An adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system

Abstract

The present invention relates to an adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system, comprising: constructing a microgrid system energy supply model; performing adaptive rolling time domain prediction within a scheduling period T to obtain a predicted load within a prediction time domain; the length of the prediction time domain is variable at different control times; inputting the predicted load into the microgrid system energy supply model, with the goal of minimizing the system comprehensive operating cost C, and obtaining the output results of each device in the microgrid system through a particle swarm optimization algorithm. The beneficial effects of the present invention are: the prediction time domain length of the present invention can be adjusted according to load fluctuations, reducing the situation where the microgrid system cannot optimize scheduling in time to meet large load fluctuations, thereby reducing the system operating cost and improving scheduling efficiency.

Description

Technical Field

[0001] The invention relates to the technical field of microgrid systems, and more specifically, to an adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system. Background Art

[0002] The CCHP microgrid system has the advantages of high primary energy utilization rate and low environmental pollution, and has become one of the hot topics in recent years. The CCHP microgrid system can generate cold energy, heat energy, heat energy and heat energy at the same time, and the primary energy utilization rate can reach 75%-80%, and the energy consumed is only 3 / 4 of the traditional heat and power supply form. A typical CCHP microgrid system includes (1) generator sets, such as gas turbines, internal combustion engines, fuel cells, etc.; (2) heating equipment, such as waste heat boilers, gas boilers, etc.; (3) refrigeration equipment, such as lithium bromide absorption refrigerators, electric refrigerators, etc.

[0003] Affected by outdoor factors such as climate change and indoor factors such as indoor personnel and equipment, the load has great uncertainty and volatility. There may be a large difference between the load data predicted a day ago and the actual load data. The cooling, heating and power supply equipment only outputs according to the hourly scheduling plan of the microgrid system for the next 24 hours obtained by the day-ahead optimization. It cannot effectively match the fluctuating demand during the day, which will lead to a decrease in user satisfaction. If the prediction deviation of the grid-balanced load is directly used, the operating economy of the system will be reduced. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide an adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system.

[0005] First, a method for predicting medium- and long-term monthly runoff of a reservoir using an improved temporal convolutional network is provided, including:

[0006] S1. Construct the energy supply model of microgrid system;

[0007] S2. Perform adaptive rolling time domain prediction within the scheduling period T to obtain the predicted load within the prediction time domain; the length of the prediction time domain is variable at different control times;

[0008] S3. Input the predicted load into the energy supply model of the microgrid system, and obtain the output results of each device in the microgrid system through a particle swarm optimization algorithm with the goal of minimizing the system comprehensive operating cost C.

[0009] Preferably, in S2, performing adaptive rolling time domain prediction to obtain the predicted load of the prediction time domain includes:

[0010] S201, setting the first control time τ 0The length L of the first prediction time domain 0 , the starting time of the first prediction time domain is the first control time τ 0 , the terminal time of the first prediction time domain is L 0 +τ 0 , and calculate the first load variance s corresponding to the first prediction time domain 0 2 ;

[0011] S202, setting the control time domain i, the second control time τ 1 =τ 0 +i, the starting time of the second prediction time domain is the second control time τ 1 , the terminal time of the first prediction time domain is L 1 +τ 1 , and calculate L 1 =L 0 The second load variance s 1 2 ;

[0012] S203: Calculate the second load variance s 1 2 and the first loading variance s 0 2 If the absolute value of the difference is less than the set variance change threshold Δs min 2 , then maintain L 1 =L 0 ; If the absolute value of the difference is greater than the variance change threshold Δs min 2 , and the variance of the second loading s 1 2 is smaller than the first load variance s 0 2 When L 1 =L 0 -1; if the absolute value of the difference is greater than the variance change threshold Δs min 2 , and the variance of the second loading s 1 2 is greater than the first load variance s 0 2 When L 1 =L 0 +1;.

[0013] S204, according to the load variance, adjust different control times τ in sequence n =τ 0 +ni, the length of the prediction time domain, up to τ n ≥T.

[0014] Preferably, in S2, the calculation formula of the load variance is:

[0015]

[0016] Wherein, L is the current prediction time domain length, in hours h; and are the average values ​​of electric load, cooling load and heating load in the current forecast time domain, in kilowatts kW; P τ 、F τ and H τ They are the electric load, cooling load and heating load corresponding to the τth moment in the current prediction time domain, in kilowatts kW; The control time is τ n The load variance in the prediction time domain; τ n It is the starting time of the current prediction time domain.

