A hierarchical energy management operation strategy considering hybrid demand response
By introducing a comprehensive objective function of user load layer and dynamic adaptive weight coefficient in a multi-micronet system, combined with a hybrid demand response strategy, the instability of the power system caused by the fluctuation of renewable energy output is solved, and the optimization of load management is achieved and reduce system operation costs.
Patent Information
- Application Number
- CN202210455249.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-28
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-04-28
AI Technical Summary
In multi-microgrid systems, the randomness and volatility of renewable energy output lead to unstable operation of the power system, and the existing technology is difficult to effectively manage load fluctuations, affecting the economy, reliability and timeliness of user electricity use.
A layered energy management operation strategy considering hybrid demand response is proposed. By introducing a user load layer at the multi-micronet system layer, a comprehensive objective function is constructed using dynamic adaptive weight coefficients, and a hybrid demand response strategy is adopted to reasonably transfer or Disconnect power loads and optimize renewable energy utilization and electricity trading.
It realizes the maximum utilization of renewable energy in multi-microgrid systems, reduces load transfer or interruption, reduces system operation costs, improves the economy and reliability of user electricity use, and improves the stability of system operation.
Smart Images

Figure CN114707896B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of hierarchical energy management, and in particular to a hierarchical energy management operation strategy considering hybrid demand response. Background Art
[0002] With the proposal of the "dual carbon" goal, my country's power industry will usher in a major opportunity for transformation and development. Renewable energy is connected to the active distribution network in the form of a multi-microgrid system, playing an important role in alleviating the strong dependence on fossil energy. However, the randomness and volatility of renewable energy output will have an adverse impact on the reliability of power system operation. Fully realizing the effective management of system energy in the active distribution network containing a multi-microgrid system is the key to ensuring the safe, stable and economic operation of the power system. In the multi-microgrid system, there are a large number of user loads inside each microgrid system, which is regarded as the electricity demand side. In demand-side management, the power generation side optimizes the configuration by using economic or technical methods, guides the demand-side users to adjust the electricity consumption mode and distribution demand, and reduces load fluctuations to achieve an economic and environmentally friendly energy management strategy. Demand response is a management strategy that is used more frequently in demand-side management. It is determined by the electricity users inside the multi-microgrid system based on time-of-use electricity price information and corresponding incentives. In hierarchical energy management, time-of-use electricity prices are often used for power transactions at the multi-microgrid system layer and the active distribution network layer. As the power market environment continues to change, considering demand response strategies in the operation of active distribution network systems containing multi-microgrid systems has become a popular direction and has been widely used. Demand response strategies guide internal users of microgrid systems to interrupt or transfer part of the load during peak hours, that is, to tap the potential for flexible adjustment of user-side loads, smooth the load fluctuations of microgrid systems, and further adjust the peak-to-valley differences of the system.
[0003] During the grid-connected operation of the multi-microgrid system, the output of renewable energy is easily affected by environmental factors, and there will be excess or insufficient output in the system. Therefore, for the hierarchical energy management model of the active distribution network containing the multi-microgrid system, it is necessary to optimize the operation strategy, reasonably transfer the power load during the period of insufficient output of renewable energy to the period of excess output, make full use of renewable energy and alleviate the power supply pressure during the peak period of electricity consumption. Summary of the invention
[0004] The purpose of the present invention is to overcome the shortcomings of the prior art, introduce a user load layer within each microgrid system, and propose a hierarchical energy management operation strategy that takes into account hybrid demand response, so that the energy management of the multi-microgrid system layer can achieve the maximum absorption of renewable energy while taking into account the participation of internal users in load demand response, thereby ensuring the economy, reliability and timeliness of users' electricity use to the greatest extent and reducing load transfer or interruption.
[0005] The present invention solves the technical problem by adopting the following technical solutions:
[0006] A hierarchical energy management operation strategy considering hybrid demand response includes the following steps:
[0007] Step 1: Set device parameters and constraints, and determine the data source;
[0008] Step 2: Select scenes from the data set;
[0009] Step 3: Divide the time periods and use the hybrid demand response strategy to analyze the electric load within the user load layer;
[0010] Step 4, solving the multi-microgrid system layer to obtain the optimal result;
[0011] Step 5: judge the optimal result. If an abnormal value appears, adjust the corresponding coefficient and return to step 4;
[0012] Step 6: If there is an internal transaction, the internal pricing strategy is executed, otherwise the lower layer result is passed to the active distribution network layer;
[0013] Step 7: The active distribution network energy decision center reads the information and feeds back the electricity sales situation to each microgrid energy management system;
[0014] Step 8: Determine the final result. If it meets expectations, save the relevant results. Otherwise, adjust the correlation coefficient and return to step 3.
