Two-stage optimization scheduling method for mine energy system

Abstract

The invention discloses a two-stage optimization scheduling method for a mine energy system. The method comprises the following steps: constructing a mine comprehensive energy system model; establishing a distributed photovoltaic output uncertainty model, quantitatively predicting error space-time correlation through a gamma function and a Gaussian mixture distribution model, and decomposing a regional error by using a node weight coefficient; a distributed ADMM coordination algorithm is designed according to a distributed photovoltaic output uncertainty model, spatial-temporal correlation compensation is realized through a local layer error transfer function and coordination layer global variable updating, and a voltage out-of-limit probability constraint is introduced; and designing a day-ahead-day two-stage optimization scheduling framework, taking the minimum sum of the power generation cost of the gas turbine in the mine integrated energy system model and the photovoltaic light abandoning penalty as a day-ahead stage target, and dynamically adjusting the power deviation through virtual energy storage in the day-ahead stage. The method can effectively solve the problem of mine energy system scheduling under distributed photovoltaic access, and is suitable for a mine comprehensive energy scene with high-proportion new energy access.

Description

Technical Field

[0001] The present invention belongs to the technical field of integrated energy system optimization, and in particular relates to a two-stage optimization scheduling method for a mine energy system. Background Art

[0002] As a highly energy-intensive industrial sector, mining energy systems must meet multiple demands for production, daily life, and environmental governance. Traditional mining energy systems primarily rely on coal-fired power generation and grid power, resulting in high carbon emissions and low energy efficiency. In recent years, with the promotion of new energy technologies, distributed photovoltaic systems have been introduced in mines to achieve a clean energy transition. However, mines are characterized by complex geographical environments, volatile loads, and multiple challenges in integrating new energy. Mines are often located in remote areas, and the output of distributed photovoltaic systems is significantly affected by weather, terrain obstruction, and dusty environments, resulting in large prediction errors. Traditional deterministic scheduling methods cannot effectively quantify the randomness of photovoltaic power generation, leading to high curtailment rates or redundant backup capacity, increased operating costs, and even safety issues such as voltage overshoot. The synergistic utilization of mining-derived resources and new energy sources is still insufficient. The thermoelectric coupling characteristics of equipment such as wind-lack oxidation power generation and gas power generation have not been fully explored. There is a lack of dynamic matching mechanisms for waste heat recovery and thermal storage devices, making it difficult to achieve flexible source-load-storage interaction. Existing error compensation methods often focus on single nodes and ignore the spatiotemporal correlations of regional photovoltaic clusters. Traditional centralized optimization algorithms struggle to adapt to the decentralized control needs of distributed energy resources in mines, and they suffer from high computational complexity and poor real-time performance. Current research focuses primarily on optimizing single energy devices or improving short-term prediction algorithms, lacking a comprehensive energy system coordination framework tailored to complex mining scenarios. Especially with the high penetration of new energy, balancing economic, safety, and environmental benefits through data-driven modeling and dynamic compensation mechanisms remains a pressing technical challenge. Summary of the Invention

[0003] The purpose of the present invention is to provide a two-stage optimization scheduling method for a mine energy system, which can effectively solve the problem of mine energy system scheduling under distributed photovoltaic access and is suitable for mine comprehensive energy scenarios with a high proportion of new energy access.

[0004] To achieve the above objectives, the present invention provides a two-stage optimization scheduling method for a mine energy system, comprising the following steps:

[0005] Construct a comprehensive mine energy system model, including a ventilation oxidation power generation unit, a gas power generation unit, a water source heat pump, a waste heat boiler, and a thermal energy storage module;

[0006] A distributed photovoltaic output uncertainty model is established, the spatiotemporal correlation of prediction errors is quantified using the gamma function and Gaussian mixture distribution model, and regional errors are decomposed using node weight coefficients.

[0007] Based on the distributed photovoltaic output uncertainty model, a distributed ADMM coordination algorithm is designed. The temporal and spatial correlation compensation is achieved through the local layer error transfer function and the coordination layer global variable update, and the voltage over-limit probability constraint is introduced.

[0008] A two-stage optimization scheduling framework, day-ahead and intraday, is designed. The sum of the gas turbine power generation cost and the photovoltaic curtailment penalty in the mine integrated energy system model is minimized as the day-ahead target. During the intraday stage, power deviation is dynamically adjusted through virtual energy storage.

