CN-121998343-A - Park comprehensive energy system scheduling method considering central air conditioner cluster energy consumption

CN121998343ACN 121998343 ACN121998343 ACN 121998343ACN-121998343-A

Abstract

The invention discloses a park comprehensive energy system scheduling method considering central air-conditioning cluster energy consumption, belonging to the technical field of energy management and optimal control, wherein the method establishes a park comprehensive energy system; generating a plurality of time sequence prediction samples in a future scheduling period, carrying out cluster analysis on each time sequence prediction sample, selecting a plurality of typical day show scene data, constructing a corresponding day-ahead scheduling optimization target based on each typical day show scene data and a comfort level deviation index, solving through a day-ahead double-layer nested optimization model based on the day-ahead scheduling optimization target to obtain a scheduling result corresponding to each typical day show scene, determining a next day reference scheduling plan, and carrying out dynamic correction by adopting a day rolling optimization and game coordination model by taking the next day reference scheduling plan as an initial plan in the operation process of the integrated energy system of a next day park. The method solves the problems of insufficient time scale coordination, limited flexible utilization of the central air conditioner and lower operation robustness of the conventional park comprehensive energy system.

Inventors

MA TIANYI
LI TING
ZHOU KUN
SHI WENQING
LIU YINGLI

Assignees

北京印刷学院

Dates

Publication Date: 20260508
Application Date: 20260127

Claims (10)