[0017] Preferably, in S3, the calculation formula of the system comprehensive operating cost C is:

[0018]

[0019] C grid (τ) = P grid (τ)·R grid (τ)·△τ

[0020]

[0021] Among them, C gas (τ), C grid (τ) and C om (τ) are natural gas cost, electricity purchase cost and operation and maintenance cost respectively;

[0022] Natural gas cost C gas In the calculation formula of (τ), P gt (τ) is the electric power released by the combustion of natural gas in the prime mover at time τ, in kilowatts kW; η gt_ele (τ) is the power generation efficiency of the prime mover, HV is the calorific value of natural gas; R gas is the price of natural gas; △τ is the system operation time interval, unit: hour;

[0023] Electricity purchase cost C grid In the calculation formula of (τ), P grid (τ) is the electric power purchased from the power grid at time τ, R grid (τ) is the electricity purchase price at time τ;

[0024] Operation and maintenance cost C om In the calculation formula of (τ), Fac (τ), F ec (τ), H wb (τ), H gb (τ), P bt_dis (τ), P bt_chr (τ), Q wt_dis (τ) and Q wt_chr (τ) are respectively the cooling power of the absorption chiller, the cooling power of the electric chiller, the power of the waste heat boiler, the power of the gas boiler, the battery discharge power, the battery charge power, the water reservoir energy release power and the water reservoir energy storage power, in kilowatts (kW); R om_gt , R om_ac , R om_ec , R om_bt , R om_wt , R om_wb and R om_gb They are the operating and maintenance prices of gas internal combustion engines, absorption refrigerators, electric refrigerators, batteries, water tanks, waste heat boilers and gas boilers.

[0025] Preferably, in S3, the optimization factors of the particle swarm optimization algorithm include grid power, prime mover power, absorption refrigerator power, electric refrigerator power, gas boiler power, waste heat boiler power and energy storage device charging and discharging power.

[0026] In a second aspect, a combined cooling, heating and power microgrid system as described in the first aspect is provided, comprising:

[0027] Power grid, prime mover, waste heat boiler, absorption refrigerator, gas boiler, electric refrigerator and energy storage system; wherein the electric energy generated by the prime mover is transmitted to the power supply grid for users through electric wires, and the power grid is used to supplement the power supply of the prime mover when it is insufficient; the heat energy generated by the prime mover is transmitted to the waste heat boiler and the absorption refrigerator respectively through the first pipeline; the heat generated by the waste heat boiler is transmitted to the heating pipeline for heating users; the cooling energy generated by the absorption refrigerator is transmitted to the cooling pipeline for cooling users; part of the heat energy generated by the gas boiler is transmitted to the absorption refrigerator through the second pipeline, and the rest is directly transmitted to the heating pipeline; the power supply pipeline is connected to an electric refrigerator, and the cooling energy generated by the electric refrigerator is transmitted to the cooling pipeline through the third pipeline.

[0028] In a third aspect, a computer storage medium is provided, and when the computer program is executed on a computer, the computer executes an adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system as described in any one of the first aspects.

[0029] In a fourth aspect, a computer program product is provided. When the computer program product is run on a computer, the computer is enabled to execute an adaptive rolling time domain optimization method for a combined heat, cooling and power microgrid system as described in any one of the first aspects.

[0030] The beneficial effects of the present invention are:

[0031] (1) Compared with the rolling time domain optimization method with a fixed time domain in the prior art, the present invention proposes a new method for determining the length of the prediction time domain. The length of the prediction time domain of the present invention can be adjusted according to the load fluctuation, and then the load information under the corresponding prediction time domain is obtained, and the particle swarm algorithm is used for optimization and solution, and the scheduling plan is determined according to the optimization results.

[0032] (2) Compared with the traditional short prediction time domain, the present invention reduces the situation where the microgrid system cannot optimize scheduling in time to meet the large fluctuations in load, thereby reducing the system operation cost.

