An optimal scheduling method for integrated energy system considering the role of carbon sources and sinks

Abstract

The present invention discloses an optimization scheduling method for an integrated energy system taking into account the role of carbon sources and sinks. The method establishes a mathematical model based on the carbon flow and energy flow process of energy coupling equipment in the integrated energy system, and analyzes the carbon source / sink role of the equipment in the system according to the characteristics of the equipment model, introduces biomass energy carbon capture and storage technology, and proposes a carbon source / sink evaluation index system for the system. Based on the index system, the corresponding incentive factors are determined and the adjustment costs are calculated. Based on this, a low-carbon scheduling model for the integrated energy system is proposed, and the optimal strategy is obtained through deep learning output prediction and DDPG algorithm training. The method proposed in the present invention reduces the carbon emissions of the system and reduces the impact of carbon reduction by subdividing the carbon emissions and capture and reuse processes in the integrated energy system, thereby improving environmental protection benefits and maintaining economic benefits. The present invention can be widely used in low-carbon scheduling of integrated energy systems.

Description

Technical Field

[0001] The present invention relates to the field of integrated energy system scheduling, and in particular to an integrated energy system optimization scheduling method taking into account the effects of carbon sources and sinks. Background Art

[0002] With the rapid development of social economy, human demand for energy is increasing, and energy shortage and environmental problems have also arisen. In this context, by organically combining various energy sources with different properties, coordinating and optimizing their production and conversion links, an integrated energy system has emerged to achieve efficient cascade utilization of energy and large-scale consumption of renewable energy. The application of the integrated energy system can realize the cascade utilization of multiple energy sources, and has the advantages of improving energy utilization efficiency, improving operation economy and reducing environmental pollution.

[0003] In recent years, the problem of optimizing the operation and dispatch of integrated energy systems has attracted widespread attention from scholars, and there are three main research ideas. First, for the capacity planning and operation of integrated energy systems, considering the impact of multiple energy storage devices such as electricity, heat and gas, time-of-use electricity prices and the comprehensive energy efficiency on the supply and demand sides, the optimal operation plan of the system is determined with multiple objectives such as economic benefits, environmental benefits and comprehensive energy efficiency; second, from the perspective of the overall planning of integrated energy systems and power grids, considering the number, location and equipment capacity configuration of energy stations, a two-level planning optimization model is proposed; third, a full-stage planning model is established for the entire life cycle of the integrated energy system, and the lowest-cost planning process and method are proposed.

[0004] In order to achieve the goals of "carbon peak" and "carbon neutrality", the energy industry accounts for a large proportion of carbon emissions and is the main force in energy conservation and emission reduction. To this end, scholars have gradually paid attention to carbon emissions in system planning, established various types of integrated energy system optimization models, and included carbon emissions in the scheduling planning of integrated energy systems in the form of transaction costs or penalty factors. However, the above-mentioned research on the day-ahead optimal scheduling of integrated energy systems only focuses on the total carbon emissions emitted by the system to the outside world, while ignoring the different roles of the equipment participating in the system's carbon flow under the multi-objective operation of the system. In the process of participating in the system's carbon emissions and carbon capture and reuse, the equipment in the integrated energy system has energy conversion and energy complementarity, so the relevant equipment can also participate in the system's consumption of new energy. In order to reduce the goal of carbon emissions, it may have too much impact on other goals, and further research is still needed. Summary of the invention

[0005] In order to solve the above problems, the present invention provides an optimization scheduling method for an integrated energy system taking into account the role of carbon source and sink. The carbon flow in the integrated energy system is considered separately according to the carbon source and carbon flow of the equipment, so as to avoid excessive impact of carbon reduction targets on other targets, which is conducive to flexible scheduling of equipment, fully reflecting the different roles of equipment under different targets, and introducing biomass carbon capture and storage technology between carbon source / sink equipment to enhance the flexibility of operation between equipment. By proposing an evaluation index for the role of carbon source / sink and adjusting the cost to affect the objective function, the regulation of equipment under multiple targets is more flexible, and the demand response role of the equipment is taken into account while the carbon source / sink role of the equipment is taken into account. The application of deep learning to predict the output range of the equipment is conducive to meeting the increasingly high requirements of the reliability and timeliness of equipment data under multi-dimensional and multi-target integrated energy systems. Finally, a more accurate identification is made through deep reinforcement learning, and the optimal decision is made, and through training, the decision time is finally reduced to achieve online regulation. The method of the present invention has great development prospects and application significance for the integrated energy system with a gradually increasing proportion of new energy systems, in terms of reducing system carbon emissions and optimizing system carbon structure while taking into account the goals of new energy consumption in the system.

[0006] In order to achieve the above object, the present invention is implemented by the following technical solutions:

[0007] The present invention is an optimization scheduling method for an integrated energy system taking into account the role of carbon sources and sinks, and the specific steps include: step 1, aiming at the multi-energy flow coupling conversion characteristics of natural gas flow-electricity flow and electricity flow-heat flow in the integrated energy system, and establishing the correlation between the carbon source / sink role of each device in the system carbon evolution process, introducing biomass energy carbon capture and storage technology, and establishing a mathematical model;

[0008] Step 2: Based on the correlation between the carbon source / sink effect of each device in the whole system and the carbon evolution process of the system, an evaluation index of the carbon source / sink effect is proposed, and the corresponding incentive factor is determined based on the index and the scheduling cost is calculated; and considering the correlation of device variables under multi-energy flow coupling, a scheduling framework including the whole system is proposed, and the scheduling cost determined by the carbon source / sink is introduced into the objective function, and a low-carbon operation model of the system is established according to the constraint conditions;

[0009] Step 3: Solve the system model using the hierarchical coordination method;

[0010] Step 4: Based on the mathematical model of the device, deep learning is applied to characterize the device at the data layer, and the device output range is predicted according to the associated constraints between the devices. The centralized training and decentralized execution strategy of deep reinforcement learning is used to implement online scheduling of each subject in the system in each cycle in a training-first-then-application manner. The further improvement of the present invention is that the mathematical model established in step 1 is specifically:

[0011] Diversion operation model of gas turbine under the action of carbon source:

[0012] The total thermal power of the gas turbine is expressed as:

[0013]

[0014] Where t is the time period, is the total thermal power consumed by the gas turbine during period t, ρ G is the density of natural gas, L G is the lower calorific value of natural gas, V t DGT is the natural gas consumption volume during period t, Δt is the time step, is the overall thermal efficiency of the gas turbine, P t DGT is the power generation of the gas turbine during period t, is the thermal-to-electricity conversion efficiency of the gas turbine, is the heating power of the gas turbine;

[0015] There is a coupling relationship between the power generation and heating power of the gas turbine. When the power is combined with heat, the ratio of the two powers is the power-to-heat ratio, and its physical model is:

[0016]

[0017]

[0018] In the formula, P t DGT They are the heating power, electric-to-heat ratio and power generation of the gas turbine during period t; the CO2 emission of the gas turbine and the power generation satisfy:

[0019]

[0020] In the formula, is the conversion ratio of gas turbine exhaust gas to natural gas consumption during period t, V t CDGT is the volume of CO2-containing gas produced by the gas turbine during period t, V t DGTC is the volume of CO2 gas emitted by the gas turbine during period t, is the CO2 concentration in the gas produced during period t, η r is the gas turbine gas cycle ratio, is the amount of CO2 produced by the gas turbine during period t, is the density of CO2;

[0021] If the CO2 concentration in the combustion chamber of the gas turbine is less than 40%, the CO2 concentration of the gas produced by the gas turbine is:

[0022]

[0023]

[0024] Carbon capture device CO2 absorption model:

[0025] The carbon capture device consumes electrical power P during the carbon capture process t CCS Electric power P from carbon capture t cCCS and the fixed electrical power P t fCCS The sum of:

[0026]

[0027] Where P t CCS is the power consumed by the carbon capture device during period t, V t O is the volume of exhaust gas captured by the carbon capture device during period t, is the power consumption coefficient affected by the capture amount and conversion process, the device fixed power P t fCS Set to 0.1 times P t cCCS .