[0015] Step 9, determine whether the stop condition is met, if it is met, output the final result, otherwise exit the operation after the iteration is met;
[0016] Step 1: Set the equipment parameters of the active distribution network including the multi-microgrid system and the upper and lower limits of the constraint conditions, and determine the data source of renewable energy output and load demand based on the relevant data sets;
[0017] Objective function:
[0018] In the multi-microgrid system layer, the user load layer is added, and the dynamic adaptive weight coefficient is used to construct the comprehensive objective function as shown in equations (1) and (2). The economic cost and environmental cost are shown in equations (3) and (6):
[0019]
[0020] Where F is the comprehensive objective function; T is the system operation cycle; g1(t) and g2(t) are the economic cost and environmental cost respectively; ω1 and ω2 are the dynamic adaptive weight coefficients;
[0021]
[0022] Where n1 and n2 are linear correlation coefficients; G1(t) and G2(t) are the per-unit values of economic cost g1(t) and environmental cost g2(t); g 1,max and g 2,max are the maximum values of the economic objective function and the environmental objective function in the entire operation cycle respectively;
[0023] min g1=C F +C M +C ADN +C MG +C LOAD +C P (3)
[0024] In the formula, C F is the natural gas consumption cost; C M is the initial investment cost, operation and maintenance cost of the equipment; C ADN is the energy interaction cost with the active distribution network layer; C MG is the electricity interaction cost with other microgrid systems; C LOAD is the compensation cost of interruptible load; C P is the penalty fee; the cost calculation is shown in formula (4) and (5):
[0025]
[0026] μ(r, l) = r·(1+r) l / [(1+r) l -1] (5)
[0027] In the formula, c gas is the unit price of natural gas; Q GT (t) is the heating power of the gas turbine at time t; Q GB (t) is the heat output of the gas boiler at time t; μ(r, l) is the depreciation coefficient; r is the discount rate of the equipment; l is the life cycle of the equipment; N K is the number of types of equipment; C 1,k and C M,k are the initial investment cost and operation and maintenance cost of the kth type of equipment; N k is the installed capacity of the kth type of equipment; P N,k Rated power of equipment of category k; c buy,MG (t) and c sell,MG (t) are the unit prices of electricity purchased or sold from the multi-microgrid system to other microgrid systems at time t; P buy,MG,a (t), P sell,MG,a (t) are the power purchased from other microgrid systems and the power sold to other microgrid systems by the microgrid system at time t; c buy,ADN(t), c sell,ADN (t) are the unit price of electricity purchased from the active distribution network and the unit price of electricity sold to the active distribution network at the multi-microgrid system layer at time t, and c buy,ADN (t) is also the unit price of electricity purchased by independent user load from the active distribution network; P buy,MG,b (t), P sell,MG,b (t) are the power purchased by the microgrid system from the active distribution network and the power sold to the active distribution network at time t; c Load is the compensation cost of interruptible load at time t; ΔP Load,III,break (t) is the load actually interrupted at time t; ξ is the penalty coefficient; P v,MG (t), P w,MG (t) are the shortfall and waste of the microgrid system energy at time t; Q v,MG (t), Q w,MG (t) are the shortfall and waste of thermal energy in the microgrid system at time t;
[0028]
[0029] Where w = 1, 2, 3, and W∈{CO2, NO, SO2} is the type of pollutant; c w The cost of treating each gram of Class W pollutants by the microgrid system; w,GT , χ w,GB , χ w,EC and χ w,ADN are the emission coefficients of the wth type of pollutants from gas turbines, gas boilers, electric refrigerators and electricity purchased from the active distribution network; P GT (t) is the power generation of the gas turbine at time t; Q GB (t) is the heat output of the gas boiler at time t; Q EC (t) is the cooling energy output by the electric refrigerator at time t; P buy,MG (t) is the power purchased by the microgrid system from the active distribution network at time t;
[0030] In the active distribution network layer, the optimization objective function with the lowest operating cost is shown in formula (7), and the operating cost of the generator set, the electric energy interaction cost and the penalty cost of the active distribution network layer are shown in formulas (8) and (9):
[0031] min f=f1+f2+f3 (7)
[0032] Where, f is the operating cost of the active distribution network layer; f1 is the operating cost of the generator set; f2 is the electric energy interaction cost; f3 is the penalty fee of the active distribution network layer;
[0033]
[0034] Where N G is the number of generating units; a q 、b q and c q is the operating parameter of the qth unit; P G,q (t) is the output power of the qth unit at time t; C MMG is the transaction cost between the active distribution network and the multi-microgrid system; C G is the transaction fee between the active distribution network and the main grid; P v,ADN (t), P w,ADN (t) are the power shortage and waste of active distribution network layer at time t;
[0035]
[0036] Where N MG is the number of microgrid systems; P buy,MG,n (t), P sell,MG,n (t) are the power purchased and sold by the nth microgrid system to the active distribution network at time t, and are related to P buy,MG (t) and P sell,MG (t) corresponding; c buy,G (t), c sell,G (t) are the unit price of electricity purchased and sold by the active distribution network to the main grid at time t; P buy,G (t), P sell,G (t) are the power purchased and sold by the active distribution network to the main grid at time t.
[0037] Constraints:
[0038] In the multi-microgrid system layer, the load balance constraint, energy storage system capacity constraint (battery capacity constraint and heat storage tank capacity constraint), equipment output constraint, equipment configuration quantity constraint, and transaction constraint with the active distribution network are shown in equations (10), (11), (12), (13), (14), and (15), respectively:
[0039]
[0040] Where P WT (t), P PV (t), P BIO (t) are the output power of wind power generation system, photovoltaic power generation system and biogas power generation system at time t; P c (t), P d (t) are the charging and discharging power of the battery at time t; P Load,after (t) is the electric load demand of the microgrid system before load transfer and load interruption at time t; P EC(t) is the electric energy consumed by the refrigerator at time t; Q WHB (t) is the heat output of the waste heat boiler at time t; Q c (t), Q d (t) are the heat storage and heat release powers of the heat storage tank at time t; Q AC,IN (t), Q HE,IN (t) are the heat energy input by the absorption refrigerator and heat converter at time t; Q AC (t), Q EC (t) are the cooling energy output by the absorption refrigerator and the electric refrigerator at time t; Q Load,c (t) is the cooling load demand of the microgrid system at time t;
[0041]
[0042] In the formula, and are the minimum and maximum energy storage capacity of the battery respectively; E BAT (t) is the remaining energy of the battery at time t; and are the minimum and maximum charging power of the battery respectively; and are the minimum and maximum discharge power of the battery respectively;
[0043]
[0044] In the formula, and are the minimum and maximum energy storage capacity of the heat storage tank respectively; Q HST (t) is the thermal energy stored in the heat storage tank at time t; and are the minimum and maximum heat storage power of the heat storage tank respectively; and are the minimum and maximum heat release powers of the heat storage tank respectively;
[0045]