[0033] (3) Compared with a longer prediction time domain, the present invention shortens the system optimization calculation time and improves the intra-day scheduling efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flowchart of an adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system;

[0035] Figure 2 It is a structural schematic diagram of a combined cooling, heating and power microgrid system;

[0036] Figure 3 It is a schematic diagram of the intraday adaptive time domain rolling control model;

[0037] Figure 4 It is a schematic diagram of the operation flow of the intraday adaptive time domain rolling control model;

[0038] Figure 5 It is a schematic diagram of equipment output results using the adaptive rolling time domain optimization method and the fixed rolling time domain optimization method;

[0039] Figure 6 It is a schematic diagram of the economic index results of the microgrid system after being optimized by the conventional rolling time domain optimization method and the optimization method of the present invention. DETAILED DESCRIPTION

[0040] The present invention is further described below in conjunction with embodiments. The description of the following embodiments is only used to help understand the present invention. It should be noted that for ordinary persons in the art, without departing from the principle of the present invention, the present invention can also be modified in some ways, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

[0041] Embodiment 1:

[0042] A combined cooling, heating and power microgrid system, such as Figure 2 As shown, including:

[0043] Power grid, prime mover (PM), waste heat boiler (WB), absorption chiller (AC), gas boiler (GB), electric chiller (EC) and energy storage system; the energy storage system includes battery (BT) and water tank (WT).

[0044] It should be noted that the prime mover can be a gas turbine (GT) or a gas internal combustion engine (ICE). Figure 2 Taking a gas internal combustion engine as an example, the prime mover generates heat or electricity by consuming natural gas. The electricity generated by the prime mover is transmitted through wires to an external power grid for powering users, providing electrical load for the system. The system can also meet the system's electricity demand by purchasing electricity from the power grid, and when the system's power production capacity is in surplus, the remaining electricity can be incorporated into the power grid. The heat energy generated by the prime mover is transmitted to the waste heat boiler and the absorption chiller through the first pipeline.

[0045] This application does not limit the connection method between the gas internal combustion engine, the waste heat boiler and the absorption chiller. In an optional implementation, Figure 2 As shown, the gas internal combustion engine, the waste heat boiler and the absorption refrigeration machine are connected in sequence through the first pipeline. In another optional implementation, after the gas internal combustion engine is connected to the waste heat boiler, it can also be directly connected to the absorption refrigeration machine without passing through the waste heat boiler.

[0046] The waste heat boiler generates heat by using hot flue gas to provide heat load for the system, and the generated heat is transported to the heating pipeline for heating users. It should be noted that the heating pipeline usually uses hot water or steam, and the function of the waste heat boiler is to heat water with high-temperature flue gas. Therefore, the present invention uses the waste heat boiler to heat users instead of the heat energy generated by the prime mover.

[0047] In addition, the absorption chiller generates cold energy by utilizing heat energy such as hot flue gas or jacket water to provide cooling load, and the generated cold energy is transported to the cooling pipeline for providing cooling for users. Therefore, in the microgrid system of the present invention, since "cold" is driven by the absorption chiller to provide heat energy through the prime mover, gas boiler or electric chiller, "cold" also refers to "heat" in a broad sense.

[0048] The gas boiler consumes natural gas to meet the heat demand of the system and provides the heat energy required for the cooling load of the absorption chiller. That is to say, part of the heat energy generated by the gas boiler is transported to the absorption chiller through the second pipeline, and the other part is directly transported to the heating pipeline. The electric chiller is connected to the power grid, and the cold energy generated by the electric chiller is transported to the cooling pipeline through the third pipeline, providing cooling load for the system and meeting part of the heat demand of the system.

[0049] In actual application, the system can also add solar power generation devices and wind power generation devices connected to the power grid as needed to provide power load for the system and meet part of the system's power demand, thereby reducing the amount of electricity purchased by the system from the power grid and the cost of purchasing electricity.

[0050] Embodiment 2:

[0051] Based on the microgrid system in Example 1, it is necessary to introduce a shorter time scale control in the microgrid system scheduling model. Based on the start and stop status and output plan of the output units obtained in the day-ahead scheduling stage, a rolling time domain control method is used in the intraday scheduling stage to optimize and obtain a more precise unit scheduling plan with smaller output deviation. Specifically, the present invention establishes the following Figure 3 The intra-day adaptive time domain rolling control model shown in the present invention is not a conventional fixed time domain rolling control, but adopts a more reasonable adaptive prediction time domain length to obtain a more economical microgrid system optimization scheduling result.