[0028] Based on the leaf CO2 absorption process of the ecosystem carbon sink, the low-energy CO2 absorption capacity of the carbon capture device in period t is proposed. is the difference between the amount of CO2 in the captured gas and the amount not absorbed, expressed as:

[0029]

[0030] In the formula, They are the CO2 non-absorption efficiency coefficient under the device absorption solution concentration, gas pressure in the absorption tower, and temperature limits;

[0031] Biomass Carbon Capture and Storage:

[0032] The plant CO2 assimilation rate shown by the biochemical process is:

[0033]

[0034] Where V t a is the actual plant CO2 assimilation rate, V t 0is the rate of plant CO2 absorption, g0(T) and g1(T) are the sensitivity coefficients of stomatal temperature, g0(T) is the stomatal conductance when the assimilation rate of plant leaves is 0, Expressed as the air humidity around the leaf, P t a , P t c represent the partial pressures of air and CO2 in the leaf, respectively;

[0035] CO2 is distributed, accumulated, and circulated within plants through assimilation, and the final amount of carbon stored is expressed as:

[0036]

[0037] In the formula, represents the amount of carbon sequestered by plants, R i Represents the CO2 produced by plant growth and catabolism; CO2 reuse model of P2G device:

[0038] The volume of natural gas produced by P2G and the volume of CO2 recycled meet the following requirements:

[0039]

[0040]

[0041] Where V t P2G is the volume of natural gas produced by P2G during period t, is the CO2 conversion ratio of P2G, V t CR represents the volume of CO2 reused by P2G during period t, T represents the total number of scheduling periods, when P2G does not obtain CO2 from the market or other sources, and the amount of CO2 reused by P2G is not greater than the amount absorbed by the carbon capture device;

[0042] The P2G device consumes electrical power to convert CO2 back into methane, which is then supplied to the natural gas grid or gas storage tank for use. The natural gas output power of P2G during period t With natural gas volume V t P2G satisfy:

[0043]

[0044] Then the power consumption of the P2G device is P t P2G It is expressed as:

[0045]

[0046] In the formula, is the P2G power-to-gas conversion efficiency during period t;

[0047] Energy storage tank model:

[0048] The dynamic characteristics of the gas tank's charging and discharging are:

[0049]

[0050] In the formula, is the gas energy stored in the gas tank during period t, b1 and b2 are 0-1 variables, b1=1 or b2=1 indicates that the gas tank is in the state of inflation or deflation, is the natural gas storage or output power of the gas tank during period t;

[0051] The heat loss of the heat storage tank is related to the stored heat, and the dynamic characteristics of thermal energy are expressed as:

[0052]

[0053] In the formula, is the heat energy stored in the heat storage tank during period t, ε is the heat loss rate of the heat storage tank, d1 and d2 are 0-1 variables, d1=1 or d2=1 indicates that the heat storage tank is in a state of absorbing or releasing heat, is the thermal power of the heat storage tank during period t;

[0054] There is power loss in the battery during the process of charging, discharging and storing electric energy. Its dynamic characteristics of charging and discharging are expressed as:

[0055]

[0056] In the formula, is the electric energy stored in the battery during period t, a1 and a2 are 0-1 variables, a1=1 or a2=1 indicates that the battery is in the charging or discharging state, P t ES is the charging and discharging power of the battery during period t, are the charge and discharge efficiency of the battery respectively, and δ is the self-discharge rate of the battery.

[0057] The further improvement of the present invention is that the specific steps of step 2 are as follows:

[0058] Step 2.1: Propose carbon source control indicators and carbon sink monitoring indicators

[0059]

[0060]

[0061] Among them, P t DGT , Pt PV , P t WT The power generated by gas turbine, photovoltaic and wind power respectively; P t P2G , P t CCS , P t ES , P t EHB The electrical power consumed by the P2G device, carbon capture device, battery, and electric boiler are respectively;

[0062] Step 2.2, calculate the dispatch cost;

[0063] The dispatch cost established by the indicator is:

[0064]

[0065]

[0066] In the formula, is the carbon emission adjustment cost of the carbon source equipment in period t, f si,t is the carbon sink adjustment cost of the carbon capture device during period t, c SR represents the unit cost of carbon emissions from the carbon source equipment, c SI It represents the incentive price per unit volume of CO2 absorbed by the carbon sink device;

[0067] Step 2.3, introduce the dispatching cost calculated in step 2.2 into the objective function, and establish a low-carbon operation model of the system according to the constraints, specifically:

[0068] Gas energy flow network optimization objective function:

[0069]

[0070] Where f1 is the main operating cost of the natural gas network, T is the total number of optimal scheduling periods, c GD is the cost of purchasing unit volume of natural gas from the external gas grid. is the gas operation and maintenance cost of the gas turbine, c P2G The cost of supplying a unit volume of natural gas to P2G, c GS The operation and maintenance costs of filling and degassing the gas tank, is the natural gas output power of the external natural gas grid during period t;

[0071] Ignoring the natural gas loss inside the equipment, the natural gas supply obtained by the system from the external gas grid, gas storage tank and P2G during period t is equal to the natural gas demand V provided to the gas turbine C and the gas load. t PLG Satisfying the balance:

[0072] ρ G V t GD +ρ G V t GS +ρ G (B1V t P2G -b2V t P2G )=ρ G V t DGT +ρ G V t PLG

[0073] Converted into the balance constraint of natural gas input and output power:

[0074]

[0075] In the formula, The natural gas output power for the gas load;

[0076] Operating constraints of each device:

[0077]

[0078] In the formula, G i,t , are the natural gas operating power of each device i in the integrated energy system during period t and its minimum and maximum values;

[0079] Gas energy flow pipeline valve pressure difference constraints:

[0080]

[0081] In the formula, are the minimum and maximum pressure differences of valves of each device i in the gas energy flow pipeline, is the natural gas flow rate of each device i in period t, μ is the flow coefficient (determined by the valve), is the pipe area;

[0082] Gas tank capacity constraints:

[0083]

[0084] In the formula, are the minimum and maximum values ​​of gas energy respectively;

[0085] Electric-thermal energy flow network optimization objective function:

[0086]