[0046] In the formula, and They are the minimum outputs of wind power generation system, photovoltaic power generation system, biogas power generation system, gas turbine and gas boiler respectively; and They are the maximum outputs of wind power generation system, photovoltaic power generation system, biogas power generation system, gas turbine and gas boiler respectively;
[0047]
[0048] In the formula, and N is the minimum number of wind turbines, photovoltaic cells, gas turbines, batteries, heat storage tanks and gas boilers. WT 、N PV 、N GT 、N BAT 、N HST and N GB They are the configuration quantities of the above corresponding devices respectively; and They are the maximum configuration quantities of the above corresponding devices respectively; and is the minimum and maximum intake volume of biogas; C BIO (t) is the intake volume of biogas at time t;
[0049]
[0050] In the formula, and are the minimum and maximum power purchase of the microgrid system respectively; and are the minimum and maximum electricity sales of the microgrid system respectively;
[0051] In the user load layer, the load constraints and electricity satisfaction constraints are shown in equations (16) and (17):
[0052]
[0053] In the formula, and are the minimum values of transferable load and interruptible load at time t respectively; and are the maximum values of the transferable load and interruptible load at time t; ΔP Load,II,trans (t)ΔP Load,III,break (t) and are the loads actually transferred and actually interrupted at time t, respectively;
[0054]
[0055] In the formula, s 1,min and 2,min are the minimum values of satisfaction with electricity consumption mode and electricity expenditure at time t; s 1,max and 2,max are the maximum values of satisfaction with electricity consumption mode and electricity expenditure at time t, respectively; s1 and s2 are the satisfaction with electricity consumption mode and electricity expenditure, respectively, as shown in formulas (18) and (19):
[0056]
[0057] Where ΔPt It represents the change of the user's electric load in each time period after the implementation of the demand response strategy; P t,0 It represents the electricity load demand of the user in each time period before the implementation of the demand response strategy;
[0058]
[0059] Where s2 is the satisfaction with electricity expenditure; ΔC t Indicates the change in the user's electricity expenditure in each time period after the implementation of the demand response strategy; C t,0 It represents the electricity expenditure of users in each time period before the implementation of the demand response strategy;
[0060] In the active distribution network layer, the load balance constraint, unit output constraint, and tie line transmission capacity constraint are shown in equations (20), (21), and (22):
[0061]
[0062] Where N G 、N MG are the number of generator sets and microgrid systems respectively; P G,j (t) is the output power of the qth unit at time t; P D (t) is the independent user load demand at the active distribution network layer at time t;
[0063]
[0064] In the formula, and are the minimum and maximum output of the jth generator set respectively;
[0065]
[0066] In the formula, and They are the minimum and maximum electricity sales from the active distribution network to the main grid; and They are the minimum and maximum power purchases from the active distribution network to the main grid;
[0067] Step 2: Select seasonal characteristics and use K-means clustering algorithm to select scenes from the data set to obtain corresponding environmental data and cold, hot, and electric load conditions;
[0068] Step 3: Based on the fuzzy clustering method, each microgrid system is divided into peak, flat and valley time periods. At the same time, an innovative hybrid demand response strategy is used to analyze the electric load within the user load layer. Each load agent transmits the load results after demand response to the corresponding microgrid agent.
[0069] The steps of the hybrid demand response strategy are as follows:
[0070] 1) Determine the initial electricity price elasticity matrix, the upper limit of interruptible load and the time-of-use electricity price of each microgrid system;
[0071] 2) Analyze the original power load data of each time period to obtain the proportion of the initial non-responsive load, transferable load and interruptible load in each time period;
[0072] 3) Based on environmental data and equipment parameters of each microgrid system, calculate the output of renewable energy in each time period;
[0073] 4) Compare the initial load demand response results with the output of renewable energy. If the output of renewable energy in the peak and flat time periods can meet the electricity demand in the time period, no load transfer or interruption will occur. If the output of renewable energy cannot meet the electricity demand in the time period, the remaining electricity load will adopt incentive-based demand response or price-based demand response under the condition that the renewable energy is fully absorbed;
[0074] 5) Modify the electricity price elasticity matrix according to the actual transferable load.
[0075] Among them, the electricity price elasticity matrix is shown in formula (23), the electricity quantity and price elasticity coefficients in the matrix are shown in formulas (24) and (25), the price-type demand response model is shown in formula (26), and the load model is shown in formulas (27) and (28):
[0076]
[0077] In the formula, p, f and l represent the peak, flat and valley time periods in the time-of-use electricity price mechanism respectively;
[0078]
[0079]
[0080] Where P c (i, i) is the elastic coefficient; P c (i, j) is the cross elastic coefficient; P i and ΔP i They represent the electric load before demand response and the load change after demand response in time period i respectively; c i and Δc i They represent the electricity price before and after the demand response in time period i; c j and Δc j They represent the electricity price before demand response and the change in electricity price after demand response in time period j respectively;
[0081]
[0082] Where P p , P f and P l are the average load in each period after price-based demand response; P p,0 , P f,0 and P l,0 are the average loads in each period before price-based demand response; Δc p , Δc f and Δc l Respectively represent the change in electricity price in each period after demand response; c p 、c f and c l Indicates the electricity price before adopting time-of-use electricity price;
[0083] P Load (t) = P Load,I (t)+P Load,II (t)+P Load,III (t) (27)
[0084] Where P Load,I (t), P Load,II (t) and P Load,III (t) are the demands of the non-responsive load, transferable load and interruptible load in the microgrid system at time t;
[0085]
[0086] Where P Load,before (t) is the electric load demand of the microgrid system after load transfer and load interruption at time t; t’,t is a variable, λ t’,t =1 means that the transferable load is transferred from time t' to time t, λ t’,t =-1 means that the transferable load is transferred from time t to time t'; σ t is a variable, σ t =1 means that a load interruption occurs at time t, σ t =0 means no load interruption occurs at time t;
[0087] Step 4: Based on the information of each parameter, the multi-microgrid system layer is solved using the sparrow search algorithm to obtain the optimal result;
[0088] Step 5, judging the optimal result of the multi-microgrid system layer obtained in step 4, if an abnormal value appears in the result, readjusting the upper and lower limits of the constraint conditions and the adaptive weight coefficient, and returning to step 4;
[0089] Step 6: When the optimal result of the multi-microgrid system layer does not show abnormal values, it is determined whether to conduct internal power trading. If internal trading is conducted, the internal pricing strategy is executed to complete the internal trading. If internal trading is not conducted, the lower layer result is directly passed to the active distribution network layer.
