Real-time optimization scheduling method for source-grid-load-storage flexibility resources based on MPC
Through the robust optimization framework and model prediction control method, a real-time optimization scheduling model is built, which solves the real-time optimization scheduling problem of source network load storage flexible resources in an uncertain environment, and realizes the robustness and economicality of real-time power generation plans. to meet the real-time peak shaving needs of the power system.
Patent Information
- Application Number
- CN202210485393.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-06
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2042-05-06
AI Technical Summary
In an uncertain environment, the real-time optimization and scheduling of flexible source and load storage resources faces challenges. The existing technology is mainly focused on the recent optimization, making it difficult to effectively respond to the real-time peak shaving needs of the power system.
Using a robust optimization framework and model prediction control (MPC) method, a real-time optimization scheduling model is constructed, combined with the source and load ultra-short-term prediction information and real-time measurement information, the real-time power generation plan, electricity price strategy is calculated through the solution of the secondary planning model. Transmission line capacity increase strategy and energy storage charging and discharging strategy.
It realizes the feasibility and economicality of source and load interaction in an uncertain environment, and can roll back the real-time power generation plan, improves the robustness and economicality of the scheduling plan, and reduces the uncertainty of system operation.
Smart Images

Figure CN114865715B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of operation, optimization dispatching and control of power systems, and in particular to a real-time optimization dispatching method for source-grid-load-storage flexibility resources based on model predictive control (MPC). Background Art
[0002] In recent years, the installed capacity of intermittent renewable energy generation has continued to grow, and the proportion of renewable energy generation represented by wind power and photovoltaics has also continued to increase. However, wind power and photovoltaic power have uncertainties, and their large-scale grid connection has increased the difficulty of peak load regulation in the power system, and the demand for flexible resources is increasing.
[0003] In the context of energy internet, the coordinated operation of flexible resources of source, grid, load and storage will become the norm. Research on the coordinated and optimized operation of flexible resources of source, grid, load and storage is currently focused on the day-ahead optimization stage. The intraday ultra-short-term prediction of extended wind power has been widely used in the active dispatch and control of wind power. Given that the time resolution of the real-time plan of power system operation is generally greater than the sampling period of real-time measurement information, the combination of real-time dispatch and control means provides a good solution to reduce the uncertainty of system operation and improve the reliability of (online) dispatch plans. It is necessary to carry out relevant exploration and research on the real-time coordinated operation strategy of flexible resources of source, grid, load and storage. Summary of the invention
[0004] The present invention aims to solve the problem of real-time optimization scheduling of source-grid-load-storage flexibility resources in an uncertain environment. Based on a robust optimization framework and an MPC method, a real-time optimization scheduling method for source-grid-load-storage flexibility resources is provided that is adaptable to dynamic updating of source-load prediction information.
[0005] The present invention solves the above technical problems by adopting the following technical solutions: obtaining the operating parameters of conventional units on the power supply side, the topological structure and line parameters on the power grid side, the load scale of the load side demand response, the operating parameters of the battery on the energy storage side, the start and stop plan of the conventional units and the charging and discharging plan of the energy storage units in the rolling optimization scheduling stage, and the ultra-short-term prediction information on the source side and the load side; within the real-time optimization scheduling cycle, based on the robust optimization framework and the MPC method, with the minimum comprehensive cost of the source, grid, load and storage flexibility resources as the objective function, and the constraints taking into account the operating characteristics of the source, grid, load and storage flexibility resources, a real-time optimization scheduling model is constructed, the real-time optimization scheduling model is converted into a quadratic programming model, and a commercial solver is called to solve the above quadratic programming problem, and the real-time output plan of conventional thermal power units, the allowable fluctuation range of renewable energy power generation and load demand response, the real-time electricity price strategy, the real-time dynamic capacity expansion strategy of the transmission line, and the real-time charging and discharging strategy of energy storage are calculated.
[0006] The present invention has the following advantages: (1) The present invention is based on robust optimization and model predictive control methods. The real-time power generation plan formulated by the present invention comprehensively considers the feasibility and economy of source-load interaction under an uncertain environment, and can be used to rollingly correct the real-time power generation plan based on the ultra-short-term source-load forecast information and real-time measurement information that are continuously updated in the real-time scheduling stage; (2) The model constructed by the present invention is ultimately converted into a quadratic programming model, which can be directly solved by calling an efficient commercial solver. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the drawings of the embodiment are briefly introduced below.
