least one, and the processor 100, communication interface 200, and memory 300 communicate with each other through communication bus 400; obviously, Figure 8 The communication connections shown for the processor 100, communication interface 200, memory 300, and communication bus 400 are optional. Optionally, the communication interface 200 can be an interface of a communication module, such as the interface of a GSM module; the processor 100 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0121] The memory 300 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0122] Specifically, the processor 100 is used to execute the application program in the memory to implement the steps of the above-mentioned scheduling optimization method for the mine microgrid.

[0123] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0124] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0125] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0126] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0127] It should be understood that the qualifiers “first,” “second,” “third,” “fourth,” “fifth,” and “sixth” used in the description of the embodiments of this application are only used to more clearly illustrate the technical solutions and are not intended to limit the scope of protection of this application.

[0128] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A scheduling optimization method of a mine micro-grid, characterized in that, The method comprises the following steps: constructing a system model of a mine area comprehensive energy system; based on the system model, constructing a mine area comprehensive energy operation model, wherein a constraint function of the mine area comprehensive energy operation model comprises a net income maximization function, a renewable energy consumption maximization function, and a carbon emission minimization function; converting the mine area comprehensive energy operation model into a Markov decision process, to determine a scheduling strategy for day-ahead scheduling control of the mine area comprehensive energy system, wherein the scheduling strategy comprises an output power range and a device state of each device of the mine area comprehensive energy system in a day-ahead scheduling period; under the limitation of a predetermined range of the scheduling strategy and a constraint condition of the mine area comprehensive energy operation model, performing day-ahead scheduling control of the mine area comprehensive energy system by using a mixed integer linear programming to minimize operation cost.

2. The method of claim 1, wherein, The method further comprises the following steps: feeding back operation cost of performing the day-ahead scheduling control to the Markov decision process, to optimize the scheduling strategy.

3. The method of claim 1, wherein, The step of constructing the system model of the mine area comprehensive energy system comprises the following steps: obtaining a system model of a mine area comprehensive energy infrastructure system as an initial system model, wherein the initial system model comprises a compressed air energy storage device submodel, a photovoltaic power generation system submodel, a wind turbine power generation system submodel, a pumped storage power station system submodel, a battery system submodel, and a diesel engine system submodel; constructing a coal gas layer model, wherein the coal gas layer model comprises a pressure swing adsorption unit submodel and an electric heat cogeneration system submodel; compositing the coal gas layer model on the initial system model, to obtain the system model of the mine area comprehensive energy system.

4. The method of claim 3, wherein, The constraint condition comprises the following steps: a power grid power balance constraint, a line transmission power constraint, a heat pump constraint, a heat pipeline constraint, a heat load constraint, and a heat network constraint.

5. The method of claim 1, wherein, The step of converting the mine area comprehensive energy operation model into a Markov decision process, to determine a scheduling strategy for day-ahead scheduling control of the mine area comprehensive energy system, comprises the following steps: converting the mine area comprehensive energy operation model into a Markov decision process, and defining a state space and an action space of an agent of the Markov decision process, and a reward function, respectively, wherein the state space comprises a current state of each device of the mine area comprehensive energy system, the action space comprises an output power and a state of each device of the mine area comprehensive energy system in a preset day-ahead time scale; training the agent; based on the trained agent, determining a scheduling strategy for day-ahead scheduling control of the mine area comprehensive energy system.

6. The method of claim 5, wherein, The reward function comprises a reward corresponding to net income maximization, renewable energy consumption maximization, and carbon emission minimization of the mine area comprehensive energy operation model, and a penalty for violating the constraint condition of the mine area comprehensive energy operation model.

7. The method of claim 6, wherein, The step of training the agent comprises the following steps: training the agent by using a double-delay deep deterministic policy gradient algorithm.

8. The method of claim 7, wherein, The step of training the agent further comprises the following steps: obtaining a first historical state space of the mine area comprehensive energy system and a first historical action space corresponding to the first historical state space in each scheduling period of the preset day-ahead time scale; A preset perturbation factor is added to the first historical state space to obtain the third historical state space; Obtain a second historical state space and a historical value of the reward function corresponding to the second historical state space. The second historical state space is a state space corresponding to the intraday scheduling control, determined within the agreed range of the first historical action space and under the constraints of the said conditions, with the goal of minimizing operating costs. The third historical state space, the first historical action space, the second historical state space, and the historical values ​​of the establishment function are used as data samples to train the agent.

9. A dispatch optimization system of a mine microgrid, characterized in that, include: The system model building unit is used to construct a system model of the integrated energy system in the mining area. The operation model construction unit is used to construct a comprehensive energy operation model for the mining area based on the system model. The constraint functions of the comprehensive energy operation model for the mining area include a net profit maximization function, a renewable energy consumption maximization function, and a carbon emission minimization function. The day-ahead scheduling control unit is used to convert the integrated energy operation model of the mining area into a Markov decision process, and to determine the scheduling strategy for day-ahead scheduling control of the integrated energy system of the mining area. The scheduling strategy includes the output power range and equipment status of each device of the integrated energy system of the mining area during the day-ahead scheduling cycle. The intraday scheduling control unit is used to perform intraday scheduling control of the integrated energy system of the mining area under the constraints of the agreed scope of the scheduling strategy and the constraints of the integrated energy operation model of the mining area, with the goal of minimizing operating costs, using mixed integer linear programming.

10. The system of claim 9, wherein, Also includes: The feedback unit is used to feed back the operating cost of executing the intraday scheduling control to the Markov decision process in order to optimize the scheduling strategy.

Images