1. A park comprehensive energy system scheduling method considering central air-conditioning cluster energy consumption is characterized by comprising the following steps: establishing a park comprehensive energy system comprising a power supply side, cold and heat source equipment, an energy storage unit and a user side air conditioner cluster; constructing a comfort level deviation index of the user side air conditioner cluster; Selecting different historical load data, historical renewable energy output data and historical meteorological data of a park comprehensive energy system, and calling a trained load and renewable energy output prediction model based on deep learning to generate a plurality of time sequence prediction samples in a future scheduling period, wherein the time sequence prediction samples comprise time sequence data of electric load, cold load and renewable energy output; The method comprises the steps of carrying out cluster analysis on each time sequence prediction sample to obtain a plurality of scene clusters, selecting the time sequence prediction sample closest to the center of the scene cluster in each scene cluster as typical day show scene data of the corresponding scene cluster; based on the data of each typical day scene and the comfort level deviation index, respectively constructing day-ahead scheduling optimization targets corresponding to each typical day scene; Solving through a day-ahead double-layer nested optimization model based on a corresponding day-ahead scheduling optimization target to obtain scheduling results corresponding to each typical day show scene; The method comprises the steps of calling a load based on deep learning and a renewable energy output prediction model which are completed by training based on latest historical data to generate a next day prediction time sequence, measuring the distance between the next day prediction time sequence and a time sequence prediction sample corresponding to each typical day show scene, and selecting a typical day scene with the smallest distance as a target scene; In the operation process of the comprehensive energy system of the next-day park, taking a next-day reference scheduling plan as an initial plan, and adopting a daily rolling optimization and game coordination model for dynamic correction; The daily double-layer nested optimization model and the daily rolling optimization and game coordination model both take an improved bloodsucking leech swarm algorithm as a solver.
2. The method for dispatching the park comprehensive energy system considering the energy consumption of the central air conditioner cluster according to claim 1, wherein the power supply side is used for supplying power and comprises a photovoltaic power generation unit, a wind power generation unit, a gas boiler, a cogeneration unit and a public power grid; the cold and heat source equipment is used for realizing the mutual conversion among electric energy, heat energy and cold energy and comprises an electric refrigerator and a lithium bromide absorption refrigerator; The energy storage unit is used for realizing time migration and peak clipping and valley filling of energy and comprises an electric energy storage system, a hot energy storage system and a cold energy storage system.
3. The method for scheduling a campus integrated energy system taking into account central air conditioning cluster energy consumption according to claim 1, wherein the operation constraint conditions of the campus integrated energy system include energy balance constraint, energy equipment operation constraint and energy storage unit dynamic constraint: Wherein, the Is a photovoltaic power generation unit Generating power at moment; Is a wind power generation unit Generating power at moment; for cogeneration units Generating power at moment; Is a public power grid Purchase power at moment; Comprehensive energy system for park The power is required by the basic electric load at the moment; Is an electric refrigerator Input power at time; For electric energy storage systems Charging power at a moment; For electric energy storage systems Discharge power at a moment; for cogeneration units Heating capacity at moment; Is a gas boiler Heating capacity at moment; For heat energy storage systems The heat release amount at the moment; Comprehensive energy system for park Thermal load demand at time; For heat energy storage systems The charging amount at the moment; is lithium bromide absorption refrigerator Driving heat input quantity at moment; Is an electric refrigerator Refrigerating output at moment; is lithium bromide absorption type refrigerator Refrigerating output at moment; is a cold energy storage system Cooling capacity at moment; Comprehensive energy system for park The time of day cold load demand; is a cold energy storage system The amount of cold accumulation at the moment; Minimum input power allowed for the electric refrigerator; Is an electric refrigerator Input power at time; Maximum input power allowed for the electric refrigerator; Minimum refrigeration output allowed for lithium bromide absorption chiller; is an absorption type cooler Refrigerating output at moment; at the moment for lithium bromide absorption chiller Is provided for the maximum allowable refrigeration output; Maximum power generated by the photovoltaic power generation unit; Maximum power generated by the wind power generation unit; The maximum power of the cogeneration unit; the heat efficiency of the cogeneration unit; The maximum heating capacity of the gas boiler is provided; The heat efficiency of the gas boiler is; Is a gas boiler Input power at time; Is an electric refrigerator The refrigeration energy efficiency ratio at the moment; the performance coefficient of the lithium bromide absorption chiller; the refrigerating output jump quantity of the lithium bromide absorption refrigerator; is lithium bromide absorption type refrigerator -Cooling output at time 1; For electric energy storage systems An energy state at time +1; For electric energy storage systems The energy state at the moment; The energy charging efficiency of the electric energy storage system is improved; is the time step; The energy release efficiency of the electric energy storage system is achieved; For heat energy storage systems An energy state at time +1; For heat energy storage systems The energy state at the moment; the energy charging efficiency of the thermal energy storage system; the energy release efficiency of the thermal energy storage system; is a cold energy storage system An energy state at time +1; is a cold energy storage system The energy state at the moment; The energy charging efficiency of the cold energy storage system is improved; The energy release efficiency of the cold energy storage system is achieved; Is the lower energy state limit of the electrical energy storage system; An upper energy state limit for the electrical energy storage system; an upper limit for charging power of the electrical energy storage system; an upper discharge power limit for the electrical energy storage system; is the lower energy state limit of the thermal energy storage system; an upper energy state limit for the thermal energy storage system; An upper limit of the charge amount for the thermal energy storage system; an upper limit of the heat release amount of the heat energy storage system; is the lower energy state limit of the cold energy storage system; An upper energy state limit for the cold energy storage system; the upper limit of the cold charge of the cold energy storage system is set; And the upper limit of the cold energy storage system is filled and discharged.