[0087] Where f2 represents the operating cost of the electricity-heat energy flow network, f e Refers to the electricity dispatch and operation costs of photovoltaic, wind power, gas turbines, batteries, carbon capture and P2G, f h Refers to the heat energy dispatching and operating costs of electric boilers and heat storage tanks, f sr is the carbon emission adjustment fee, f si is the carbon sink adjustment fee, c PV 、c WT 、c DGT 、c CCS 、c P2G are the unit power generation costs of photovoltaic, wind power, gas turbine, carbon capture and P2G in period t, The operating cost of battery charging and discharging, are the unit heat storage and release operating costs of the electric boiler and the heat storage tank respectively. t PV , P t WT are the power generation of photovoltaic power and wind power in period t respectively;

[0088] Ignoring the internal power loss of the equipment, the system power balance relationship during period t is:

[0089] P t PV +P t WT +P t DGT +a1P t ES -a2P t ES =P t PLE +P t CCS +P t P2G +P t EHB

[0090] Where P t PLE is the power load value;

[0091] Ignoring the internal heat loss of the equipment, the heat supply and heat load demand of the system gas turbine, electric boiler and heat storage tank in period t are balanced:

[0092]

[0093] In the formula, is the heat load demand value.

[0094] Limit constraints for each equipment operation constraint:

[0095]

[0096] Where P i,t ,H i,t are the electrical power and thermal power of different equipment i in period t during the optimization process, are the minimum and maximum values ​​of the operating power of these devices in period t. At the same time, the operation of these devices meets the constraints;

[0097] Thermal energy flow pipeline valve pressure difference constraints:

[0098]

[0099] In the formula, are the minimum and maximum pressure differences of the valves of each device i in the heat flow pipeline, respectively, Q is the density of hot water, is the hot water flow rate of each device i in time period t;

[0100] Thermal energy flow pipeline temperature constraints:

[0101]

[0102] Where, T i min ,T i max are the minimum and maximum values ​​of the pipe temperature of each device i in the heat flow pipeline, respectively. i,t is the thermal power of each device i in time period t, C is the specific heat capacity of water, and T0 is the initial temperature of water;

[0103] Battery and heat storage tank capacity constraints:

[0104]

[0105] In the formula, are the minimum and maximum energy storage values ​​of the battery and the heat storage tank, respectively. A further improvement of the present invention is that step 3 is specifically:

[0106] Step 3.1, optimize the scheduling problem of the lower-level subsystem;

[0107] According to the initial forecast values ​​of the day-ahead dispatch of wind power generation, photovoltaic power generation, energy storage equipment and various energy loads in the integrated energy system, the augmented objective function of the gas energy flow network is established after adopting the Lagrange relaxation strategy for the network constraints:

[0108]

[0109] Among them, λv is the Lagrange multiplier vector of the gas energy flow network, σ v is its penalty factor, R v1 (x v ), R vi (x v )) are the equality constraints and inequality constraints of the gas energy flow network, i=2,···,m;

[0110] Establish the augmented objective function of the electrical-thermal energy flow network:

[0111]

[0112] Among them, λ e is the Lagrange multiplier vector of the electrical-thermal energy flow network, σ e is its penalty factor, R e1 (x e ), R e2 (x e ), R ei (x e ) are the equality constraints and inequality constraints of the electric energy flow network and the thermal energy flow network, respectively, j = 3, ..., l; after solving the optimization problems of the lower gas energy flow network and the electric-thermal energy flow network, the optimization solution is recorded as

[0113] Step 3.2, optimize the upper overall problem:

[0114] Feedback the optimization results of each network to the upper layer for overall Lagrange multiplier optimization correction, and establish the dual optimization problem of the upper system:

[0115]

[0116] When solving the upper-level dual optimization problem, the gradient method is used to update the unknown quantity λ v ,λ e :

[0117]

[0118] Step 3.3, check the convergence conditions:

[0119] Solve the lower-level subproblem and the upper-level overall dual problem, and optimize the system by alternating the upper and lower-level calculation processes until a global optimal solution is found;

[0120] Set the termination condition to:

[0121]

[0122] If φ k≤ε, ε>0 is the control error, then the iteration is stopped. Due to the existence of the duality gap, the obtained solution is a feasible solution for the global optimization of the original problem.

[0123] The beneficial effects of the present invention are: (1) The present invention converts the carbon source / sink effect index into a scheduling cost that affects the model objective function, and solves it based on the obtained model. The scheduling scheme obtained thereby will reduce the operating power of the gas turbine and increase the carbon capture device and the P2G device to a certain extent, and enhance the flexibility of the operating mode of each device under the demand of biomass carbon capture and storage technology, and effectively complement the energy storage equipment of each energy source to maintain the stable operation of the system, thereby reducing the carbon emissions of the system as a whole and improving its environmental benefits.

[0124] (2) The present invention proposes an integrated energy system optimization scheduling method that takes into account the role of carbon sources and sinks. When introducing the integrated energy system optimization scheduling model using relevant scheduling costs, due to the constraints of carbon source control and carbon sink monitoring indicators at the current moment, the carbon source / sink structure ratio at the current moment is optimized to a certain extent. The carbon capture device meets the demand response function while reducing the system's carbon emissions, improving the stability of the integrated energy system under the output of new energy, and preventing the occurrence of a small probability scenario that ignores the system's operating costs. While meeting the requirements for the operation of the integrated energy system, the carbon emissions and emission rate of the integrated energy system are effectively reduced, and the environmental protection effect of the system is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0125] Figure 1 Flow chart for implementing the present invention.

[0126] Figure 2 This is a schematic diagram of the basic structure of the integrated energy system;

[0127] Figure 3 Solve the flow chart for the integrated energy system dispatch model;

[0128] Figure 4 This is a comparison chart of system carbon emissions under various scenarios.

[0129] Figure 5 A comparison chart of the system carbon capture device operation under various scenarios. DETAILED DESCRIPTION

[0130] In order to make the purpose, technical solution and innovation of the present invention clearer, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0131] The present invention is a method for optimizing the scheduling of an integrated energy system taking into account the effects of carbon sources and sinks. A mathematical model is established that takes into account the multi-energy coupling characteristics of the equipment in the integrated energy system and the effects of the equipment in the carbon flow. Based on the mathematical model of the equipment, a systematic evaluation index for carbon sources / sinks is proposed. The adjustment cost is calculated through relevant incentive factors and introduced into the objective function of the system's low-carbon operation scheduling model. Deep reinforcement learning is used to centrally train the model and a feasible strategy for decentralized scheduling is proposed.

[0132] Take the gas-electricity-heat integrated energy system as an example. Figure 2 As shown in the figure, the system is mainly composed of wind power generation system, photovoltaic power generation system, gas turbine power generation system, carbon capture device, power-to-gas device (P2G), electric boiler, various energy storage equipment (batteries, heat storage tanks and gas storage tanks) and various energy loads (electricity load, heat load and gas load) of users in the system. The system is connected to the natural gas grid and the power grid to interact with natural gas and electricity. The specific steps are as follows:

[0133] Step 1: Aiming at the multi-energy flow coupling conversion characteristics of natural gas flow-electricity flow and electricity flow-heat flow in the integrated energy system, and establishing the correlation between the carbon source / sink of each device in the carbon evolution process of the system, introduce biomass energy carbon capture and storage technology, and establish a mathematical model;

[0134] Step 2: Based on the correlation between the carbon source / sink effect of each device in the whole system and the carbon evolution process of the system, an evaluation index of the carbon source / sink effect is proposed. Based on the index, the corresponding incentive factor is determined and the scheduling cost is calculated. The equipment variable correlation under multi-energy flow coupling is considered, and a scheduling framework including the whole system is proposed. The scheduling cost determined by the carbon source / sink is introduced into the objective function, and a low-carbon operation model of the system is established according to the constraint conditions.