[0009] As a further solution of the present invention: In the mine integrated energy system model, catalytic oxidation technology is used to treat low-concentration gas in mine ventilation air. The reaction releases heat and produces high-temperature flue gas. The waste heat is recovered to provide a heat source for subsequent links. A ventilation air oxidation power generation unit is constructed with a steam turbine unit. Its output power is expressed as:

[0010]

[0011] Among them, P TFRR,t is the output power of the wind-deficient oxidation power generation unit at time t, η TFRR is the power generation efficiency of the wind-lack oxidation power generation unit, Gas calorific value, δ V,t is the gas concentration in the ventilation air at time t, V V,t is the lack of air flow at time t.

[0012] As a further solution of the present invention: In the mine integrated energy system model, for the high concentration of gas in the mine, a gas turbine is selected as the power generation equipment to construct a gas power generation unit, and its output power is expressed as:

[0013]

[0014] Among them, P GT,t is the gas turbine output power at time t, η GET is the gas turbine power generation efficiency, is the gas consumption at time t, η loss is the self-consumption rate of mining gas turbine, It is the low calorific value of gas.

[0015] As a further solution of the present invention: In the mine integrated energy system model, mine water is used as the cold and hot sources of the water source heat pump. The heating and cooling power of the water source heat pump are expressed as:

[0016]

[0017] Among them, Q heat,t is the heating power at time t, Q cool,t is the cooling power at time t, COP heatis the heating efficiency coefficient, COP cool is the refrigeration energy efficiency coefficient, P HP,e,t is the power consumption of the water source heat pump at time t.

[0018] As a further solution of the present invention: In the mine integrated energy system model, the waste heat boiler collects a large amount of waste heat generated by mine gas and wind power generation, and uses a hydrothermal cycle to supply the heat demand load. The waste heat boiler recovery power is expressed as:

[0019]

[0020] Among them, Q WH,t is the waste heat boiler recovery power at time t, η B is the thermal efficiency of the waste heat boiler, Q GT,W,t is the waste heat of gas turbine exhaust, η GT It is the heat utilization conversion efficiency of gas turbine.

[0021] As a further solution of the present invention: In the mine integrated energy system model, the operation constraints of the thermal energy storage module are expressed as follows:

[0022]

[0023] Among them, H sto,t is the heat storage capacity of the thermal energy storage module at time t, H sto,max is the upper limit of heat storage capacity of the thermal energy storage module at time t, H stod,t is the heat output of the thermal energy storage module at time t, H stod,max is the maximum heat output of the thermal energy storage module, Q sto,t is the thermal energy equivalent in the thermal energy storage module at time t, Q stod,t is the thermal energy output equivalent of the thermal energy storage module at time t, Q sto,max is the upper limit of thermal energy equivalent of the thermal energy storage module at time t, ΔQ sto,t+1 To predict the thermal energy dispatch demand at time t+1, Δt is the unit time interval, μ sto is the energy conversion coefficient of the thermal energy storage module, and T is the full optimization period;

[0024]

[0025] Considering the output delay constraint of thermal energy storage, ensure that the output power of the thermal energy storage module is within the delay τ BE After the effect, use Represents the output power after time delay, and introduces the auxiliary variable H of thermal energy storage output state sto,de,t .

[0026] As a further solution of the present invention: the distributed photovoltaic output uncertainty model is uniformly controlled by the mine integrated energy system model, and the distributed photovoltaic output expression at node i is:

[0027]

[0028] Among them, P PV,i,t is the distributed photovoltaic output at node i at time t, r t is the environmental impact factor at time t, α PV , β PV and ε PV is the fitting parameter, Γ(·) is the gamma function

[0029] As a further solution of the present invention: a distributed ADMM coordination algorithm is designed to further quantify the short-term prediction error characteristics of regional photovoltaic clusters based on the unified control of the distributed photovoltaic output uncertainty model by considering the mine integrated energy system model, and to establish an error compensation mechanism;

[0030] Assuming there are N photovoltaic nodes in the region, the prediction error can be decomposed into:

[0031]

[0032] Where ΔP PV,i,t is the photovoltaic output prediction error at node i, N is the total number of nodes, ω i is the weight coefficient of the photovoltaic installation node in the system, and the environmental impact factor r at time t t Related, ε corr,t is the coupling error;

[0033] The error probability density function can be obtained. Based on the gamma function Γ(·) in the photovoltaic fitting function, an improved gamma function and Gaussian mixture distribution model is constructed:

[0034]

[0035] Among them, λ is the gamma distribution weight, k PV and θ PV is the fitting parameter, is the standard deviation of the Gaussian term, which characterizes the measurement noise; N(·) is the Gaussian distribution function.