4. The method for dispatching a park comprehensive energy system considering energy consumption of a central air-conditioning cluster according to claim 1, wherein the expression of the comfort level deviation index of the user side air-conditioning cluster is: Wherein, the Air conditioning cluster for user side A time comfort level deviation index; Time is; The number of areas scheduled for participating in the user side air conditioner cluster; indexing the region numbers; Is the first Individual areas The actual temperature at the moment; Is the first Individual areas Setting a temperature at the moment; Is an absolute value; is a comfortable lower temperature limit; is the upper comfort temperature limit.
5. The method for dispatching a park comprehensive energy system considering energy consumption of a central air conditioner cluster according to claim 1, wherein the expression of the day-ahead dispatching optimization target is: Wherein, the Optimizing objectives for day-ahead scheduling, in particular objective function values Minimum; Is the electricity purchasing cost coefficient; Is a public power grid Purchase power at moment; Is a fuel cost factor; Is that Fuel consumption at the moment; Maintaining a cost coefficient for the operation of the equipment output; Is that The equipment output at moment; Cost coefficient per carbon emission; Is that Equivalent carbon emissions at time; the weight coefficient is a comfort level deviation index; Is that A time comfort level deviation index; is the total number of time steps in the scheduling period.
6. The method for dispatching the park comprehensive energy system considering the energy consumption of the central air conditioner cluster according to claim 1, wherein the upper layer optimization of the day-ahead double-layer nested optimization model aims at source load stability, minimizes fluctuation of the power purchasing power of the public power grid in a dispatching period, and aims at minimizing an objective function value The lower layer of the double-layer nested optimization model before the day to comprehensively optimize the target And the minimization is aimed, and the cooperative adjustment of response distribution and comfort deviation constraint of each sub-cluster is realized by coordinating each air-conditioning sub-cluster obtained by dividing the air-conditioning clusters at the user side through non-cooperative game: Wherein, the Optimizing for dual layer nesting before date an upper layer optimization target of the model; for the total number of time steps in the scheduling period; Is a public power grid Purchase power at moment; The average value of the power purchase power in the dispatching cycle is obtained; Is an absolute value; Optimizing model for double-layer nesting before date is a lower layer comprehensive optimization target; Is that Weight coefficient of (2); An economic cost target value calculated from the electricity purchase cost and the gas cost; Is that Weight coefficient of (2); equivalent carbon emission generated by electricity purchase and gas consumption in a dispatching cycle; Is that Weight coefficient of (2); Is composed of Accumulating the comfort level deviation indexes at the moment to obtain a comfort level deviation target value; Is the electricity purchasing cost coefficient; Is a fuel cost factor; Is that Fuel consumption at the moment; Maintaining a cost coefficient for the operation of the equipment output; Is that The equipment output at moment; Cost coefficient per carbon emission; Is that Equivalent carbon emissions at time; Is that And a comfort level deviation index at the moment.
7. The method for dispatching the park comprehensive energy system considering the energy consumption of the central air conditioner cluster according to claim 1, wherein the method for dynamically correcting by adopting a daily rolling optimization and game coordination model is specifically as follows: And triggering an intra-day rolling optimization and game coordination model to dynamically correct the equipment output, the electricity purchasing power and the stored energy charging and discharging power of the remaining time steps in the day when the deviation between the actual load or the renewable energy output and the next day reference scheduling plan exceeds a preset threshold or the energy storage unit SOC exceeds a preset boundary threshold.
8. The method for dispatching the park comprehensive energy system considering the energy consumption of the central air conditioner cluster according to claim 7, wherein the upper layer laminated game participants of the intra-day rolling optimization and game coordination model comprise the total power of power grid interaction, energy storage and air conditioning, and the common optimization targets of the upper layer laminated game participants are as follows: Wherein, the To optimize the goal for the upper layer, in particular to minimize the goal value ; To be from the current time step By the end time step of the scheduling period Is a target value of economic cost; to be from the current time step By the end time step of the scheduling period A carbon emission cost target value of (2); the lower layer of the daily rolling optimization and game coordination model independently optimizes each air conditioner sub-cluster under the power constraint through non-cooperative games, and the lower layer optimization targets of the daily rolling optimization and game coordination model are as follows: Wherein, the Is a carbon emission factor term; Is the first Equivalent carbon emission of each air conditioner sub-cluster in the current period; is a comfort level deviation term coefficient; Is the first Comfort level deviation values of the air conditioner sub-clusters in the current period; punishing a weight factor for air conditioning sub-group comfort bias; Is the first Constraint penalty factors for each air conditioner sub-cluster in the current iteration.
9. The method for dispatching the park comprehensive energy system considering the energy consumption of the central air conditioner cluster according to claim 1 is characterized in that the improved optimization algorithm of the blood sucking leech cluster is characterized in that a constraint penalty factor self-adaption mechanism, a self-adaption step length adjustment mechanism, a proportion self-adjustment mechanism and an elite solution reservation mechanism are introduced into the optimization algorithm of the blood sucking leech cluster, and the elite solution reservation mechanism is used for reserving a high-adaptability individual preferentially based on a set reservation proportion when target fluctuation or individual feasibility is reduced.
10. The method for dispatching the park comprehensive energy system considering the energy consumption of the central air conditioner cluster according to claim 9, wherein expressions of the constraint penalty factor self-adaption mechanism, the self-adaption step length adjustment mechanism and the proportion self-adaption mechanism are respectively as follows: Wherein, the Is the first Constraint adjustment coefficients for +1 iterations; Is the first Constraint adjustment coefficients of the second iteration; is a nonlinear adjusting function; Is the first Constraint violation degree indexes in the secondary iteration; Is the first The iteration is used for controlling the step length adjustment quantity of the search update amplitude; Is the first +1 Iterations are used to control the step size adjustment of the search update amplitude; clipping a function for the power correction boundary; A step size lower limit for adjusting the minimum amplitude limit; an upper step size limit for adjusting the maximum amplitude limit; to adjust parameters; Is a natural constant; Standard deviation of the current population; Is the first A preset variance threshold in the secondary iteration; First, the A step attenuation coefficient of the secondary iteration; judging a threshold value for convergence; is a scaling factor.