[0135] Step 3: Solve the system model using the hierarchical coordination method;

[0136] Step 4: Based on the mathematical model of the equipment, deep learning is used to characterize the equipment at the data layer, and the equipment output range is predicted according to the correlation constraints between the equipment. The centralized training and decentralized execution strategy of deep reinforcement learning is used to implement online scheduling of each subject in the system in each cycle in a training-first-then-application manner.

[0137] The model established in step 1 is as follows:

[0138] The specific model is as follows:

[0139] 1) Diversion operation model of gas turbine under the action of carbon source.

[0140] Gas turbines convert gas energy into electrical energy and thermal energy. They completely burn natural gas to generate thermal power. and natural gas consumption volume Vt DGT satisfy:

[0141]

[0142] Where t is the time period, is the total thermal power consumed by the gas turbine during period t, ρ G is the density of natural gas, L G is the lower calorific value of natural gas, V t DGT is the volume of natural gas consumed during period t, and Δt is the time step.

[0143] The total thermal power of the gas turbine is expressed as:

[0144]

[0145] In the formula, is the overall thermal efficiency of the gas turbine, P t DGT is the power generation of the gas turbine during period t, is the thermal-to-electricity conversion efficiency of the gas turbine, is the heating power of the gas turbine.

[0146] There is a coupling relationship between the power generation and heating power of the gas turbine. When the power is combined with heat, the ratio of the two powers is the power-to-heat ratio, and its physical model is:

[0147]

[0148]

[0149] In the formula, λ t DGT , P t DGT They are the heating power, electric-to-heat ratio and power generation power of the gas turbine during period t.

[0150] As the main carbon source equipment, gas turbines will produce a large amount of greenhouse gas with high CO2 concentration while outputting electricity and heat energy. The volume of CO2-containing gas produced and the volume of natural gas consumed meet the following requirements:

[0151]

[0152] In the formula, is the conversion ratio of gas turbine exhaust gas to natural gas consumption during period t, V t CDGT is the volume of CO2-containing gas produced by the gas turbine during period t.

[0153] At this time, the volume of CO2 emitted by the gas turbine includes the volume of CO2 produced and the volume of CO2 in the recycled exhaust gas:

[0154]

[0155] Where V t DGTC is the volume of CO2 gas emitted by the gas turbine during period t. is the CO2 concentration in the gas produced during period t, η r is the gas turbine gas cycle ratio.

[0156] From the above formula, it can be obtained that the CO2 emission of the gas turbine and the power generation satisfy:

[0157]

[0158] In the formula, is the amount of CO2 produced by the gas turbine during period t, is the density of CO2.

[0159] At the same time, the CO2 concentration in the combustion chamber of the gas turbine should be less than 40% as much as possible to reduce the generation of pollutants. The CO2 concentration of the gas produced by the gas turbine should meet the following requirements:

[0160]

[0161] 2) CO2 absorption model of carbon capture device.

[0162] The carbon capture device consumes electrical power P during the carbon capture process t CCS Electric power P from carbon capture t cCCS and the fixed electrical power P t fCCS The sum of:

[0163]

[0164] Where P t CCS is the power consumed by the carbon capture device during period t, V t O is the volume of exhaust gas captured by the carbon capture device during period t, is the power consumption coefficient affected by the capture amount and conversion process, the device fixed power P t fCS Set to 0.1 times P t cCCS .

[0165] Based on the leaf CO2 absorption process of the ecosystem carbon sink, the low-energy CO2 absorption capacity of the carbon capture device in period t is proposed. is the difference between the amount of CO2 in the captured gas and the amount not absorbed, expressed as:

[0166]

[0167] In the formula, are the CO2 non-absorption efficiency coefficient under the device absorption solution concentration, gas pressure in the absorption tower, and temperature limits. The formula shows that when the proportion of CO2 in the exhaust gas exceeds 50%, it is beneficial to the low-energy capture of the carbon capture device. When the proportion of CO2 is lower than 50%, it will increase the difficulty of CO2 absorption in the carbon capture device and increase energy consumption.

[0168] 3) Biomass carbon capture and storage.

[0169] Plant photosynthesis depends on the CO2 utilization efficiency of the photosynthetic biochemical process and the supply of C7O2 diffused into the intercellular spaces of the leaves through the stomata. The plant CO2 assimilation rate shown by the biochemical process is:

[0170]

[0171] Where V t a is the actual plant CO2 assimilation rate, V t 0 is the rate at which plants absorb CO2, g0(T) and g1(T) are the sensitivity coefficients of stomata to temperature, and g0(T) is the stomatal conductance when the assimilation rate of plant leaves is 0. Expressed as the air humidity around the leaf, P t a , P t c Respectively represent the partial pressures of air and CO2 in the leaves.

[0172] CO2 is distributed, accumulated, and circulated within plants through assimilation, and the final amount of carbon stored can be expressed as:

[0173]

[0174] In the formula, represents the amount of carbon sequestered by plants, R i Represents CO2 produced by plant growth and catabolism.

[0175] The role of biomass carbon capture and storage is mainly to reduce the requirement for carbon capture devices to capture CO2 produced by gas turbine combustion operations, allowing equipment that acts as a carbon source / sink in the system to operate in a more flexible manner.

[0176] 4) CO2 reuse model of P2G device.

[0177] The CO2 captured by the carbon capture device can be used as raw material for the P2G device. The volume of natural gas produced by P2G and the volume of CO2 reused meet the following requirements:

[0178]

[0179]

[0180] Where V t P2G is the volume of natural gas produced by P2G during period t, is the CO2 conversion ratio of P2G, V t CR It represents the volume of CO2 reused by P2G in period t, and T represents the total number of scheduling periods. When P2G does not obtain CO2 from the market or other sources, the amount of CO2 reused by P2G is not greater than the amount absorbed by the carbon capture device.

[0181] The P2G device consumes electrical power to convert CO2 back into methane, which is then supplied to the natural gas grid or gas storage tank for use. The natural gas output power of P2G during period t With natural gas volume V t P2G satisfy:

[0182]

[0183] Then the power consumption of the P2G device is P t P2G It is expressed as:

[0184]

[0185] In the formula, is the P2G power-to-gas conversion efficiency during period t.

[0186] 5) Model of the energy storage tank.

[0187] The dynamic characteristics of the gas tank's charging and discharging are:

[0188]

[0189] In the formula, is the gas energy stored in the gas tank during period t, b1 and b2 are 0-1 variables, b1=1 or b2=1 indicates that the gas tank is in the state of inflation or deflation, is the natural gas storage or output capacity of the gas tank during period t.