[0036] In the intraday stage, based on the real-time observation value of the error, the virtual energy storage in the mine integrated energy system model is dynamically corrected, and the correction power of the virtual energy storage is:

[0037]

[0038] Among them, P adj,t is the virtual energy storage correction power, η VS is the virtual energy storage adjustment coefficient, T c is the error-related time constant;

[0039] The distributed ADMM coordination algorithm optimizes the local layer, considers minimizing the prediction error, approximating global consistency, and updating the global variables of the coordination layer, and realizes the temporal and spatial correlation compensation of the regional photovoltaic prediction error. The distributed ADMM coordination regional node model is expressed as:

[0040]

[0041] Among them, Φ i (·) is the prediction error transfer function of node i, which is associated with the gamma function Γ(·), z k is the coordination layer global variable, is the dual variable of node i, ρ is the penalty coefficient, z k+1 Update variables for the coordination layer;

[0042] Based on the mine power system node model, the voltage deviation caused by the prediction error is derived:

[0043]

[0044] Where, ΔV i,t is the voltage deviation of node i at time t; ΔP PV,i,t is the photovoltaic prediction error; R l is the line resistance; V n is the reference voltage, is the symbol of partial derivative, r t is the environmental impact factor at time t;

[0045] Considering the impact of distributed power generation on power supply security, node voltage limit indicators are set:

[0046]

[0047] Among them, R V is the node voltage limit indicator, σ V is the voltage Gaussian distribution variance, Pr is the probability function symbol, and e is a natural constant.

[0048] As a further solution of the present invention, a two-stage optimization scheduling framework of day-ahead and intraday is designed. When scheduling the mine integrated energy system model, the objective function of the day-ahead economic scheduling stage is to minimize the sum of the day-ahead economic scheduling cost and the scheduling cost under real-time operation. The expression is:

[0049]

[0050] Among them, α GT , β GT 、c GT They represent the power generation cost function coefficients of the gas turbine unit, P GT,t is the output of the gas turbine unit, δPV is the penalty coefficient for photovoltaic curtailment, is the predicted value of photovoltaic power output at node i at time t;

[0051] The objective function of the intraday real-time scheduling stage is to minimize the scheduling cost. The objective function expression is:

[0052]

[0053] Among them, ρ i,t is the distributed photovoltaic output weight value of node i, C i,t is the unit scheduling cost of node i, y i,t is the real-time control value of node i.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The present invention can target the operating characteristics of the multi-source coupled energy system in mines, combine the spatiotemporal correlation of distributed photovoltaic output and the hysteresis of the dynamic response of energy storage, and construct a photovoltaic prediction error compensation mechanism based on the gamma function, Gaussian mixture distribution model, and distributed ADMM collaborative algorithm optimization theory to improve the error probability density fitting accuracy.

[0056] This invention builds a comprehensive mine energy system model to quantify the synergistic effect of wind-lack oxidation power generation, gas power generation waste heat recovery and water source heat pumps, thereby improving the waste heat utilization rate.

[0057] Aiming at the scenario of high penetration of new energy in mines, the present invention establishes a two-stage optimization framework of day-ahead economic scheduling and intraday real-time scheduling. Through virtual energy storage, it corrects power deviations in real time and constrains the probability of voltage exceeding the limit, thereby reducing the overall operating cost while ensuring power supply security. It solves the problems of low multi-energy coordination efficiency, high photovoltaic abandonment rate and delayed dynamic response of traditional mine energy systems, and provides an efficient and low-carbon energy scheduling solution for green mine construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is an implementation flow chart of the two-stage optimization scheduling method of the mine energy system of the present invention.

[0059] Figure 2 This is a flow chart of solving a two-stage problem using a method combining a column constraint generation algorithm with an ADMM algorithm. DETAILED DESCRIPTION

[0060] The present invention will be further described below by way of examples.

[0061] like Figure 1 As shown, a two-stage optimization scheduling method for a mine energy system includes the following steps:

[0062] Step 101: Based on the energy output mechanism of the mine energy system and the application of mine-derived resources, and in accordance with the mine production process, a mine comprehensive energy system model is constructed, including a ventilation oxidation power generation unit, a gas power generation unit, a water source heat pump, a waste heat boiler, and a thermal energy storage module.