[0090] The total supply and demand in the multi-microgrid system is shown in formula (29), and the supply-demand ratio and demand-supply ratio are shown in formula (30):
[0091]
[0092] Where P sell (t) and P buy (t) are the total energy supply and total energy demand of the multi-microgrid system at time t;
[0093]
[0094] Where Y(t) and V(t) are the supply-demand ratio and demand-supply ratio of the multi-microgrid system at time t, respectively;
[0095] 1) When there is no supply and demand relationship, P sell (t) and P buy (t) are all 0, and the internal electricity price is shown in formula (31):
[0096]
[0097] 2) When supply and demand are balanced, P sell (t) and P buy (t) is equal and not 0, and the internal electricity price is as shown in formula (32):
[0098]
[0099] 3) When supply is less than demand, P sell (t) is less than P buy (t), the internal electricity price is shown in formula (33):
[0100]
[0101] 4) When supply exceeds demand, P sell (t) greater than P buy (t), the internal electricity price is shown in formula (34):
[0102]
[0103] Step 7: The active distribution network energy decision center reads the parameter information of the generator set and the independent load, uses the sparrow search algorithm to solve and obtain the optimal result of the active distribution network layer, and feeds back the power sales situation from the upper layer to the lower layer to the energy management systems of each microgrid at the lower layer;
[0104] Step 8: judge the final result of hierarchical energy management. If it can effectively reduce the operating cost of the active distribution network containing multiple microgrid systems and improve the reliability of system operation, save the relevant results. If the expected goals are not achieved, it is still necessary to readjust the upper and lower limits of the constraints and the adaptive weight coefficients, and return to step 3.
[0105] Step 9, determine whether the active distribution network operation strategy containing the multi-microgrid system meets the stop condition. The stop condition is that the number of daily operation cycles T has been reached. If the stop condition is met, the final result is output; if the stop condition is not met, the current daily operation time is increased by 1, the time period is updated, and step 3 is re-entered. It is repeatedly iterated and solved until the stop condition is met and the operation is exited;
[0106] The above-mentioned hierarchical energy management operation strategy considering hybrid demand response, the microgrid agent, the active distribution network energy decision center, and the microgrid energy management system are existing technologies and are well known to those skilled in the art;
[0107] The above-mentioned hierarchical energy management operation strategy considering hybrid demand response, the K-means clustering algorithm, fuzzy clustering method, and sparrow search algorithm are existing methods;
[0108] The above-mentioned hierarchical energy management operation strategy considering hybrid demand response, the various powers, loads, etc. are well known to those skilled in the art;
[0109] The advantages and positive effects of the present invention are:
[0110] 1. In the hierarchical energy management model of the active distribution network containing multiple microgrid systems, the present invention introduces a user load layer inside each microgrid system, and applies the demand response strategy to the hierarchical energy management, so as to reasonably transfer the electric load in the time period of insufficient output of renewable energy to the time period of excess output, which can fully utilize renewable energy and alleviate the power supply pressure during the peak power consumption period to a certain extent.
[0111] 2. The present invention adopts dynamic adaptive weight coefficients to construct a comprehensive objective function, converting the multi-objective optimization problem into a single-objective optimization problem, which not only reflects the certainty of the microgrid system for the optimization of important objectives, but also gives the microgrid system a certain active adjustment capability;
[0112] 3. The present invention designs a hybrid demand response strategy, which enables the energy management of the multi-microgrid system layer to achieve the maximum consumption of renewable energy while considering the participation of internal users in load demand response, and to maximize the economic efficiency, reliability and timeliness of users' electricity consumption, reduce load transfer or interruption, and effectively reduce the operating cost of the entire system and the amount of electricity purchased from the active distribution network layer;
[0113] 4. The present invention designs an internal transaction pricing strategy for electric energy that takes into account transaction fairness and interactive rationality, better mobilizes the enthusiasm of each microgrid system to participate in internal transactions, and realizes the optimal management of system energy, thereby achieving the purpose of peak shaving, valley filling, cost reduction and improved operational stability. BRIEF DESCRIPTION OF THE DRAWINGS
[0114] The present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0115] Figure 1 It is a flow chart of a hierarchical energy management operation strategy considering hybrid demand response;
[0116] Figure 2 It is a structural diagram of a hierarchical energy management model considering hybrid demand response;
[0117] Figure 3 It is a schematic diagram of the operation strategy of an improved microgrid system based on heat to determine electricity;
[0118] Figure 4 It is a model structure diagram of a combined cooling, heating and power microgrid system;
[0119] Figure 5 This is a result diagram of selecting application scenarios for the environmental data of this embodiment in spring;
[0120] Figure 6 This is a diagram showing the selection results of application scenarios for different loads in spring days in this embodiment;
[0121] Figure 7 is a diagram of the internal transaction price of electric energy at the multi-microgrid system layer of this embodiment; DETAILED DESCRIPTION
[0122] Figure 1The process of a hierarchical energy management operation strategy considering hybrid demand response is as follows: start → set and initialize parameters → select scenarios for data sets → analyze the electric load within the user load layer → solve the multi-microgrid system layer → determine the optimal result and handle abnormal values → if there is an internal transaction, execute the internal pricing strategy, otherwise pass the lower layer result to the active distribution network layer → the active distribution network energy decision center reads the information and feeds back the electricity sales situation to each microgrid energy management system → determine whether the final result meets expectations and processes → determine whether the stop condition is met → output the result → end;
[0123] Figure 2 It is shown that a hierarchical energy management model considering a hybrid demand response strategy includes a user load layer, a multi-microgrid system layer, and an active distribution network layer. The user load layer is a lower-layer model, and a corresponding load agent is set up to transmit the user's load information to the microgrid agent and receive the demand response request sent by the microgrid agent. In each microgrid system, the microgrid agent and the load agent can formulate scheduling strategies and operation strategies for each unit according to environmental information, load demand information, and the configuration of distributed power units, and then use the sparrow search algorithm to obtain the optimal microgrid system energy management results, and send scheduling instructions to each distributed unit and user load. In addition, the microgrid energy management system monitors the output of renewable energy and the capacity of the energy storage system in real time in each time period. When the output of renewable energy is less than the load demand, a demand response request is sent to the load agent through the microgrid agent. After receiving the request, the load agent transfers or interrupts the user's electric load according to the actual situation. Among them, the multi-microgrid system layer is the middle layer and core layer of the entire hierarchical energy management model. It has the ability to transmit energy and information with the upper and lower models, and has the ability to formulate internal electricity prices and demand response strategies. Among them, EDC is the energy decision center, EMS is the energy management system, MMGA is the multi-microgrid agent, MGA is the microgrid agent, MG1 and MG2 are demand-type microgrid systems, MG3 is the energy supply microgrid system, and LA is the load agent;