[0052] The present invention provides a method for predicting medium- and long-term monthly runoff of a reservoir by improving a temporal convolutional network. Figure 1 As shown, including:

[0053] S1. Construct a mathematical model of each device in the microgrid system, and construct an energy supply model of the microgrid system based on the mathematical model of each device.

[0054] Specifically, the mathematical models of each device include:

[0055] 1. Prime mover (gas internal combustion engine) model

[0056] The natural gas consumption of internal combustion engines is expressed as:

[0057]

[0058] Where G is the amount of natural gas consumed by the internal combustion engine, unit: m 3 ; P is the power generation of the internal combustion engine, unit kW; η e is the electrical efficiency of the internal combustion engine; CV is the calorific value of natural gas, in kWh / m 3 .

[0059] 2. As the main heat generating equipment in the CCHP system, the heat generating model of the waste heat boiler is:

[0060] H out,boiler =H in,boiler ×η boiler

[0061] Among them, H out,boiler is the heat output of the waste heat boiler, in kW; H in,boiler is the heat entering the waste heat boiler, in kW; η boiler is the thermal efficiency of the waste heat boiler.

[0062] 3. The absorption chiller is the main cooling equipment in the trigeneration system. The flue gas with higher temperature enters the chiller for double-effect refrigeration cycle, and the jacket water with lower temperature enters the chiller for single-effect refrigeration cycle. The mathematical model of the chiller can be expressed as follows:

[0063] C ac =H in,ac ×COP ac

[0064] In the formula, C out,ac is the cooling capacity of the absorption chiller, in kW; H in,ac is the flue gas heat entering the refrigerator, in kW; COP ac is the refrigeration efficiency of the absorption chiller.

[0065] 4. Gas boilers are used as supplementary heating equipment in the trigeneration system to supplement when the heat production and heat storage of waste heat boilers are insufficient. The model is as follows:

[0066] H out,gas_boiler =H in,gas_boiler ×η gas_boiler

[0067] In the formula, H out,gas_boiler is the heat output of the gas boiler, in kW; H in,gas-boiler is the heat entering the gas boiler, unit kW; η gas_boiler is the thermal efficiency of gas boiler.

[0068] 5. The electric refrigerator is used as a supplementary cooling device for the system. It provides cooling when the refrigerator and cold storage are insufficient. The model is as follows:

[0069] C hp =E in,hp ×COP hp

[0070] In the formula, C hp E is the cooling capacity of the electric refrigerator, in kW; in,hp The power entering the electric refrigerator, in kW; COP hp is the cooling efficiency of the electric refrigerator.

[0071] Based on the mathematical models of all equipment, the energy supply model of the microgrid system under the target operation mode is constructed.

[0072] S2. Within the scheduling period T, an adaptive rolling time domain prediction is performed to obtain the predicted load within the prediction time domain; the length of the prediction time domain is variable at different control times.

[0073] The calculation formula for the load variance required to adjust the prediction time domain length is:

[0074]

[0075] Where, L is the current prediction time domain length, in hours; and are the average values ​​of electric load, cooling load and heating load in the current forecast time domain, in kilowatts kW; P τ 、F τ and H τ They are the electric load, cooling load and heating load corresponding to the τth moment in the current prediction time domain, in kilowatts kW; The control time is τ n The load variance in the prediction time domain; τ n It is the starting time of the current prediction time domain.

[0076] S2 specifically includes:

[0077] S201, setting the first control time τ 0 The length L of the first prediction time domain 0 , the starting time of the first prediction time domain is the first control time τ 0 , the terminal time of the first prediction time domain is L 0 +τ 0 , and calculate the first load variance s corresponding to the first prediction time domain 0 2 ;

[0078] S202, setting the control time domain i, the second control time τ 1 =τ 0 +i, the starting time of the second prediction time domain is the second control time τ 1 , the terminal time of the first prediction time domain is L 1 +τ 1 , and calculate L 1 =L 0 The second load variance s 1 2 ;

[0079] S203, calculating the second load variance s 1 2 and the first loading variance s0 2 If the absolute value of the difference is less than the set variance change threshold Δs min 2 , it means that the load fluctuation is not significant, and the L 1 =L 0 ; If the absolute value of the difference is greater than the variance change threshold Δs min 2 , and in the second load variance s 1 2 Less than the first load variance s 0 2 When L 1 =L 0 -1; if the absolute value of the difference is greater than the variance change threshold Δs min 2 , and in the second load variance s 1 2 Greater than the first load variance s 0 2 When L 1 =L 0 +1;.