[0190] The heat loss of the heat storage tank is related to the stored heat, but the heat loss rate is very small under good insulation conditions. The dynamic characteristics of thermal energy can be expressed as:

[0191]

[0192] In the formula, is the heat energy stored in the heat storage tank during period t, ε is the heat loss rate of the heat storage tank, d1 and d2 are 0-1 variables, d1=1 or d2=1 indicates that the heat storage tank is in a state of absorbing or releasing heat, is the thermal power of the heat storage tank during period t.

[0193] There is power loss in the battery during the process of charging, discharging and storing electric energy. Its dynamic characteristics of charging and discharging can be expressed as:

[0194]

[0195] In the formula, is the electric energy stored in the battery during period t, a1 and a2 are 0-1 variables, a1=1 or a2=1 indicates that the battery is in the charging or discharging state, P t ES is the charging and discharging power of the battery during period t, are the charge and discharge efficiency of the battery respectively, and δ is the self-discharge rate of the battery.

[0196] In step 2, the initial values ​​of each device are randomly generated based on the parameter settings of each device in the system, as shown in Table 1, and the carbon source / sink index is determined, and the scheduling cost determined by it is introduced into the system operation objective function;

[0197] Table 1 Parameter settings of each device in the system

[0198]

[0199]

[0200] The specific carbon source / sink indicators are:

[0201] 1) Carbon source control index. Gas turbines are the main source of CO2 emissions to the atmosphere within the system. At the same time, gas turbines, distributed photovoltaics, and wind power provide electricity to the entire system. In order to reduce carbon emissions and promote clean energy consumption, a carbon source control index is proposed to measure the ratio of carbon source equipment output power to distributed photovoltaic and wind power consumption, and a gas turbine carbon source control index for period t is proposed. To measure its carbon impact:

[0202]

[0203] The smaller the value of this indicator is, the lower the proportion of gas turbine usage in period t is. The smaller the proportion of carbon source is, the better the energy structure of the entire system is.

[0204] In order to promote carbon sources to reduce carbon emissions and the consumption of photovoltaic and wind power, it is proposed to use carbon source control indicators as incentive factors to calculate the carbon emission adjustment costs of carbon source equipment in the day-ahead dispatch period t.

[0205]

[0206] In the formula, c SR Represents the unit cost of carbon emissions from the carbon source equipment.

[0207] 2) Carbon sink monitoring indicators. The carbon capture device in the system mainly captures CO2 produced by gas turbines and some gas loads, such as chemical plants and steel plants. Not all CO2 in the system will enter the gas pipeline of the carbon capture device, but will be directly discharged into the atmosphere. At this time, it can only rely on ecological carbon sinks (such as greening). Therefore, the indicator is proposed To measure the carbon sink impact of carbon capture during period t:

[0208]

[0209] The smaller the value of the carbon sink monitoring index is, the better the capture effect of the carbon capture device is, indicating that the carbon sink effect of the carbon capture device is better under this scheduling operation mode.

[0210] In order to motivate its carbon sink performance in the system optimization scheduling, it is proposed to use the carbon sink monitoring index as an incentive factor to calculate the carbon sink adjustment cost f of the carbon capture device in period t. si,t :

[0211]

[0212] In the formula, c SI It represents the incentive price per unit volume of CO2 absorbed by the carbon sink device.

[0213] In step 2, the dispatching cost under the carbon source / sink evaluation index is introduced into the integrated energy system dispatching framework, specifically:

[0214] 1) Establish an optimization objective function in the gas energy flow network to minimize the natural gas purchase cost, gas turbine, P2G and gas tank operation cost:

[0215]

[0216] Where f1 is the main operating cost of the natural gas network, T is the total number of optimal scheduling periods, c GD is the cost of purchasing unit volume of natural gas from the external gas grid. is the gas operation and maintenance cost of the gas turbine, c P2G The cost of supplying a unit volume of natural gas to P2G, c GS The operation and maintenance costs of filling and degassing the gas tank, is the natural gas output power of the external natural gas grid during period t.

[0217] Ignoring the natural gas loss inside the equipment, the natural gas supply obtained by the system from the external gas grid, gas storage tank and P2G during period t is equal to the natural gas demand V provided to the gas turbine C and the gas load. t PLG Satisfying the balance:

[0218] ρ G V t GD +ρ G V t GS +ρ G (B1V t P2G -b2V t P2G )=ρ G V t DGT +ρ G V t PLG

[0219] Converted into the balance constraint of natural gas input and output power:

[0220]

[0221] In the formula, The natural gas output power for the gas load.

[0222] In the objective function of gas energy flow network optimization, the decision variables are Transfer variables from gas energy flow to electric-thermal energy flow network Obtained through energy conversion in the energy coupling device gas turbine.

[0223] Operating constraints of each device:

[0224]

[0225] In the formula, G i,t , They are the natural gas operating power of each device i in the integrated energy system during period t and its minimum and maximum values.

[0226] Gas energy flow pipeline valve pressure difference constraints:

[0227]

[0228] In the formula, are the minimum and maximum pressure differences of valves of each device i in the gas energy flow pipeline, is the natural gas flow rate of each device i in period t, μ is the flow coefficient (determined by the valve), is the pipe area.

[0229] Gas tank capacity constraints:

[0230]

[0231] In the formula, are the minimum and maximum values ​​of gas energy respectively.

[0232] 2) In the electric-thermal energy flow network, the carbon source equipment gas turbine outputs electric power, which leads to an increase in carbon emissions, and the carbon capture and P2G carbon sink equipment consumes electric power to absorb carbon. In order to promote low-carbon economic scheduling, it is proposed to take into account the carbon emission adjustment cost of carbon source / sink equipment sr and carbon sink adjustment costs si The objective function of the electric-thermal energy flow network optimization is:

[0233]

[0234] Where f2 represents the operating cost of the electricity-heat energy flow network, f e Refers to the electricity dispatch and operation costs of photovoltaic, wind power, gas turbines, batteries, carbon capture and P2G, f h Refers to the heat energy dispatching and operating costs of electric boilers and heat storage tanks, f sr is the carbon emission adjustment fee, f si is the carbon sink adjustment fee. PV 、c WT 、c DGT 、c CCS 、c P2G are the unit power generation costs of photovoltaic, wind power, gas turbine, carbon capture and P2G in period t, The operating cost of battery charging and discharging, are the unit heat storage and release operating costs of the electric boiler and the heat storage tank respectively. t PV , P t WT are the photovoltaic and wind power generation powers in period t respectively.

[0235] In the objective function of the electric and thermal energy flow network optimization, the decision variables in the t period are The transfer variable is β e (t) = {V t DGT ,V t CCS ,V t P2G}, obtained through the energy conversion of energy coupling equipment gas turbine and P2G, according to carbon emissions, it can be obtained through the conversion of energy coupling source equipment in the network, and the gas turbine power generation power P t DGT The required V t DGT , the power consumption of carbon capture device P t CCS Volume of CO2 captured P2G device power consumption P t P2G The volume of natural gas output is V t P2G .