[0124] Figure 3 It shows that an improved microgrid system operation strategy based on heat is: the waste heat of the biogas power generation system should match the demand of the absorption chiller and heat exchanger at any time. If the waste heat is excessive, the heat storage tank will store the excess heat energy, resulting in energy waste after reaching the upper limit of the energy storage; if the waste heat is insufficient, the shortage will be supplemented by gas boilers, heat storage tanks and electric chillers. The electricity generated by wind, photovoltaic and biogas power generation systems at any time should match the demand of electrical load and electric chillers. If the electricity is insufficient, the electricity shortage will be supplemented by batteries and active distribution networks, or the active distribution network will purchase electricity from the main grid; if the electricity is excessive, the excess electricity will be absorbed by batteries and active distribution networks;
[0125] Figure 4A model of a combined cooling, heating and power microgrid system is shown, in which the main distributed power sources include wind power generation system, photovoltaic power generation system, biogas power generation system and gas turbine; the energy storage system includes batteries and heat storage tanks; the auxiliary energy supply equipment includes waste heat boilers, gas boilers, heat converters, absorption refrigerators and electric refrigerators, etc., and the user loads inside the microgrid system include electrical load, thermal load and cooling load.
[0126] Example
[0127] The present invention proposes a hierarchical energy management operation strategy considering hybrid demand response. The model is built using a PC as a platform, where the CPU is i5-8300H 2.3GHz, the installed memory is 8G, the operating system is Windows 10-64 bit, and the MATLAB R2016a version is used. The process of the hierarchical energy management operation strategy is as follows: Figure 1 As shown in Figure 2, the model structure of hierarchical energy management is as follows: Figure 2 The hierarchical energy management model adopts a combined cooling, heating and power microgrid system. The operation strategy of the microgrid system is as follows: Figure 3 As shown in Figure 2, the model structure of the microgrid system is as follows: Figure 4 shown.
[0128] Step 1: Set device parameters and constraints, and determine the data source;
[0129] Step 2: Select scenes from the data set;
[0130] Step 3: Divide the time periods and use the hybrid demand response strategy to analyze the electric load within the user load layer;
[0131] Step 4, solving the multi-microgrid system layer to obtain the optimal result;
[0132] Step 5: judge the optimal result. If an abnormal value appears, adjust the corresponding coefficient and return to step 4;
[0133] Step 6: If there is an internal transaction, the internal pricing strategy is executed, otherwise the lower layer result is passed to the active distribution network layer;
[0134] Step 7: The active distribution network energy decision center reads the information and feeds back the electricity sales situation to each microgrid energy management system;
[0135] Step 8: Determine the final result. If it meets expectations, save the relevant results. Otherwise, adjust the correlation coefficient and return to step 3.
[0136] Step 9, determine whether the stop condition is met, if it is met, output the final result, otherwise exit the operation after the iteration is met;
[0137] Step 1: Set the objective function, upper and lower limits of constraints, equipment parameters, etc. of the active distribution network containing multiple microgrid systems, and determine the data source of renewable energy output and load demand based on relevant data sets;
[0138] First, establish the objective function: In the multi-microgrid system layer with the user load layer added, first use equations (3)-(5) to establish the economic cost objective function, but do not consider C when each microgrid system is running independently. MG , use formula (6) to establish the environmental cost objective function, and then use formulas (1) and (2) to construct the comprehensive objective function; in the active distribution network layer, use formulas (8) and (9) to represent the operating cost of the generator set, the electric energy interaction cost and the penalty fee of the active distribution network layer, and use formula (7) to establish the optimization objective function with the lowest operating cost;
[0139] Then, the constraint conditions are established: in the multi-microgrid system layer, the load balance constraint is established by formula (10), the energy storage system capacity constraint (battery capacity constraint and heat storage tank capacity constraint) is established by formula (11) and (12), the equipment output constraint is established by formula (13), the equipment configuration quantity constraint is established by formula (14), and the transaction constraint with the active distribution network is established by formula (15); in the user load layer, the load constraint condition is established by formula (16), the power satisfaction constraint condition is established by formula (17), and the power consumption mode satisfaction s1 and power consumption expenditure satisfaction s2 in formula (17) are calculated by formula (18) and formula (19); in the active distribution network layer, the load balance constraint is established by formula (20), the unit output constraint is established by formula (21), and the interconnection line transmission capacity constraint is established by formula (22);
[0140] In the formulas (1)-(22) of this embodiment, the linear correlation coefficient is n1+n2=0.8; the natural gas unit price c gas Take 2.05 yuan / m 3; The operation cycle of the system is T = 24h; The operating parameters of each generator set in the active distribution network layer are shown in Table 1; The relevant parameters and costs of various equipment in the multi-microgrid system layer are shown in Table 2, where WT is a wind power generation system, PV is a photovoltaic power generation system, BIO is a biogas power generation system, GT is a gas turbine, GB is a gas boiler, WHB is a waste heat boiler, HE is a heat converter, AC is an absorption refrigerator, EC is an electric refrigerator, BAT is a battery, HST is a heat storage tank, there are two types of gas turbines, batteries and heat storage tanks, and the rest of the equipment is one type; The depreciation coefficient μ(r, l) of the equipment is set to 0.08; The maximum and minimum charge states of the energy storage system are 0.9 and 0.2 respectively; The unit prices of the treatment of pollutants CO2, NO and SO2 are 0.000 21 yuan / g, 0.063 0 yuan / g and 0.014 8 yuan / g; the emission of different pollutants is shown in Table 3; the time-of-use electricity prices of the multi-microgrid system layer and the active distribution network layer, and the active distribution network layer and the main grid are shown in Table 4; the air intake per unit time of the biogas power generation system is set to the uniform air intake mode, and the lower limit of the biogas intake per unit time is 50m 3 / h, the upper limit is 70m 3 / h, and the number of biogas generators is configured according to the total amount of biogas power generation; for the equipment composition and equipment quantity constraints of each microgrid system, MG1 is the same as MG2, but different from MG3, and the minimum number of equipment is 1, as shown in Table 5; the minimum value s of the satisfaction with electricity consumption mode and electricity expenditure at time t in the user load layer 1,min and 2,min Take 0.8, and the corresponding maximum value is 1;
[0141] Table 1 Generator parameters at the active distribution network layer
[0142]
[0143] Table 2 Equipment parameters of multi-microgrid system layer
[0144]
[0145]
[0146] Table 3 Pollutant emission coefficients
[0147]
[0148] Table 4 Time-of-use electricity price
[0149]
[0150] Table 5 Equipment composition and upper limit of quantity of each microgrid system
[0151]
[0152] Step 2: Select wind speed, ambient temperature, light intensity, daily load rate, number of daily load cycles, and cold / hot load ratio as seasonal characteristics of the application scenario, and use the K-means clustering algorithm to select scenes from the data set to obtain environmental data and cold, hot, and electric load conditions for the four seasons.