[0080] S204, according to the load variance, adjust different control times τ in sequence n =τ 0 +ni, the length of the prediction time domain, up to τ n ≥T.

[0081] For example, please refer to Figure 3 , set the first control time τ 0 is 0h, the length of the first prediction time domain is L 0 is 3h; the control time domain i is 1h, the first prediction time domain is rolled back 1h to obtain the second prediction time domain, and the length of the second prediction time domain is temporarily set to 3h; due to the second load variance s 1 2 and the first loading variance s 0 2 The difference is greater than the set variance change threshold Δs min 2 , and the load variance s in the second prediction time domain 1 2 Greater than the first forecast time domain load variance s 0 2 , increase the length of the second prediction time domain by one hour.

[0082] Similarly, please refer to Figure 3 , the second prediction time domain is rolled back 1h to obtain the third prediction time domain, and the length of the third prediction time domain is temporarily set to 4h; due to the third load variance s 2 2and the second loading variance s 1 2 The difference is greater than the set variance change threshold Δs min 2 , and the load variance s in the third prediction time domain 2 2 Less than the second forecast time domain load variance s 1 2 , reduce the length of the third prediction time domain by one hour.

[0083] By analogy, at the control time τ n When the load variance of the corresponding forecast time domain is equal to the previous control time τ n-1 The load variance of the forecast time domain under the current time domain is compared to determine the length of the forecast time domain at that moment. Figure 4 As shown, when the control time domain is 1 hour, the control moment and the prediction time domain are continuously rolled back by 1 hour until the terminal moment of the prediction time domain exceeds the end moment of the given scheduling period T (ie, L+τ≥T). In addition, when τ≥T, it means that the rolling prediction of all control moments within the scheduling period T has been completed. For example, the scheduling period T is generally set to 24 hours, that is, one day. In addition, the initial length of the prediction time domain can be any value from 3h to 9h.

[0084] S3. Input the predicted load into the microgrid system energy supply model, with the goal of minimizing the system comprehensive operating cost C, and obtain the output results of each device in the microgrid system through the particle swarm optimization algorithm.

[0085] In S3, the calculation formula of the system comprehensive operating cost C is:

[0086]

[0087] C grid (τ) = P grid (τ)·R grid (τ)·△τ

[0088]

[0089] Among them, C gas (τ), C grid (τ) and C om (τ) are respectively natural gas cost, electricity purchase cost and operation and maintenance cost, in yuan;

[0090] Natural gas cost C gas In the calculation formula of (τ), P gt (τ) is the electric power released by the combustion of natural gas in the prime mover at time τ, in kilowatts kW; η gt_ele(τ) is the power generation efficiency of the prime mover, HV is the calorific value of natural gas 38931 kJ / m3; R gas is the price of natural gas, in RMB / m3; △τ is the system operation time interval, in hours;

[0091] Electricity purchase cost C grid In the calculation formula of (τ), P grid (τ) is the electric power purchased from the power grid at time τ, R grid (τ) is the electricity purchase price at time τ;

[0092] Operation and maintenance cost C om In the calculation formula of (τ), F ac (τ), F ec (τ), H wb (τ), H gb (τ), P bt_dis (τ), P bt_chr (τ), Q wt_dis (τ) and Q wt_chr (τ) are respectively the cooling power of the absorption chiller, the cooling power of the electric chiller, the power of the waste heat boiler, the power of the gas boiler, the battery discharge power, the battery charge power, the water reservoir energy release power and the water reservoir energy storage power, in kilowatts (kW); R om_gt , R om_ac , R om_ec , R om_bt , R om_wt , R om_wb and R om_gb They are the operating and maintenance prices of gas internal combustion engines, absorption chillers, electric chillers, batteries, water reservoirs, waste heat boilers and gas boilers, in yuan / kWh.