[0236] Ignoring the internal power loss of the equipment, the system power balance relationship during period t is:

[0237] P t PV +P t WT +P t DGT +a1P t ES -a2P t ES =P t PLE +P t CCS +P t P2G +P t EHB

[0238] Where P t PLE is the power load value.

[0239] Ignoring the internal heat loss of the equipment, the heat supply and heat load demand of the system gas turbine, electric boiler and heat storage tank in period t are balanced:

[0240]

[0241] In the formula, is the heat load demand value.

[0242] Limit constraints for each equipment operation constraint:

[0243]

[0244] Where P i,t ,H i,t are the electrical power and thermal power of different equipment i in period t during the optimization process, are the minimum and maximum operating powers of these devices during period t. At the same time, the operation of these devices meets the constraints.

[0245] Thermal energy flow pipeline valve pressure difference constraints:

[0246]

[0247] In the formula, are the minimum and maximum pressure differences of the valves of each device i in the heat flow pipeline, respectively, Q is the density of hot water, is the hot water flow of each device i in period t.

[0248] Thermal energy flow pipeline temperature constraints:

[0249]

[0250] Where, T i min ,T i max are the minimum and maximum values ​​of the pipe temperature of each device i in the heat flow pipeline, respectively. i,t is the thermal power of each device i in time period t, C is the specific heat capacity of water, and T0 is the initial temperature of water.

[0251] Battery and heat storage tank capacity constraints:

[0252]

[0253] In the formula, They are the minimum and maximum energy storage of the battery and heat storage tank respectively.

[0254] like Figure 3 As shown, step 3 is to solve the system scheduling problem in a hierarchical and coordinated manner. The specific steps are as follows:

[0255] 1) Optimize the scheduling of lower-level subsystems

[0256] According to the initial forecast values ​​of the day-ahead dispatch of wind power generation, photovoltaic power generation, energy storage equipment and various energy loads in the integrated energy system, the augmented objective function of the gas energy flow network is established after adopting the Lagrange relaxation strategy for the network constraints:

[0257]

[0258] Among them, λ v is the Lagrange multiplier vector of the gas energy flow network, σ v is its penalty factor, R v1 (x v ), R vi (xv )) are the equality constraints and inequality constraints of the gas energy flow network, i=2,···,m.

[0259] Establish the augmented objective function of the electrical-thermal energy flow network:

[0260]

[0261] Among them, λ e is the Lagrange multiplier vector of the electrical-thermal energy flow network, σ e is its penalty factor, R e1 (x e ), R e2 (x e ), R ei (x e ) are the equality constraints and inequality constraints of the electric energy flow network and the thermal energy flow network respectively, j=3,…,l.

[0262] After solving the optimization problems of the lower gas energy flow network and the electric-thermal energy flow network, the optimization solution is recorded as

[0263] 2) Optimize the overall problem of the upper layer

[0264] Feedback the optimization results of each network to the upper layer for overall Lagrange multiplier optimization correction, and establish the dual optimization problem of the upper system:

[0265]

[0266] When solving the upper-level dual optimization problem, the gradient method is used to update the unknown quantity λ v ,λ e :

[0267]

[0268] 3) Check convergence conditions

[0269] Solve the lower-level subproblem and the upper-level overall dual problem, and optimize the system by alternating and iterating the upper and lower-level calculation processes until a global optimal solution is found.

[0270] Set the termination condition to:

[0271]

[0272] If φ k ≤ε, ε>0 is the control error, then the iteration is stopped. Due to the existence of the duality gap, the obtained solution is a feasible solution for the global optimization of the original problem.

[0273] In step 4, deep learning is applied to predict the output range of some equipment. Specifically, when applying deep learning, the initial conversion variables between equipment, the current equipment operating status, and the operating status of related equipment are set as data layer A, and the final value of the conversion variable between equipment is set as data layer B. After training and calibration inspection with measured data, it is applied. And the centralized training and decentralized execution strategy of deep reinforcement learning is used, specifically:

[0274] First, take 80% of the measured data A and B layers and give them to the neural network for learning until the training converges. Secondly, take 20% of the measured data A layer and input them into the neural network in turn, and the network outputs the B layer, and then compares it with the measured data B layer to calculate the difference. If the difference rate between the output of each measured device and the output predicted by deep learning is set to be less than the fixed value interval, the network learning effect is considered good. If the error cannot meet the requirements, more data samples need to be collected for learning. Finally, use the trained network directly, input the known A layer data into the network, and the neural network will give the B layer data.

[0275] The DeepMind team combined Actor-Critic (AC) and deep Q-learning algorithms to develop the DDPG algorithm. Based on the DDPG, the system's partial equipment is divided into two categories according to the carbon source / sink properties. Each category has its own A network. For each A network, the input dimension is greatly reduced. The two types of equipment share a C network. Under the condition that the individual and the whole meet the constraints, the target expectation is maximized as much as possible. The improved DDPG relies on the AC structure and uses the policy gradient algorithm. In the algorithm, the Actor no longer outputs the probability of each action, but gives a specific action, which is more helpful for the machine to learn in the continuous action space during training. Take advantage of the action in the continuous space and change the random strategy to a deterministic strategy.

[0276] Among them, the neural network parameters are set as follows: the input layer of the Actor network is an N×2 sequence, the pytorch framework is used to build the module, the hidden layer includes 2 layers, the number of neurons is 2560 and 320 respectively, the output layer dimension is N-1, the activation function is tanh, the loss function is mean square error, and the optimization method is RMSprop. The input of the Critic network consists of two parts: the first part is the state observed by the agent; the second part is the action taken by the agent accordingly. The hidden layer includes 2 layers, the number of neurons is 2560 and 320 respectively, and the number of neurons in the output layer is 1, which represents the Q value obtained by the Critic network when taking a certain action under this state. Except for the output layer, the other activation functions use tanh, and the activation function of the output layer is relu. After obtaining the Q value, the agent adopts the ε-greedy strategy, that is, there is a probability of 1-ε to choose π *(s) = argmaxQ(s,a), and the probability of generating random actions is ε, and the action value of generating random actions follows the normal distribution, μ = π * (s t |θ)=a,σ=[Q(s,a)-argmaxQ(s,a)] 2 .

[0277] During the training process, the observation values ​​of the devices in the system are combined as the input of the Critic part through AC, so that the Critic module of each agent can use the observation information and action information of all agents to evaluate the current S. The Actor selects an a based on the current S, and the Critic calculates a Q value based on Sa as feedback to the Actor's action. The Critic is trained based on the estimated Q value and the actual Q value, and the Actor updates the strategy based on the Critic's feedback.

[0278] The following is a simulation of a specific example. Some of the results are as follows: Figure 4 , Figure 5 shown.

[0279] Three scenarios were set up: S1 (no restrictions on carbon emissions), S2 (considering a simple penalty on total carbon emissions) and S3 (taking carbon sources / sinks into account to determine the dispatch weight factor). The dispatch schemes under the three considerations were obtained by particle swarm algorithm. The operating costs of the S1, S2 and S3 systems were 10698.27 yuan, 12007.45 yuan and 11839.49 yuan respectively. Although the system operating cost in S3 has increased, the environmental protection benefit is the best. In this scenario, the carbon source / sink index is used to evaluate the structure of carbon sources and the proportion of carbon sinks in the system during the current period. When photovoltaic power generation is high at noon, the carbon source control index constraint is reduced, the overall power of controllable equipment as demand response is increased, and the carbon sink detection index is constrained to increase the proportion of carbon capture device power. When the basic power load increases in the evening, the constraints of the two indicators are opposite to those at noon.