[0153] In this embodiment, the environmental data of the energy management application scenario is derived from the Meteonorm 8.0 meteorological data set. The environmental data of each microgrid system is consistent. The load data is derived from the load data of 16 commercial reference buildings in different regions released by the U.S. Department of Energy. This embodiment uses the spring day scene in Los Angeles as an example, and uses the restaurants, supermarkets, and shopping malls in the data set as MG1, the hotels and apartments as MG2, and the office area as MG3. The application scenario selection results of the environmental data are shown in the figure below. Figure 5 As shown in the figure, the application scenario selection results of MG1, MG2, MG3 loads and active distribution network independent loads Figure 6 As shown;
[0154] Step 3: Based on the fuzzy clustering method, each microgrid system is divided into peak, flat and valley time periods. At the same time, an innovative hybrid demand response strategy is used to analyze the electric load within the user load layer. Each load agent transmits the load result after demand response to the corresponding microgrid agent. The steps of the hybrid demand response strategy are as follows:
[0155] 1) Determine the initial electricity price elasticity matrix, the upper limit of interruptible load and the time-of-use electricity price of each microgrid system;
[0156] 2) Analyze the original power load data of each time period to obtain the proportion of the initial non-responsive load, transferable load and interruptible load in each time period;
[0157] 3) Based on environmental data and equipment parameters of each microgrid system, calculate the output of renewable energy in each time period;
[0158] 4) Compare the initial load demand response results with the output of renewable energy. If the output of renewable energy in the peak and flat time periods can meet the electricity demand in the time period, no load transfer or interruption will occur. If the output of renewable energy cannot meet the electricity demand in the time period, the remaining electricity load will adopt incentive-based demand response or price-based demand response under the condition that the renewable energy is fully absorbed;
[0159] 5) Modify the electricity price elasticity matrix according to the actual transferable load;
[0160] In this embodiment, the electricity price elasticity matrix is established using formula (23), and the electricity price self-elasticity coefficient P in the matrix is established using formulas (24) and (25). c (i, i) and the cross elastic coefficient P c (i, j), use formula (26) to establish a price-based demand response model, and use formulas (27) and (28) to establish a load model; among them, the peak-flat-valley time division results of each microgrid system in spring are shown in Table 6. The maximum proportion of interruptible load per unit time is 0.1, and its compensation cost is 0.3 yuan / kWh;
[0161] Table 6 Peak-average-valley time division results of each microgrid system in spring
[0162]
[0163] Step 4: Based on the information of each parameter, the multi-microgrid system layer is solved using the sparrow search algorithm to obtain the optimal result;
[0164] Step 5, judging the optimal result of the multi-microgrid system layer obtained in step 4, if an abnormal value appears in the result, readjusting the upper and lower limits of the constraint conditions and the adaptive weight coefficient, and returning to step 4;
[0165] Step 6: When the optimal result of the multi-microgrid system layer does not show abnormal values, it is determined whether to conduct internal power trading. If internal trading is conducted, the internal pricing strategy is executed to complete the internal trading. If internal trading is not conducted, the lower layer result is directly passed to the active distribution network layer.
[0166] In this embodiment, the total energy supply P of the multi-microgrid system at time t is calculated using formula (29): sell (t) and total demand P buy (t), use formula (30) to calculate the supply-demand ratio Y(t) and demand-supply ratio V(t) of the multi-microgrid system at time t; for the internal pricing strategy, when there is no supply-demand relationship, use formula (31) to calculate the internal electricity price; when supply and demand are balanced, use formula (32) to calculate the internal electricity price; when supply is less than demand, use formula (33) to calculate the internal electricity price; when supply is greater than demand, use formula (33) to calculate the internal electricity price, and obtain the internal transaction price of electricity at the multi-microgrid system level as follows: Figure 7 As shown;
[0167] Step 7: The active distribution network energy decision center reads the parameter information of the generator set and the independent load, uses the sparrow search algorithm to solve and obtain the optimal result of the active distribution network layer, and feeds back the power sales situation from the upper layer to the lower layer to the energy management systems of each microgrid at the lower layer;
[0168] Step 8: judge the final result of hierarchical energy management. If it can effectively reduce the operating cost of the active distribution network containing multiple microgrid systems and improve the reliability of system operation, save the relevant results. If the expected goals are not achieved, it is still necessary to readjust the upper and lower limits of the constraints and the adaptive weight coefficients, and return to step 3.