[0008] Figure 1 An ultra-short-term forecast power diagram of wind power and load provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0009] A real-time optimization scheduling method for source-grid-load-storage flexibility resources based on MPC, comprising the following steps:
[0010] Step 1: Obtain the operating parameters of conventional units on the power supply side, the topological structure and line parameters on the grid side, the demand response load scale on the load side, the operating parameters of batteries on the energy storage side, the start and stop plan of conventional units and the charging and discharging plan of energy storage units in the rolling optimization scheduling stage, and the ultra-short-term prediction information on the source side and the load side;
[0011] Step 2: Construct the objective function and constraints of the real-time optimization scheduling model, and establish the real-time optimization scheduling model for source-grid-load-storage flexibility resources. Please refer to steps 2.1 to 2.2 for details.
[0012] Step 2.1, establish the objective function of the real-time optimization scheduling model of source-grid-load-storage flexibility resources: min{F 1s +F 2s +F 3s +F 4s +F 5s} (1)
[0013]
[0014]
[0015]
[0016]
[0017] F 5s =F ops -F pss (6)
[0018]
[0019]
[0020]
[0021] Where, T s is the period of real-time scheduling; N T is the number of thermal power units, N R is the number of renewable energy sites, N P is the number of nodes where the price-based demand response (PDR) load is located, N N is the number of transmission lines, N B is the number of energy storage units; mt is the current period; F 1s is the deviation cost of the output of adjustable units relative to the rolling optimization result in the real-time scheduling time domain, F 2s is the reserve cost of thermal power units in the real-time optimization stage, F 3s is the sum of the penalty cost for ultra-short-term renewable energy generation and the penalty cost for load demand response, F 4s is the dynamic capacity increase cost of the transmission line in the real-time scheduling stage, F 5s is the total operating cost of the energy storage power station, F ops is the operation and maintenance cost of the energy storage unit, F pss The frequency regulation benefit of the energy storage unit; and is the coefficient matrix of deviation cost, and are the adjustable unit output plan and state variable row vector in the intraday rolling optimization scheduling stage and the real-time scheduling stage, respectively, Δu j,mt+t is the row vector of control variables of the adjustable unit; and are the upward and downward rotation reserve capacities of thermal power unit j in time period t in the real-time optimization scheduling stage, and are the corresponding upward and downward spinning reserve cost coefficients respectively; and are the wind / solar power abandonment and load shedding power of renewable energy station r in period t, and are the penalty cost coefficients for the corresponding abandoned wind / solar power and load shedding power respectively; and are the upward and downward demand response power deviations of PDR load n in time period t, and are the penalty cost coefficients for the corresponding upward and downward demand response power deviations, respectively; is the over-limit power of line m in time period t, is the corresponding unit power dynamic capacity increase cost; are the rated power, rated capacity, unit power and unit capacity operating costs of energy storage unit b respectively; α bs , r are the actual cycle life, discount rate, and capital recovery rate of energy storage unit b, respectively; are the upward adjustment output of the energy storage unit b in the charging state in time period t, the downward adjustment output in the discharging state and its profit coefficient, They are respectively the downward adjustment output of the energy storage unit b in the charging state in time period t, the upward adjustment output in the discharging state and its profit coefficient.
[0022] Step 2.2, construct the constraints of the real-time optimization dispatch model of source-grid-load-storage flexibility resources, mainly including: conventional thermal power unit constraints, wind farm / photovoltaic power station power constraints, PDR load power constraints, energy storage power station operation constraints, system node power balance constraints, system spinning reserve constraints, network transmission capacity security constraints. Among them, the conventional thermal power unit constraints include technical output constraints, unit climbing constraints, and spinning reserve release constraints; the energy storage power station operation constraints include energy storage unit charging / discharging power constraints, energy storage unit charge state constraints, and energy storage power station life constraints.
[0023] The technical output constraints of the adjustable units are:
[0024]
[0025] Where P a,mt,0 It is a row vector composed of the real-time sampling values of the adjustable unit a in the mt period. and They are the row vectors consisting of the upper and lower limits of the technical output of adjustable unit a respectively.
[0026] The adjustable unit climbing constraint:
[0027] -R dn ≤Δu a,mt+τ ≤R up ,τ∈[1,t] (11)
[0028] In the formula, Δu a,mt+τ is the row vector of the adjustable unit control variables. up and R dn They are row vectors composed of the ramp-up / ramp-down rates of the adjustable units in the real-time scheduling stage.
[0029] The energy storage unit charge state constraint:
[0030]
[0031] In the formula, SoC b,tis the row vector of the charge state of the energy storage unit b in time period t, is the row vector formed by the maximum and minimum charge states of the energy storage unit b; are row vectors composed of real-time sampling values of the energy storage unit b in the charging and discharging states during the mt period; and are the charging and discharging efficiencies of energy storage station b, and They are the row vectors of the control variables of the energy storage unit b in the charging and discharging states during the mt+τ period.