[0280] The method proposed in the present invention establishes a mathematical model that takes into account the multi-energy coupling characteristics of the equipment in the comprehensive energy system and its role in the carbon flow, proposes a systematic evaluation index on carbon sources / sinks based on the mathematical model of the equipment, calculates the adjustment cost through relevant incentive factors, introduces it into the objective function of the system's low-carbon operation scheduling model, and uses deep reinforcement learning to centrally train the model and propose a feasible strategy for decentralized scheduling. Avoiding excessive impact of carbon reduction targets on other targets is conducive to flexible scheduling of equipment, fully reflecting the different roles of equipment under different targets, and has great development prospects and application significance in reducing system carbon emissions and optimizing system carbon structure while taking into account the consumption of new energy within the system.

[0281] The above is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been described above with a preferred embodiment, it is not used to limit the present invention. Any technician familiar with the profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A method for optimizing the scheduling of an integrated energy system taking into account the role of carbon sources and sinks, characterized in that: The following steps are involved: Step 1: Aiming at the multi-energy flow coupling conversion characteristics of natural gas flow-electricity flow and electricity flow-heat flow in the integrated energy system, and establishing the correlation between the carbon source / sink of each device in the carbon evolution process of the system, introduce biomass energy carbon capture and storage technology, and establish a mathematical model; specifically: Diversion operation model of gas turbine under the action of carbon source: The total thermal power of the gas turbine is expressed as: Where t is the time period, is the total thermal power consumed by the gas turbine during period t, ρ G is the density of natural gas, L G is the lower calorific value of natural gas, is the natural gas consumption volume during period t, Δt is the time step, is the overall thermal efficiency of the gas turbine, is the power generation of the gas turbine during period t, is the thermal-to-electricity conversion efficiency of the gas turbine, is the heating power of the gas turbine; There is a coupling relationship between the power generation and heating power of the gas turbine. When the power is combined with heat, the ratio of the two powers is the power-to-heat ratio, and its physical model is: In the formula, They are the heating power, electric-to-heat ratio and power generation of the gas turbine during period t; the CO2 emission of the gas turbine and the power generation satisfy: In the formula, is the conversion ratio of gas turbine exhaust gas to natural gas consumption during period t, V t CDGT is the volume of CO2-containing gas produced by the gas turbine during period t, V t DGTC is the volume of CO2 gas emitted by the gas turbine during period t, is the CO2 concentration in the gas produced during period t, η r is the gas turbine gas cycle ratio, is the amount of CO2 produced by the gas turbine during period t, is the density of CO2; If the CO2 concentration in the combustion chamber of the gas turbine is less than 40%, the CO2 concentration of the gas produced by the gas turbine is: Carbon capture device CO2 absorption model: The carbon capture device consumes electrical power P during the carbon capture process t CCS Electric power P from carbon capture t cCCS and the fixed electrical power P t fCCS The sum of: Where P t CCS is the power consumed by the carbon capture device during period t, V t O is the volume of exhaust gas captured by the carbon capture device during period t, is the power consumption coefficient affected by the capture amount and conversion process, the device fixed power P t fCS Set to 0.1 times P t cCCS ; Based on the leaf CO2 absorption process of the ecosystem carbon sink, the low-energy CO2 absorption capacity of the carbon capture device in period t is proposed. is the difference between the amount of CO2 in the captured gas and the amount not absorbed, expressed as: In the formula, They are the CO2 non-absorption efficiency coefficient under the device absorption solution concentration, gas pressure in the absorption tower, and temperature limits; Biomass Carbon Capture and Storage: The plant CO2 assimilation rate shown by the biochemical process is: Where V t a is the actual plant CO2 assimilation rate, V t 0 is the rate of plant CO2 absorption, g0(T) and g1(T) are the sensitivity coefficients of stomatal temperature, g0(T) is the stomatal conductance when the assimilation rate of plant leaves is 0, Expressed as the air humidity around the leaf, P t a , P t c represent the partial pressures of air and CO2 in the leaf, respectively; CO2 is distributed, accumulated, and circulated within plants through assimilation, and the final amount of carbon stored is expressed as: In the formula, represents the amount of carbon sequestered by plants, R i Represents the CO2 produced by plant growth and catabolism; CO2 reuse model of P2G device: The volume of natural gas produced by P2G and the volume of CO2 recycled meet the following requirements: Where V t P2G is the volume of natural gas produced by P2G during period t, is the CO2 conversion ratio of P2G, V t CR represents the volume of CO2 reused by P2G during period t, T represents the total number of scheduling periods, when P2G does not obtain CO2 from the market or other sources, and the amount of CO2 reused by P2G is not greater than the amount absorbed by the carbon capture device; The P2G device consumes electrical power to convert CO2 back into methane, which is then supplied to the natural gas grid or gas storage tank for use. The natural gas output power of P2G during period t With natural gas volume V t P2G satisfy: Then the power consumption of the P2G device is P t P2G It is expressed as: In the formula, is the P2G power-to-gas conversion efficiency during period t; Energy storage tank model: The dynamic characteristics of the gas tank's charging and discharging are: In the formula, is the gas energy stored in the gas tank during period t, b1 and b2 are 0-1 variables, b1=1 or b2=1 indicates that the gas tank is in the state of inflation or deflation, is the natural gas storage or output power of the gas tank during period t; The heat loss of the heat storage tank is related to the stored heat, and the dynamic characteristics of thermal energy are expressed as: In the formula, is the heat energy stored in the heat storage tank during period t, ε is the heat loss rate of the heat storage tank, d1 and d2 are 0-1 variables, d1=1 or d2=1 indicates that the heat storage tank is in a state of absorbing or releasing heat, is the thermal power of the heat storage tank during period t; There is power loss in the battery during the process of charging, discharging and storing electric energy. Its dynamic characteristics of charging and discharging are expressed as: In the formula, is the electric energy stored in the battery during period t, a1 and a2 are 0-1 variables, a1=1 or a2=1 indicates that the battery is in the charging or discharging state, P t ES is the charging and discharging power of the battery during period t, are the charging and discharging efficiencies of the battery, and δ is the self-discharge rate of the battery; Step 2: Based on the correlation between the carbon source / sink effect of each device in the whole system and the carbon evolution process of the system, an evaluation index of the carbon source / sink effect is proposed, and the corresponding incentive factor is determined based on the index and the scheduling cost is calculated; and Considering the correlation of equipment variables under multi-energy flow coupling, a scheduling framework that includes all system entities is proposed, and the scheduling costs determined by carbon sources / sinks are introduced into the objective function, and a low-carbon operation model of the system is established according to the constraints; Step 3: Solve the system model using the hierarchical coordination method; Step 4: Based on the mathematical model of the equipment, deep learning is used to characterize the equipment at the data layer, and the equipment output range is predicted according to the correlation constraints between the equipment. The centralized training and decentralized execution strategy of deep reinforcement learning is used to implement online scheduling of each subject in the system in each cycle in a training-first-then-application manner.