[0169] Step 9, determine whether the active distribution network operation strategy containing multiple microgrid systems meets the stop condition. The stop condition is that the number of daily operation cycles T has been reached. If the stop condition is met, the final result is output; if the stop condition is not met, the current daily operation time is increased by 1, the time period is updated, and step 3 is re-entered. The iterative operation is repeated until the stop condition is met, and the operation is exited to obtain the total operation cost and total power purchase amount of each system, as shown in Table 7 Scenario IV;
[0170] In this embodiment, in order to analyze the rationality and feasibility of the operation strategy of the present invention, four scenarios I, II, III, and IV are set up for comparison, and the total operation cost and total power purchase of each system are calculated as shown in Table 7:
[0171] (1) Scenario I: Each microgrid system operates independently, without considering the division of peak-flat-valley time periods and the demand response of user loads;
[0172] (2) Scenario II: Each microgrid system adopts an interconnected operation mode, taking into account the internal transaction of electric energy, but not considering the division of peak-flat-valley time periods and the demand response of user loads;
[0173] (3) Scenario III: Each microgrid system operates independently, taking into account the division of peak-flat-valley time periods and the demand response strategy of user loads;
[0174] (4) Scenario IV: Each microgrid system adopts an interconnected operation mode, taking into account the internal transaction of electric energy, and considering the division of peak-flat-valley time periods and the hybrid demand response strategy of user loads;
[0175] Table 7 Comparison results of different scenarios
[0176]
[0177] According to Table 7, a hierarchical energy management operation strategy considering hybrid demand response can effectively reduce the total operating cost of the entire system and the total power purchase at the multi-microgrid system level, and effectively improve the economic benefits of system operation.
[0178] In the above embodiments, the microgrid agent, active distribution network energy decision center, and microgrid energy management system are existing technologies and are well known to technicians in this technical field; the K-means clustering algorithm, fuzzy clustering method, and sparrow search algorithm are existing methods.
[0179] It should be emphasized that the embodiments described in the present invention are illustrative rather than restrictive, and therefore the present invention is not limited to the embodiments described in the specific implementation manners. Any other implementation manners derived by those skilled in the art based on the technical solutions of the present invention also fall within the scope of protection of the present invention.
Claims
1. A hierarchical energy management operation method considering hybrid demand response, characterized in that The following steps are involved: Step 1: Set device parameters and constraints and determine data sources; Step 2: Select scenes from the dataset; Step 3: Divide the time period and use the hybrid demand response strategy to analyze the electric load within the user load layer; Step 4: Solve the multi-microgrid system layer to obtain the optimal result; Step 5: Determine the optimal result. If an abnormal value occurs, adjust the corresponding coefficient and return to step 4. Step 6: If there is an internal transaction, the internal pricing strategy is executed; otherwise, the lower layer result is passed to the active distribution network layer; Step 7: The active distribution network energy decision center reads the information and feeds back the electricity sales situation to the energy management systems of each microgrid; Step 8: Determine the final result. If it meets expectations, save the relevant results. Otherwise, adjust the correlation coefficient and return to step 3. Step 9: Determine whether the stopping condition is met. If it is met, output the final result. Otherwise, exit the operation after the iteration is met. Step 1: Set the equipment parameters and upper and lower limits of the constraints of the active distribution network containing multiple microgrids, and determine the data source of renewable energy output and load demand based on relevant data sets; The user load layer is added to the multi-microgrid system layer, and the dynamic adaptive weight coefficient is used to construct the comprehensive objective function as shown in formula (1), where the economic cost and environmental cost are shown in formulas (2) and (4): Where F is the comprehensive objective function; T is the operating period of the system; g1(t) and g2(t) are the economic cost and environmental cost respectively; ω1 and ω2 are the dynamic adaptive weight coefficients; min g1=C F +C M +C ADN +C MG +C LOAD +C P (2) Where C F is the natural gas consumption cost; C M is the initial investment cost, operation and maintenance cost of the equipment; C ADN is the energy interaction cost with the active distribution network layer; C MG is the electricity interaction cost with other microgrid systems; C LOAD is the compensation cost of interruptible load; C P is the penalty fee; the calculation of each cost is shown in formula (3): Where C gas is the unit price of natural gas; Q GT (t) is the heating power of the gas turbine at time t; Q GB (t) is the heat output of the gas boiler at time t; μ(r, l) is the depreciation coefficient; r is the discount rate of the equipment; l is the life cycle of the equipment; N K is the number of types of equipment; C I,k and C M,k are the initial investment cost and operation and maintenance cost of the kth type of equipment respectively; N k is the installed quantity of the kth type of equipment; P N,k Rated power of equipment of category k; c buy,MG (t) and c sell,MG (t) are the unit prices of electricity purchased or sold from the multi-microgrid system to other microgrid systems at time t; P buy,MG,a (t), P sell,MG,a (t) are the power purchased from other microgrid systems and the power sold to other microgrid systems at time t; c buy,ADN (t), c sell,ADN (t) are the unit price of electricity purchased from the active distribution network and the unit price of electricity sold to the active distribution network at the multi-microgrid system layer at time t, and c buy,ADN (t) is also the unit price of electricity purchased by independent user load from the active distribution network; P buy,MG,b (t), P sell,MG,b (t) are the power purchased from the active distribution network and the power sold to the active distribution network by the microgrid system at time t; c Load is the compensation cost of the interruptible load at time t; ΔP Load,III,break (t) is the load actually interrupted at time t; ξ is the penalty coefficient; P v,MG (t), P w,MG (t) are the shortfall and waste of the microgrid system’s electric energy at time t; Q v,MG (t), Q w,MG (t) are the shortfall and waste of thermal energy in the microgrid system at time t; Where w = 1, 2, 3, and W∈{CO2, NO, SO2} is the type of pollutant; c w The cost of treating each gram of Class W pollutants by the microgrid system; w,GT , χ w,GB 、X w,EC and χ w,ADN are the emission coefficients of the wth type of pollutants from gas turbines, gas boilers, electric refrigerators, and electricity purchased from the active distribution network; P GT (t) is the power generation of the gas turbine at time t; Q GB (t) is the heat output of the gas boiler at time t; Q EC (t) is the cooling