[0093] In S3, the optimization factors of the particle swarm optimization algorithm include grid power, prime mover power, absorption chiller power, electric chiller power, gas boiler power, waste heat boiler power, and energy storage device charging and discharging power.

[0094] In summary, if the microgrid system is optimized by the adaptive rolling time domain optimization method, the load variance s at the control time τ and its next time is first calculated. 0 2 and 1 2 Compare the difference between the two to see if it exceeds the range Δs defined by the variance change min 2 If it does not exceed, it means that the load fluctuation is not significant. At this time, the prediction time domain length L remains unchanged. If the predicted load variance Less than At this time, the predicted time domain length L 1 In the previous prediction time domain length L0 Based on the time domain length of 1 hour, that is, L 1 =L 0 -1; if the forecast load variance Greater than Prediction length L 1 In the previous prediction time domain length L 0 On the basis of the time domain length, the time domain length is increased by 1 hour, that is, L 1 =L 0 +1; the system heat load is met by the prime mover, waste heat boiler, gas boiler, absorption chiller and electric chiller, etc. The total natural gas consumption of the system is the natural gas consumption of the prime mover and gas boiler, and the economic index C of the combined heat, cooling and power system is obtained.

[0095] Embodiment 3:

[0096] The method in Example 2 is used to optimize the output power of the main equipment of the microgrid system, and the main equipment operating power corresponding to the lowest comprehensive operating cost of the system can be obtained as follows: Figure 5 As shown in the figure, in addition to the optimization scheduling results of the intraday adaptive time domain, there are also the corresponding optimization scheduling results of the fixed 3-hour time domain.

[0097] Experiment 1): The time domain duration in the algorithm is fixed to 3 hours, and the equipment power of the CCHP system is optimized. The results are as follows: Figure 5 shown.

[0098] Experiment 2): The time domain duration in the algorithm is fixed to 7 hours, and the equipment power of the CCHP system is optimized. The results are as follows: Figure 5 shown.

[0099] Figure 6 The total operating cost obtained in experiments 1)-2) and in the embodiments of the present invention is shown. Figure 6 It can be seen that compared with the traditional short prediction time domain, the present invention reduces the situation where the microgrid system cannot optimize scheduling in time to meet the large fluctuations in load, thereby reducing the system operation cost; compared with the long prediction time domain, the present invention reduces the increase in system calculation time and the waste of resources, thereby improving the model efficiency.

[0100] In summary, the prediction time domain length of the present invention can be adjusted according to the load fluctuation situation, and then the load information under the corresponding prediction time domain is obtained, and the particle swarm algorithm is used for optimization and solution, and the scheduling plan is determined according to the optimization results; compared with the traditional short prediction time domain, the situation where the microgrid system cannot optimize the scheduling in time to meet the large fluctuations in the load is reduced, thereby reducing the system operation cost; compared with the longer prediction time domain, the system optimization calculation time is shortened, and the daily scheduling efficiency is improved.

Claims

1. An adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system. It is characterized in that The following steps are involved: S1. Construct the energy supply model of microgrid system; S2. Perform adaptive rolling time domain prediction within the scheduling period T to obtain the predicted load within the prediction time domain; the length of the prediction time domain is variable at different control times; In S2, the adaptive rolling time domain prediction is performed to obtain the predicted load in the prediction time domain, including: S201, setting the first control time τ 0 The length L of the first prediction time domain 0 , the starting time of the first prediction time domain is the first control time τ 0 , the terminal time of the first prediction time domain is L 0 +τ 0 , and calculate the first load variance s corresponding to the first prediction time domain 0 2 ; S202, setting the control time domain i, the second control time τ 1 =τ 0 +i, the starting time of the second prediction time domain is the second control time τ 1 , the terminal time of the first prediction time domain is L 1 +τ 1 , and calculate L 1 =L 0 The second load variance s 1 2 ; S203: Calculate the second load variance s 1 2 and the first loading variance s 0 2 If the absolute value of the difference is less than the set variance change threshold Δs min 2 , then maintain L 1 =L 0 ; If the absolute value of the difference is greater than the variance change threshold Δs min 2 , and the variance of the second loading s 1 2 is smaller than the first load variance s 0 2 When L 1 =L 0 -1; if the absolute value of the difference is greater than the variance change threshold Δs min 2 , and the variance of the second loading s 1 2 is greater than the first load variance s 0 2 When L 1 =L 0 +1; S204, according to the load variance, adjust different control times τ in sequence n =τ 0 +ni, the length of the prediction time domain, up to τ n ≥T; In S2, the calculation formula for load variance is: Wherein, L is the current prediction time domain length, in hours h; and are the average values ​​of electric load, cooling load and heating load in the current forecast time domain, in kilowatts kW; P τ 、F τ and H τ They are the electric load, cooling load and heating load corresponding to the τth moment in the current prediction time domain, in kilowatts kW; The control time is τ n The load variance in the prediction time domain; τ n is the starting time of the current prediction time domain; S3. Input the predicted load into the energy supply model of the microgrid system, and obtain the output results of each device in the microgrid system through a particle swarm optimization algorithm with the goal of minimizing the system comprehensive operating cost C.