2. The method for optimizing the scheduling of an integrated energy system taking into account the carbon source and sink effect according to claim 1, characterized in that: Step 2 The specific steps are as follows: Step 2.1: Propose carbon source control indicators and carbon sink monitoring indicators Among them, P t DGT , P t PV , P t WT The power generated by gas turbine, photovoltaic and wind power respectively; P t P2G , P t CCS , P t ES , P t EHB The electrical power consumed by the P2G device, carbon capture device, battery, and electric boiler are respectively; Step 2.2, calculate the dispatch cost; The dispatch cost established by the indicator is: In the formula, is the carbon emission adjustment cost of the carbon source equipment in period t, f si,t is the carbon sink adjustment cost of the carbon capture device during period t, c SR represents the unit cost of carbon emissions from the carbon source equipment, c SI It represents the incentive price per unit volume of CO2 absorbed by the carbon sink device; Step 2.3, introduce the dispatching cost calculated in step 2.2 into the objective function, and establish a low-carbon operation model of the system according to the constraints, specifically: Gas energy flow network optimization objective function: Where f1 is the main operating cost of the natural gas network, T is the total number of optimal scheduling periods, c GD is the cost of purchasing unit volume of natural gas from the external gas grid. is the gas operation and maintenance cost of the gas turbine, c P2G The cost of supplying a unit volume of natural gas to P2G, c GS The operation and maintenance costs of filling and degassing the gas tank, is the natural gas output power of the external natural gas grid during period t; Ignoring the natural gas loss inside the equipment, the natural gas supply obtained by the system from the external gas grid, gas storage tank and P2G during period t is equal to the natural gas demand V provided to the gas turbine C and the gas load. t PLG Satisfying the balance: Converted into the balance constraint of natural gas input and output power: In the formula, The natural gas output power for the gas load; Operating constraints of each device: In the formula, G i,t , are the natural gas operating power of each device i in the integrated energy system during period t and its minimum and maximum values; Gas energy flow pipeline valve pressure difference constraints: In the formula, are the minimum and maximum pressure differences of valves of each device i in the gas energy flow pipeline, is the natural gas flow rate of each device i in time period t, μ is the flow coefficient, which is determined by the valve, is the pipe area; Gas tank capacity constraints: In the formula, are the minimum and maximum values ​​of gas energy respectively; Electric-thermal energy flow network optimization objective function: Where f2 represents the operating cost of the electricity-heat energy flow network, f e Refers to the electricity dispatch and operation costs of photovoltaic, wind power, gas turbines, batteries, carbon capture and P2G, f h Refers to the heat energy dispatching and operating costs of electric boilers and heat storage tanks, f sr is the carbon emission adjustment fee, f si is the carbon sink adjustment fee, c PV 、c WT 、c DGT 、c CCS 、c P2G are the unit power generation costs of photovoltaic, wind power, gas turbine, carbon capture and P2G in period t, The operating cost of battery charging and discharging, are the unit heat storage and release operating costs of the electric boiler and the heat storage tank, respectively, t PV , P t WT are the power generation of photovoltaic power and wind power in period t respectively; Ignoring the internal power loss of the equipment, the system power balance relationship during period t is: P t PV +P t WT +P t DGT +a1P t ES -a2P t ES =P t PLE +P t CCS +P t P2G +P t EHB Where P t PLE is the power load value; Ignoring the internal heat loss of the equipment, the heat supply and heat load demand of the system gas turbine, electric boiler and heat storage tank in period t are balanced: In the formula, is the heat load demand value; Limit constraints for each equipment operation constraint: Where P i,t ,H i,t are the electrical power and thermal power of different equipment i in period t during the optimization process, are the minimum and maximum values ​​of the operating power of these devices in period t, respectively. At the same time, the operation of these devices meets the constraints; Thermal energy flow pipeline valve pressure difference constraints: In the formula, are the minimum and maximum pressure differences of the valves of each device i in the heat flow pipeline, respectively, Q is the density of hot water, is the hot water flow rate of each device i in time period t; Thermal energy flow pipeline temperature constraints: Where, T i min ,T i max are the minimum and maximum values ​​of the pipe temperature of each device i in the heat flow pipeline, respectively. i,t is the thermal power of each device i in time period t, C is the specific heat capacity of water, and T0 is the initial temperature of water; Battery and heat storage tank capacity constraints: In the formula, They are the minimum and maximum energy storage of the battery and heat storage tank respectively.

3. The method for optimizing the scheduling of an integrated energy system taking into account the carbon source and sink effect according to claim 1, characterized in that: Step 3 is as follows: Step 3.1, optimize the scheduling problem of the lower-level subsystem; According to the initial forecast values ​​of the day-ahead dispatch of wind power generation, photovoltaic power generation, energy storage equipment and various energy loads in the integrated energy system, the augmented objective function of the gas energy flow network is established after adopting the Lagrange relaxation strategy for the network constraints: Among them, λ v is the Lagrange multiplier vector of the gas energy flow network, σ v is its penalty factor, R v1 (x v ), R vi (x v )) are the equality constraints and inequality constraints of the gas energy flow network, i=2,···,m; Establish the augmented objective function of the electrical-thermal energy flow network: Among them, λ e is the Lagrange multiplier vector of the electrical-thermal energy flow network, σ e is its penalty factor, R e1 (x e ), R e2 (x e ), R ei (x e ) are the equality constraints and inequality constraints of the electric energy flow network and the thermal energy flow network, respectively, j = 3, ..., l; after solving the optimization problems of the lower gas energy flow network and the electric-thermal energy flow network, the optimization solution is recorded as Step 3.2, optimize the upper overall problem: Feedback the optimization results of each network to the upper layer for overall Lagrange multiplier optimization correction, and establish the dual optimization problem of the upper system: When solving the upper-level dual optimization problem, the gradient method is used to update the unknown quantity λ v ,λ e : Step 3.3, check the convergence conditions: Solve the lower-level subproblem and the upper-level overall dual problem, and optimize the system by alternating the upper and lower-level calculation processes until a global optimal solution is found; Set the termination condition to: If φ k ≤ε, ε>0 is the control error, then the iteration is stopped. Due to the existence of the duality gap, the obtained solution is a feasible solution for the global optimization of the original problem.

4. The method for optimizing the scheduling of an integrated energy system taking into account the carbon source and sink effect according to claim 1, characterized in that: In step 3, when deep learning is applied, the initial conversion variables between devices, the current device operating status, and the operating status of related devices are set as data layer A, and the final value of the conversion variable between devices is set as data layer B. After training and calibration inspection with measured data, it is applied. 80% of the measured data A and B layers are given to the neural network for learning. After the training converges, 20% of the measured data A layer is input into the neural network in turn, and the network outputs the B layer, which is then compared with the measured data B layer to calculate the difference. The difference rate between the output of each measured device and the output predicted by deep learning is set to be less than the fixed value interval, which means that the network learning effect is good. If the error cannot meet the requirements, more data samples need to be collected for learning. The trained network is directly used to input the known A layer data into the network, and the neural network gives the B layer data.

Images