[0032] The system power balance constraint:
[0033]
[0034] Where N L is the number of load nodes, p a,mt,0 is the real-time sampling value of adjustable unit a during the mt period, Δu a,mt+τ is the control variable of adjustable unit a in the period mt+τ; The ultra-short-term predicted output of renewable energy station r in time period t; is the predicted value of the load at node l in period t, is the predicted response power of PDR load n in time period t.
[0035] The network transmission capacity security constraint is:
[0036]
[0037] In the formula, λ a,m , n,m and λ r,m are the quasi-steady-state sensitivity factors of adjustable unit a, PDR load and renewable energy station r to line m respectively; is the response range of the PDR load at node n in the tth period; and is the bidirectional transmission capacity limit of line m; Δp a,t is the regulated power of adjustable unit a in period t, is the fluctuating power of renewable energy station r in period t; is the capacity increase power of line m in time period t; is the capacity increase limit of line m.
[0038] Step 3: Convert the network transmission capacity security constraint (14) into the following deterministic constraint:
[0039]
[0040] In the formula, λ ε,m is the set of quasi-steady-state sensitivity factors of renewable energy station r and PDR load to line m; are the upper and lower limits of the forecast error of the power that can be absorbed by the renewable energy station r in time period t, are the upper and lower limits of the permissible response power of the PDR at node n in time period t, respectively; is the set of the upper limit of the forecast error of the renewable energy station r in the period t and the lower limit of the allowable response power of the PDR load n in the period t; is the set of the lower limit of the forecast error of the renewable energy station r in the period t and the upper limit of the allowable response power of the PDR load n in the period t; Auxiliary variables introduced for real-time optimization scheduling stage.
[0041] Step 4: Call a commercial solver (such as CPLEX or GUROBI) to solve the above quadratic programming problem, and calculate the real-time output plan of conventional thermal power units, the allowable fluctuation range of renewable energy power generation and load demand response, the real-time electricity price strategy, the real-time dynamic capacity expansion strategy of transmission lines, the real-time charging and discharging strategy of energy storage, etc.
[0042] An embodiment is based on the IEEE 30-node test system, with thermal power units connected to nodes 1, 2, 22, 23, and 27, with installed capacities of 160MW, 160MW, 100MW, 60MW, and 110MW, respectively. The load demand response node is located at node 8, and the response quantity prediction error does not exceed ±4.75%. The maximum charge and discharge power of the energy storage battery is 180MW, and the capacity is 450MWh. The energy storage installation costs are 20,000 yuan / MW and 30,000 yuan / MWh, respectively. The penalty cost coefficient for abandoned wind power and the penalty cost coefficient for downward fluctuation deviation of load demand response are 400 yuan / MW, and the penalty cost coefficient for load shedding power, the penalty cost coefficient for upward fluctuation deviation of load demand response, and the dynamic capacity increase cost coefficient of transmission lines are 600 yuan / MW. The wind farm is connected to node 5, with an equivalent installed capacity of 300MW. The wind power control center updates the ultra-short-term forecast information of wind power for the next 15 minutes every 5 minutes. Taking the rolling updated wind power forecast data from 11:00 to 12:00 (recorded as time period 1 to 15) as an example, the real-time optimization scheduling result analysis is carried out.
[0043] The ultra-short-term forecast information of wind power and load within the optimization period is as follows: Figure 1 As shown, the following four optimization models are defined: Model 1: a deterministic optimization model based on MPC; Model 2: a robust optimization model based on MPC, i.e., the method of the present invention; Model 3: a deterministic optimization model; Model 4: a robust optimization model.
[0044] Table 1 Comparison of calculation results of different models
[0045]
[0046] As shown in Table 1, under the same wind power consumption capacity, Model 2 and Model 4 based on robust optimization have stronger ability to cope with the uncertainty of load demand response than Model 1 and Model 3 based on deterministic optimization. Moreover, without considering MPC, Model 3 and Model 4 require more operating costs of thermal power units and energy storage power stations, resulting in higher comprehensive costs than Model 1 and Model 2. Compared with Model 1 and Model 4, Model 2 has made an effective trade-off in dealing with source-load uncertainty and system operation economy, verifying the feasibility and effectiveness of the method of the present invention.
[0047] It can be seen from the above test results that the method of the present invention can effectively solve the problem of real-time power generation planning for source-grid-load-storage flexibility resources taking into account source-load uncertainty, which is conducive to achieving the robustness and economy of decision-making results.