energy output by the electric refrigerator at time t; P buy,MG (t) is the power purchased by the microgrid system from the active distribution network at time t; In the multi-microgrid system layer, load balance constraints, energy storage system capacity constraints, equipment output constraints, equipment configuration quantity constraints, and transaction constraints with the active distribution network are established. The load balance constraint is shown in formula (5): Where, P WT (t), P PV (t), P BIO (t) are the output power of wind power generation system, photovoltaic power generation system and biogas power generation system at time t; P c (t), P d (t) are the charging and discharging power of the battery at time t; P Load,after (t) is the electric load demand of the microgrid system before load transfer and load interruption at time t; P EC (t) is the electric energy consumed by the refrigerator at time t; Q WHB (t) is the heat output of the waste heat boiler at time t; Q c (t), Q d (t) are the heat storage and heat release powers of the heat storage tank at time t; Q AC,IN (t), Q HE,IN (t) are the heat energy input by the absorption refrigerator and heat converter at time t; Q AC (t), Q EC (t) are the cooling energy output by the absorption refrigerator and the electric refrigerator at time t; Q Load,c (t) is the cooling load demand of the microgrid system at time t; In the user load layer, load constraints and electricity satisfaction constraints are established as shown in equations (6) and (7): Where, and are the minimum values of transferable load and interruptible load at time t respectively; and They are the maximum values of the load that can be transferred and the load that can be interrupted at any moment; ΔP Load,II,trans (t)ΔP Load,III,break (t) and (t) are the loads actually transferred and actually interrupted at time t, respectively; Where s 1,min and s 2,min are the minimum values of satisfaction with electricity consumption mode and electricity expenditure at time t; s 1,max and s 2,max are the maximum values of satisfaction with electricity consumption mode and electricity expenditure at time t; s1 and s2 are the satisfaction with electricity consumption mode and electricity expenditure, respectively, as shown in formulas (8) and (9): Where, AP t It represents the change of the user's electric load in each time period after the implementation of the demand response strategy; P t,0 It represents the user's electricity load demand in each time period before the implementation of the demand response strategy; Where s2 is the satisfaction with electricity expenditure; ΔC t Indicates the change in the user's electricity expenditure in each time period after the implementation of the demand response strategy; C t,0 It represents the electricity expenditure of users in each time period before the implementation of the demand response strategy; Step 2: Select seasonal characteristics and use the K-means clustering algorithm to select scenes from the data set to obtain the corresponding environmental data and cooling, heating, and electricity load conditions; Step 3: Based on the fuzzy clustering method, each microgrid system is divided into peak, flat, and off-peak periods. At the same time, an innovative hybrid demand response strategy is used to analyze the electrical load within the user load layer. Each load agent transmits the load results after demand response to the corresponding microgrid agent. The load model is established as shown in equations (10) and (11): P Load (t)=P Load,I (t)+P Load,II (t)+P Load,III (t) (10) Where, P Load,I (t), P Load,II (t) and P Load,III (t) are the demands of the non-responsive load, transferable load and interruptible load within the microgrid system at time t; Where, P Load,before (t) is the electric load demand of the microgrid system after load transfer and load interruption at time t; λ t’,t is a variable, λ t’,t =1 means the transferable load is transferred from time t' to time t, λ t’,t =-1 means the transferable load is transferred from time t to time t'; σ t is a variable, σ t =1 means that a load interruption occurs at time t, σ t =0 means no load interruption occurs at time t; Step 4: Based on the parameter information, the sparrow search algorithm is used to solve the multi-microgrid system layer to obtain the optimal result; Step 5: Determine the optimal result of the multi-microgrid system layer obtained in step 4. If an abnormal value appears in the result, readjust the upper and lower limits of the constraint conditions and the adaptive weight coefficient, and return to step 4. Step 6: When the optimal result of the multi-microgrid system layer does not show any abnormal value, it is determined whether to conduct internal energy trading. If internal trading is to be conducted, the internal pricing strategy is implemented and the internal trading is completed. If internal trading is not to be conducted, the lower-level result is directly passed to the active distribution network layer. Step 7: The active distribution network energy decision center reads the parameter information of the generator set and independent load, uses the sparrow search algorithm to solve and obtain the optimal result of the active distribution network layer, and simultaneously feeds back the power sales situation from the upper layer to the lower layer to the energy management system of each microgrid in the lower layer; Step 8: Determine the final result of the hierarchical energy management. If it can effectively reduce the operating cost of the active distribution network containing multiple microgrids and improve the reliability of system operation, save the relevant results. If it does not achieve the expected goals, it is still necessary to readjust the upper and lower limits of the constraints and the adaptive weight coefficients, and return to step 3. Step 9: Determine whether the active distribution network operation method including the multi-microgrid system meets the stopping condition, where the stopping condition is that the number of daily operation cycles T has been reached. If the stopping condition is met, output the final result; If the stop condition is not met, the current daily running time is increased by 1, the time period is updated, and step 3 is re-entered. The solution is iterated repeatedly until the stop condition is met and the operation is exited.
2. A hierarchical energy management operation method considering hybrid demand response according to claim 1, characterized in that: The steps of the hybrid demand response strategy are as follows: 1) Determine the initial electricity price elasticity matrix, the upper limit of interruptible load, and the time-of-use electricity price of each microgrid system; 2) Analyze the original power load data of each time period to obtain the proportion of the initial non-responsive load, transferable load and interruptible load in each time period; 3) Calculate the output of renewable energy in each time period based on environmental data and equipment parameters of each microgrid system; 4) Compare the initial load demand response results with the output of renewable energy. If the output of renewable energy during peak and off-peak periods can meet the electricity demand during that period, no load shifting or interruption will occur. If the output of renewable energy cannot meet the electricity demand during that period, incentive-based demand response or price-based demand response will be adopted for the remaining electricity load, provided that the renewable energy is fully absorbed. 5) Modify the electricity price elasticity matrix according to the actual transferable load.