[0063] As a mining-derived resource, ventilation air is treated with catalytic oxidation technology in the mine integrated energy system model. The reaction releases heat and produces high-temperature flue gas, which is then recovered through waste heat to provide a heat source for subsequent steps. The ventilation air oxidation power generation unit is combined with a steam turbine unit. Its output power is expressed as:

[0064]

[0065] Among them, P TFRR,t is the output power of the wind-deficient oxidation power generation unit at time t, η TFRR is the power generation efficiency of the wind-lack oxidation power generation unit, Gas calorific value, δ V,t is the gas concentration in the ventilation air at time t, V V,t is the lack of air flow at time t.

[0066] In response to the high concentration of gas in mines, gas turbines are selected as power generation equipment to build gas power generation units. The output power is expressed as:

[0067]

[0068] Among them, P GT,t is the gas turbine output power at time t, η GET is the power generation efficiency of the gas turbine, which is taken as 0.45, is the gas consumption at time t, η loss is the self-consumption rate of mining gas turbine, The low calorific value of gas is 35.8MJ / m 3 .

[0069] During mining, a large amount of mine water is discharged directly without being treated, causing serious damage to the surrounding ecological environment. The temperature of mine water is stable at around 15°C all year round. Using mine water as the cooling and heating source of water source heat pumps can avoid the consumption of traditional fossil energy and achieve emission reduction effects. The heating and cooling power of water source heat pumps are:

[0070]

[0071] Among them, Q heat,t is the heating power at time t, Q cool,t is the cooling power at time t, COP heat is the heating efficiency coefficient, take 3.6, COPcool is the refrigeration energy efficiency coefficient, take 2.8, P HP,e,t is the power consumption of the water source heat pump at time t.

[0072] The waste heat boiler is preferably a mining-type waste heat boiler without afterburning, which collects a large amount of waste heat generated by mine gas and wind power generation, and uses a hydrothermal cycle to supply the heat demand load; the waste heat boiler recovery power is expressed as:

[0073]

[0074] Among them, Q WH,t is the waste heat boiler recovery power at time t, η B is the thermal efficiency of the waste heat boiler, which is 0.85, Q GT,W,t is the waste heat of gas turbine exhaust, η GT It is the heat utilization conversion efficiency of gas turbine.

[0075] Thermal energy storage modules are typically deployed near heat loads to ensure thermal power balance. Thermal energy storage modules in mining systems function as transient flow modules in heat transmission, supplementing the heat network's supply capacity and providing both translation and regulation. The operational constraints of a mine thermal energy storage module are expressed as:

[0076]

[0077] Among them, H sto,t is the heat storage capacity of the thermal energy storage module at time t, H sto,max is the upper limit of heat storage capacity of the thermal energy storage module at time t, H stod,t is the heat output of the thermal energy storage module at time t, H stod,max is the maximum heat output of the thermal energy storage module, Q sto,t is the thermal energy equivalent in the thermal energy storage module at time t, Q stod,t is the thermal energy output equivalent of the thermal energy storage module at time t, Q sto,max is the upper limit of thermal energy equivalent of the thermal energy storage module at time t, ΔQ sto,t+1 To predict the thermal energy dispatch demand at time t+1, Δt is the unit time interval, μ sto is the energy conversion coefficient of the thermal energy storage module, and T is the full optimization period;

[0078]

[0079] Considering the output delay constraint of thermal energy storage, ensure that the output power of the thermal energy storage module is within the delay τ BE After the effect, use Represents the output power after time delay, and introduces the auxiliary variable H of thermal energy storage output state sto,de,t .

[0080] Step 102 : Establish a distributed photovoltaic output uncertainty model, quantify the spatiotemporal correlation of prediction errors using a gamma function and a Gaussian mixture distribution model, and decompose regional errors using node weight coefficients.

[0081] The distributed photovoltaic units around the mine have a large installed capacity and can provide clean electricity for the mine. However, they are affected by various external environmental factors, making it difficult to accurately determine the output power and thus impossible to formulate a reasonable scheduling plan. The distributed photovoltaic output uncertainty model is uniformly controlled by the mine integrated energy system model. The distributed photovoltaic output expression at node i is:

[0082]

[0083] Among them, P PV,i,t is the distributed photovoltaic output at node i at time t, r t is the environmental impact factor at time t, α PV , β PV and ε PV is the fitting parameter, Γ(·) is the gamma function.