Claims
1. A real-time optimization scheduling method for source-grid-load-storage flexibility resources based on MPC, characterized in that: include: Obtain the operating parameters of conventional units on the power supply side, the topology and line parameters on the grid side, the demand response load scale on the load side, the operating parameters of batteries on the energy storage side, the start and stop plans of conventional units and the charging and discharging plans of energy storage units in the rolling optimization scheduling stage, and ultra-short-term prediction information on the source side and the load side; In the real-time optimization scheduling cycle, based on the robust optimization framework and MPC method, with the minimum comprehensive cost of source-grid-load-storage flexibility resources as the objective function, and the constraints taking into account the operating characteristics of source-grid-load-storage flexibility resources, a real-time optimization scheduling model is constructed, and the real-time optimization scheduling model is converted into a quadratic programming model. The commercial solver is called to solve the quadratic programming model, and the real-time output plan of conventional thermal power units, the allowable fluctuation range of renewable energy power generation and load demand response, the real-time electricity price strategy, the real-time dynamic capacity expansion strategy of transmission lines, and the real-time charging and discharging strategy of energy storage are calculated; The objective function is described as: min{F 1s +F 2s +F 3s +F 4s +F 5s } F 5s =F ops -F pss Where, T s is the period of real-time scheduling; N T is the number of thermal power units, N R is the number of renewable energy sites, N P is the number of nodes where the PDR load is located, N N is the number of transmission lines, N B is the number of energy storage units; mt is the current period; F 1s is the deviation cost of the output of adjustable units relative to the rolling optimization result in the real-time scheduling time domain, F 2s is the reserve cost of thermal power units in the real-time optimization stage, F 3s is the sum of the penalty cost for ultra-short-term renewable energy generation and the penalty cost for load demand response, F 4s is the dynamic capacity increase cost of the transmission line in the real-time scheduling stage, F 5s is the total operating cost of the energy storage power station, F ops is the operation and maintenance cost of the energy storage unit, F pss The frequency regulation benefit of the energy storage unit; and is the coefficient matrix of deviation cost, and are the adjustable unit output plan and state variable row vector in the intraday rolling optimization scheduling stage and the real-time scheduling stage, respectively, Δu j,mt+t is the row vector of control variables of the adjustable unit; and are the upward and downward rotation reserve capacities of thermal power unit j in time period t in the real-time optimization scheduling stage, and are the corresponding upward and downward spinning reserve cost coefficients respectively; and are the wind / solar power abandonment and load shedding power of renewable energy station r in period t, and are the penalty cost coefficients for the corresponding abandoned wind / solar power and load shedding power respectively; and are the upward and downward demand response power deviations of PDR load n in time period t, and are the penalty cost coefficients for the corresponding upward and downward demand response power deviations, respectively; is the over-limit power of line m in time period t, is the corresponding unit power dynamic capacity increase cost; are the rated power, rated capacity, unit power and unit capacity operating costs of energy storage unit b respectively; r are the actual cycle life, discount rate, and capital recovery rate of energy storage unit b; are the upward adjustment output of the energy storage unit b in the charging state in time period t, the downward adjustment output in the discharging state and its profit coefficient, They are respectively the downward adjustment output of the energy storage unit b in the charging state in time period t, the upward adjustment output in the discharging state and its profit coefficient; The constraints include conventional thermal power unit constraints, wind farm / photovoltaic power station power constraints, PDR load power constraints, energy storage power station operation constraints, system node power balance constraints, system spinning reserve constraints, and network transmission capacity safety constraints. The conventional thermal power unit constraints include technical output constraints, unit climbing constraints, and spinning reserve release constraints. The energy storage power station operation constraints include energy storage unit charging / discharging power constraints, energy storage unit charge state constraints, and energy storage power station life constraints.
2. The method according to claim 1, characterized in that The network transmission capacity security constraint is transformed into the following deterministic constraint: In the formula, λ ε,m is the set of quasi-steady-state sensitivity factors of renewable energy station r and PDR load to line m; are the upper and lower limits of the forecast error of the power that can be absorbed by the renewable energy station r in time period t, are the upper and lower limits of the permissible response power of the PDR at node n in time period t, respectively; is the set of the upper limit of the forecast error of the renewable energy station r in the period t and the lower limit of the allowable response power of the PDR load n in the period t; is the set of the lower limit of the forecast error of the renewable energy station r in the period t and the upper limit of the allowable response power of the PDR load n in the period t; Auxiliary variables introduced for real-time optimization scheduling stage.
Citation Information
Patent Citations
Source network load storage flexibility resource day-ahead optimization method
CN113869608A
Multi-target day-ahead scheduling optimization method and system for source network load storage power system
CN114362255A
Power grid mixing and rolling scheduling method that considers clogging and energy-storing time-of-use price
WO2020143104A1