2. The adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system according to claim 1, It is characterized in that In S3, the calculation formula of the system comprehensive operating cost C is: C grid (τ)=P grid (t)·R grid (τ)·△τ Among them, C gas (τ), C grid (τ) and C om (τ) are natural gas cost, electricity purchase cost and operation and maintenance cost respectively; Natural gas cost C gas In the calculation formula of (τ), P gt (τ) is the electric power released by the combustion of natural gas in the prime mover at time τ, in kilowatts kW; η gt_ele (τ) is the power generation efficiency of the prime mover, HV is the calorific value of natural gas; R gas is the price of natural gas; △τ is the system operation time interval, unit: hour; Electricity purchase cost C grid In the calculation formula of (τ), P grid (τ) is the electric power purchased from the power grid at time τ, R grid (τ) is the electricity purchase price at time τ; Operation and maintenance cost C om In the calculation formula of (τ), F ac (τ), F ec (τ), H wb (τ), H gb (τ), P bt_dis (τ), P bt_chr (τ), Q wt_dis (τ) and Q wt_chr (τ) are respectively the cooling power of the absorption chiller, the cooling power of the electric chiller, the power of the waste heat boiler, the power of the gas boiler, the battery discharge power, the battery charge power, the water reservoir energy release power and the water reservoir energy storage power, in kilowatts (kW); R om_gt , R om_ac , R om_wb , R om_gb , R om_ec , R om_bt , R om_wt , R om_wb and R om_gb They are the operating and maintenance prices of gas internal combustion engines, absorption chillers, waste heat boilers, gas boilers, electric chillers, batteries, water reservoirs, waste heat boilers and gas boilers.

3. The adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system according to claim 2, It is characterized in that In S3, the optimization factors of the particle swarm optimization algorithm include power of the power grid, power of the prime mover, power of the absorption refrigerator, power of the electric refrigerator, power of the gas boiler, power of the waste heat boiler and power of charging and discharging of the energy storage device.

4. A combined cooling, heating and power microgrid system as claimed in claim 1, It is characterized in that include: A power grid, a prime mover, a waste heat boiler, an absorption refrigerator, a gas boiler, an electric refrigerator and an energy storage system; wherein the electric energy generated by the prime mover is transmitted to a power grid for supplying electricity to users through electric wires, and the power grid is used to supplement the power supply of the prime mover when the power supply is insufficient; the heat energy generated by the prime mover is transmitted to the waste heat boiler and the absorption refrigerator respectively through a first pipeline; the heat generated by the waste heat boiler is transmitted to a heating pipeline for supplying heat to users; The cooling energy generated by the absorption refrigerator is delivered to the cooling pipeline for providing cooling to users; part of the heat energy generated by the gas boiler is delivered to the absorption refrigerator through the second pipeline, and the rest is directly delivered to the heating pipeline; an electric refrigerator is connected to the power supply network, and the cooling energy generated by the electric refrigerator is delivered to the cooling pipeline through the third pipeline.

5. A computer storage medium, It is characterized in that The computer storage medium stores a computer program; when the computer program is executed on a computer, the computer executes the adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system as claimed in any one of claims 1 to 3.

6. A computer program product, It is characterized in that When the computer program product is run on a computer, the computer is enabled to execute the adaptive rolling time domain optimization method for a combined cooling, heating and power microgrid system as claimed in any one of claims 1 to 3.

Images