[0084] Step 103 : Based on the distributed photovoltaic output uncertainty model, a distributed ADMM coordination algorithm is designed to achieve spatiotemporal correlation compensation through the local layer error transfer function and the coordination layer global variable update, and introduce voltage over-limit probability constraints.

[0085] Figure 2 This is a flow chart of solving a two-stage problem using a method combining a column constraint generation algorithm with an ADMM algorithm.

[0086] Step 201, initialization phase, initializing the upper and lower convergence limits, number of iterations and allowable deviation;

[0087] Step 202, setting pre-scheduling decision variables;

[0088] Step 203, data preprocessing stage, generates initial uncertainty probabilities and constructs output uncertainty sets;

[0089] Step 204: Optimize the day-ahead scheduling cost and check the deviation to determine whether the scheduling behavior is feasible.

[0090] Step 205: determine the uncertain probability set and reallocate the node scheduling amount to optimize the overall objective function;

[0091] Step 206: Update the number of iterations and determine if the parameter exceeds the limit.

[0092] Specifically, based on the unified regulation of the distributed photovoltaic output uncertainty model by the mine integrated energy system model, the short-term prediction error characteristics of the regional photovoltaic cluster are further quantified, and an error compensation mechanism is established;

[0093] Assuming there are N photovoltaic nodes in the region, the prediction error can be decomposed into:

[0094]

[0095] Where ΔP PV,i,t is the photovoltaic output prediction error at node i, N is the total number of nodes, ω i is the weight coefficient of the photovoltaic installation node in the system, and the environmental impact factor r at time t t Related, ε corr,t is the coupling error;

[0096] The error probability density function can be obtained. Based on the gamma function Γ(·) in the photovoltaic fitting function, an improved gamma function and Gaussian mixture distribution model is constructed:

[0097]

[0098] Among them, λ is the gamma distribution weight, k PV and θ PV is the fitting parameter, is the standard deviation of the Gaussian term, which characterizes the measurement noise; N(·) is the Gaussian distribution function.

[0099] In the intraday stage, based on the real-time observation value of the error, the virtual energy storage (such as ventilation system and ventilation system) in the mine integrated energy system model is dynamically corrected. The correction power of the virtual energy storage is:

[0100]

[0101] Among them, P adj,t is the virtual energy storage correction power, η VS is the virtual energy storage adjustment coefficient, T c is the error-related time constant;

[0102] The distributed ADMM coordination algorithm optimizes the local layer, considers minimizing the prediction error, approximating global consistency, and updating the global variables of the coordination layer, and realizes the temporal and spatial correlation compensation of the regional photovoltaic prediction error. The distributed ADMM coordination regional node model is expressed as:

[0103]

[0104] Among them, Φ i (·) is the prediction error transfer function of node i, which is associated with the gamma function Γ(·), z kis the coordination layer global variable, is the dual variable of node i, ρ is the penalty coefficient, z k+1 Update variables for the coordination layer;

[0105] Based on the mine power system node model, the voltage deviation caused by the prediction error is derived:

[0106]

[0107] Where, ΔV i,t is the voltage deviation of node i at time t; ΔP PV,i,t is the photovoltaic prediction error; R l is the line resistance; V n is the reference voltage, is the symbol of partial derivative, r t is the environmental impact factor at time t;

[0108] Considering the impact of distributed power generation on power supply security, node voltage limit indicators are set:

[0109]

[0110] Among them, R V is the node voltage limit indicator, σ V is the voltage Gaussian distribution variance, Pr is the probability function symbol, and e is a natural constant.

[0111] Step 104 , design a two-stage optimization scheduling framework, day-ahead and intraday, with minimizing the sum of the gas turbine power generation cost and the photovoltaic curtailment penalty in the mine integrated energy system model as the day-ahead target. In the intraday stage, dynamic power deviation is adjusted through virtual energy storage.