Description

Park comprehensive energy system scheduling method considering central air conditioner cluster energy consumption Technical Field The invention belongs to the technical field of energy management and optimal control, and particularly relates to a park comprehensive energy system scheduling method considering central air conditioner cluster energy consumption. Background With the deep implementation of the 'double carbon' target, the energy system is accelerated to the evolution towards the multi-energy complementation and intelligent direction. The traditional single energy system has the defects of single energy form, slow response and low operation efficiency, and can not meet the requirements of the park-level comprehensive energy system on high efficiency, cleanliness and flexible scheduling. The park Comprehensive energy system (compact INTEGRATED ENERGY SYSTEM, CIES) realizes layered supply and cascade utilization of energy through integrating multiple energy sources such as electricity, heat, cold and the like, and becomes an important form of intelligent energy development in the future. However, the CIES operation process has the characteristics of multi-time scale coupling, various load types, high operation uncertainty and the like, so that the scheduling and the optimization of the CIES operation process are very challenging. Investigation shows that the energy consumption of the central air-conditioning cluster is generally more than 40% of the total energy consumption of the park, and the operation characteristics of the central air-conditioning cluster have decisive influence on the peak-valley distribution, the energy efficiency level and the operation economy of the system. The central air conditioner has remarkable flexible regulation potential, and peak clipping, valley filling and energy efficiency improvement can be effectively realized by reasonably setting temperature, power and start-stop strategies. However, the following major problems remain in the current research and practice: (1) The time scale is single, and the coordination between the day front and the day inside is lacking. The existing scheduling mostly adopts a staged independent strategy, the day-ahead optimization focuses on long-term economy, the real-time scheduling focuses on short-term fluctuation response, the two lack of cooperation, and the global and local targets are difficult to be considered. Especially when the central air conditioning cluster is considered, the delay characteristic of the central air conditioning cluster cannot be utilized to realize cross-time scale adjustment, and the flexibility of the system is limited. (2) The central air conditioner is not fully utilized in flexibility. In the existing model, an air conditioner is mostly regarded as a rigid load, only power tracking or simple reduction control is carried out, and adjustability in a comfort level allowable range is ignored, so that the air conditioner is in a passive response state in a system, and both energy utilization efficiency and user experience are affected. (3) The system optimization and robustness are insufficient. The comprehensive energy system relates to multi-objective (economical efficiency, low carbon and comfort) and multi-constraint characteristics, the traditional algorithm is slow to converge and easy to sink into local optimum under a high-dimensional situation, and the renewable energy output fluctuation and load prediction error further aggravate scheduling risks. Therefore, a collaborative scheduling method which can achieve both long-term planning and short-term response, fully mine the flexible potential of the central air conditioner and has strong robustness optimization capability is needed, so that efficient, low-carbon and stable operation of the park comprehensive energy system is realized. Disclosure of Invention Aiming at the defects in the prior art, the scheduling method of the park comprehensive energy system considering the energy consumption of the central air conditioner cluster solves the problems of insufficient time scale coordination, limited flexible utilization of the central air conditioner and lower operation robustness of the conventional park comprehensive energy system. In order to achieve the above purpose, the technical scheme adopted by the invention is a park comprehensive energy system scheduling method considering central air-conditioning cluster energy consumption, comprising the following steps: establishing a park comprehensive energy system comprising a power supply side, cold and heat source equipment, an energy storage unit and a user side air conditioner cluster; constructing a comfort level deviation index of the user side air conditioner cluster; Selecting different historical load data, historical renewable energy output data and historical meteorological data of a park comprehensive energy system, and calling a trained load and renewable energy output prediction model