[0112] When scheduling the mine integrated energy system model, the objective function of the day-ahead economic scheduling stage is to minimize the sum of the day-ahead economic scheduling cost and the scheduling cost under real-time operation. The expression is:

[0113]

[0114] Among them, α GT , β GT 、c GT They represent the power generation cost function coefficients of the gas turbine unit, P GT,t is the output of the gas turbine unit, δ PV is the penalty coefficient for photovoltaic curtailment, is the predicted value of photovoltaic power output at node i at time t;

[0115] The objective function of the intraday real-time scheduling stage is to minimize the scheduling cost. The objective function expression is:

[0116]

[0117] Among them, ρ i,t is the distributed photovoltaic output weight value of node i, C i,t is the unit scheduling cost of node i, y i,t is the real-time control value of node i.

[0118] According to the mine comprehensive energy system model and the space-time compensation mechanism based on the distributed ADMM algorithm, the present invention utilizes the day-ahead-intraday two-stage optimization scheduling framework to determine a two-stage optimization scheduling method for the mine energy system that takes into account the prediction error of the distributed photovoltaic random quantity.

Claims

1. A two-stage optimization scheduling method for a mine energy system, characterized in that: The following steps are involved: Construct a comprehensive mine energy system model, including a ventilation oxidation power generation unit, a gas power generation unit, a water source heat pump, a waste heat boiler, and a thermal energy storage module; A distributed photovoltaic output uncertainty model is established, the spatiotemporal correlation of prediction errors is quantified using the gamma function and Gaussian mixture distribution model, and regional errors are decomposed using node weight coefficients. Based on the distributed photovoltaic output uncertainty model, a distributed ADMM coordination algorithm is designed. The temporal and spatial correlation compensation is achieved through the local layer error transfer function and the coordination layer global variable update, and the voltage over-limit probability constraint is introduced. A two-stage optimization scheduling framework, day-ahead and intraday, is designed. The sum of the gas turbine power generation cost and the photovoltaic curtailment penalty in the mine integrated energy system model is minimized as the day-ahead target. During the intraday stage, power deviation is dynamically adjusted through virtual energy storage.

2. A two-stage optimization scheduling method for a mine energy system according to claim 1, characterized in that: In the mine integrated energy system model, catalytic oxidation technology is used to treat low-concentration gas in mine ventilation air. The reaction releases heat and produces high-temperature flue gas. The waste heat is recovered to provide a heat source for subsequent links. A steam turbine unit is used to construct a ventilation air oxidation power generation unit. Its output power is expressed as: Among them, P TFRR,t is the output power of the wind-deficient oxidation power generation unit at time t, η TFRR is the power generation efficiency of the wind-lack oxidation power generation unit, is the gas calorific value, δ V,t is the gas concentration in the ventilation air at time t, V V,t is the lack of air flow at time t.

3. A two-stage optimization scheduling method for a mine energy system according to claim 2, characterized in that: In the mine integrated energy system model, gas turbines are used as power generation equipment to construct a gas power generation unit to address the high concentration of gas in the mine shaft. The output power is expressed as: Among them, P GT,t is the gas turbine output power at time t, η GET is the gas turbine power generation efficiency, is the gas consumption at time t, η loss is the self-consumption rate of mining gas turbine, It is the low calorific value of gas.

4. A two-stage optimization scheduling method for a mine energy system according to claim 1, characterized in that: In the mine integrated energy system model, mine water is used as the heat and cold source of the water source heat pump. The heating and cooling power of the water source heat pump are expressed as: Among them, Q heat,t is the heating power at time t, Q cool,t is the cooling power at time t, COP heat is the heating efficiency coefficient, COP cool is the refrigeration energy efficiency coefficient, P HP,e,t is the power consumption of the water source heat pump at time t.

5. A two-stage optimization scheduling method for a mine energy system according to claim 3, characterized in that: In the mine integrated energy system model, the waste heat boiler collects a large amount of waste heat generated by mine gas and wind power generation, and uses a hydrothermal cycle to supply the heat demand load. The waste heat boiler recovery power is expressed as: Among them, Q WH,t is the waste heat boiler recovery power at time t, η B is the thermal efficiency of the waste heat boiler, Q GT,W,t is the waste heat of gas turbine exhaust, η GT It is the heat utilization conversion efficiency of gas turbine.

6. A two-stage optimization scheduling method for a mine energy system according to claim 1, characterized in that: In the mine integrated energy system model, the operation constraints of the thermal energy storage module are expressed as: Among them, H sto,t is the heat storage capacity of the thermal energy storage module at time t, H sto,max is the upper limit of heat storage capacity of the thermal energy storage module at time t, H stod,t is the heat output of the thermal energy storage module at time t, H stod,max is the maximum heat output of the thermal energy storage module, Q sto,t is the thermal energy equivalent in the thermal energy storage module at time t, Q stod,t is the thermal energy output equivalent of the thermal energy storage module at time t, Q sto,max is the upper limit of thermal energy equivalent of the thermal energy storage module at time t, ΔQ sto,t+1 To predict the thermal energy dispatch demand at time t+1, Δt is the unit time interval, μ sto is the energy conversion coefficient of the thermal energy storage module, and T is the full optimization period; Considering the output delay constraint of thermal energy storage, ensure that the output power of the thermal energy storage module is within the delay τ BE After the effect, use Represents the output power after time delay, and introduces the auxiliary variable H of thermal energy storage output state sto,de,t .

7. A two-stage optimization scheduling method for a mine energy system according to claim 1, characterized in that: The distributed photovoltaic output uncertainty model is uniformly controlled by the mine integrated energy system model, and the distributed photovoltaic output expression at node i is: Among them, P PV,i,t is the distributed photovoltaic output at node i at time t, r t is the environmental impact factor at time t, α PV , β PV and ε PV is the fitting parameter, Γ(·) is the gamma function.

8. A two-stage optimization scheduling method for a mine energy system according to claim 7, characterized in that: Design a distributed ADMM coordination algorithm, taking into account the unified control of the distributed photovoltaic output uncertainty model of the mine integrated energy system model, further quantify the short-term prediction error characteristics of the regional photovoltaic cluster, and establish an error compensation mechanism; Assuming there are N photovoltaic nodes in the region, the prediction error can be decomposed into: Where ΔP PV,i,t is the photovoltaic output prediction error at node i, N is the total number of nodes, ω i is the weight coefficient of the photovoltaic installation node in the system, and the environmental impact factor r at time t t Related, ε corr,t is the coupling error; The error probability density function can be obtained. Based on the gamma function Γ(·) in the photovoltaic fitting function, an improved gamma function and Gaussian mixture distribution model is constructed: Among them, λ is the gamma distribution weight, k PV and θ PV is the fitting parameter, is the standard deviation of the Gaussian term, which characterizes the measurement noise; N(·) is the Gaussian distribution function. In the intraday stage, based on the real-time observation value of the error, the virtual energy storage in the mine integrated energy system model is dynamically corrected, and the correction power of the virtual energy storage is: Among them, P adj,t is the virtual energy storage correction power, η VS is the virtual energy storage adjustment coefficient, T c is the error-related time constant; The distributed ADMM coordination algorithm optimizes the local layer, considers minimizing the prediction error, approximating global consistency, and updating the global variables of the coordination layer, and realizes the temporal and spatial correlation compensation of the regional photovoltaic prediction error. The distributed ADMM coordination regional node model is expressed as: Among them, Φ i (·) is the prediction error transfer function of node i, which is associated with the gamma function Γ(·), z k is the coordination layer global variable, is the dual variable of node i, ρ is the penalty coefficient, z k+1 Update variables for the coordination layer; Based on the mine power system node model, the voltage deviation caused by the prediction error is derived: Where, ΔV i,t is the voltage deviation of node i at time t; ΔP PV,i,t is the photovoltaic prediction error; R l is the line resistance; V n is the reference voltage, is the symbol of partial derivative, r t is the environmental impact factor at time t; Considering the impact of distributed power generation on power supply security, node voltage limit indicators are set: Among them, R V is the node voltage limit indicator, σ V is the voltage Gaussian distribution variance, Pr is the probability function symbol, and e is a natural constant.

9. A two-stage optimization scheduling method for a mine energy system according to claim 1, characterized in that: A two-stage optimization scheduling framework, day-ahead and intraday, is designed. When scheduling the mine integrated energy system model, the objective function of the day-ahead economic scheduling stage is to minimize the sum of the day-ahead economic scheduling cost and the scheduling cost under real-time operation. The expression is: Among them, α GT , β GT 、c GT They represent the power generation cost function coefficients of the gas turbine unit, P GT,t is the output of the gas turbine unit, δ PV is the penalty coefficient for photovoltaic curtailment, is the predicted value of photovoltaic power output at node i at time t; The objective function of the intraday real-time scheduling stage is to minimize the scheduling cost. The objective function expression is: Among them, ρ i,t is the distributed photovoltaic output weight value of node i, C i,t is the unit scheduling cost of node i, y i,t is the real-time control value of node i.

Images