A method and system for energy dispatching in a photovoltaic-storage-DC-flexible distribution network based on load forecasting

Abstract

This invention provides a method and system for energy dispatching in a photovoltaic-storage-DC-flexible distribution network based on load forecasting. The method includes: acquiring load forecasting data and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network; inputting the load forecasting data and photovoltaic power generation data into a preset outer-layer energy management model, so that the outer-layer energy management model generates a first energy dispatching scheme according to an outer-layer objective equation and decision variables, and updates the decision variables according to the first energy dispatching scheme; inputting the first energy dispatching scheme into an inner-layer energy management model, so that the inner-layer energy management model optimizes the first energy dispatching scheme according to an inner-layer objective equation and the decision variables, generates a second energy dispatching scheme, and updates the decision variables according to the second energy dispatching scheme; and performing energy dispatching within the current dispatching cycle according to the second energy dispatching scheme to improve the operational stability and efficiency of the photovoltaic-storage-DC-flexible distribution network.

Description

Technical Field

[0001] This invention relates to the field of power systems and their automation, and in particular to a method and system for energy dispatching of photovoltaic-storage-DC-flexible distribution networks based on load forecasting. Background Technology

[0002] The construction and operation of a photovoltaic-storage-DC-flexible distribution network will enhance the distributed absorption capacity of clean energy, improve the overall operating efficiency of the energy system, reduce the proportion of power generation from high-emission fossil fuels, alleviate my country's energy shortage problem, provide flexible and energy-saving integrated energy services for end users, improve the living environment of residents, and generate significant indirect economic benefits. Photovoltaic integration and DC transformation have altered the power flow distribution of the distribution network and the electricity consumption characteristics of users. The increasing complexity of the distribution network significantly increases the difficulty of precise energy management. Accurate energy management is essential for rationally formulating power generation plans to achieve supply and demand balance. Energy management can better reflect future electricity demand in power system planning, thereby optimizing resource allocation and planning power generation, transmission, and distribution facilities. This helps avoid resource waste and improve system efficiency. It is of great significance for increasing energy utilization efficiency and improving economic and social benefits.

[0003] Currently, commonly used energy management strategies mainly fall into two categories: First, energy management strategies based on mathematical algorithms. These methods generally rely on algorithms derived through rigorous mathematical derivation based on certain accepted theories, primarily including mixed-integer linear programming, interior-point methods, and consensus algorithms. Intelligent algorithms require initialization; while these algorithms are rigorously derived mathematically and have high reliability, their robustness is relatively poor. Second, these methods consider the impact of load structure, the relative size of photovoltaic output and energy storage with the load, and peak-valley electricity pricing on electricity economy, but neglect user comfort during peak hours.

[0004] Second, there are energy management strategies based on multiple time scales. These methods include intraday and day-ahead forecasting. The accuracy of these forecasts is affected by environmental factors, leading to discrepancies between the distribution network's dispatch plan and the actual operating state of the grid. However, accumulated discrepancies can negatively impact the stable operation of the grid. Furthermore, the selection of different time scales within these multiple time scales can increase the solution time for the dispatch plan, while too few time scales can increase the pressure for timely adjustments. Summary of the Invention

[0005] This invention provides an energy dispatching method and system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting. The method optimizes the resource allocation and energy management of the photovoltaic-storage-DC-flexible distribution network according to load forecasting data, thereby improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible distribution network.

[0006] In a first aspect, the present invention provides an energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting, comprising:

[0007] Within the current scheduling cycle, acquire load forecast data and photovoltaic power generation data for the photovoltaic-storage-DC-flexible distribution network;

[0008] The load forecast data and photovoltaic power generation data are input into a preset outer energy management model, so that the outer energy management model generates a first energy dispatch scheme according to the outer objective equation and decision variables, updates the decision variables according to the first energy dispatch scheme, and then transmits the decision variables to a preset inner energy management model.

[0009] The first energy scheduling scheme is input into the inner energy management model, so that the inner energy management model optimizes the first energy scheduling scheme according to the inner objective equation and the decision variables, generates a second energy scheduling scheme, updates the decision variables according to the second energy scheduling scheme, and then passes the decision variables to the outer energy management model.

[0010] According to the second energy dispatch scheme, the energy dispatch of the photovoltaic-storage-DC-flexible power distribution network is carried out within the current dispatch cycle.

[0011] This invention provides an energy dispatch method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting. Within each dispatch cycle, based on load forecasting data and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, an outer energy management model generates a corresponding first energy dispatch scheme. This first energy dispatch scheme maximizes the satisfaction of the preset outer objective equation, performing preliminary optimization and energy management of the photovoltaic-storage-DC-flexible distribution network. Furthermore, considering the potential error between load forecasting data and actual load data, and the uncertainty of photovoltaic power generation, the first energy dispatch scheme generated by the outer energy management model can be further optimized. Therefore, the first energy dispatch scheme is input into an inner energy management model and optimized through the inner objective equation to generate a second energy dispatch scheme. This reduces the impact of photovoltaic power generation fluctuations and load forecasting errors on energy management, improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible distribution network. In addition, in this embodiment of the invention, decision variables are set to record the calculation results of certain key parameters after each run of the outer and inner energy management models. These variables are used as set values ​​during the operation of the outer and inner energy management models to guide the models in calculations, thereby further reducing the impact of uncertainties in photovoltaic power generation and load forecasting errors on energy dispatch.

[0012] In one possible implementation, the outer energy management model is constructed based on a Min-max robust model predictive controller, comprising:

[0013] With minimizing operating costs as the optimization objective, and in conjunction with the Min-max robust model predictive controller, the outer objective equation is established, wherein the operating costs include power loss costs, battery aging costs, and user comfort costs.

[0014] Based on the power supply structure, load forecast data, and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, multiple constraint equations are established, including power constraint equations, battery differential equations, and battery SOC limiting equations.

[0015] This invention provides a method for constructing an outer-layer energy management model. First, with minimizing operating costs as the optimization objective, an outer-layer objective equation is established using a Min-max robust model predictive controller. During the optimization process, the power loss cost of the distribution network, the aging cost of batteries, and the user comfort cost are comprehensively considered, thereby improving user comfort while reducing operating costs. Furthermore, by constructing power constraint equations, battery differential equations, and battery SOC constraint equations, the optimization boundary of the model is established to prevent over-optimization in pursuit of the optimization objective, which could affect the normal operation of the power grid. The power constraint equations ensure that all equipment in the power grid operates within its normal power range, while the battery differential equations and battery SOC constraint equations ensure that the battery's electrical energy and state of charge are within preset ranges. This invention, by constructing an outer-layer energy management model based on a Min-max robust model predictive controller, performs preliminary optimization configuration and energy management for a photovoltaic-storage-DC-flexible distribution network, improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible distribution network.

[0016] Furthermore, the specific formula for the outer objective equation is as follows:

[0017]

[0018]

[0019]

[0020] in, For power loss costs, For the cost of battery aging, A * Cost for user comfort; m m The unit cost of photovoltaic power generation, For photovoltaic power generation, m B The unit cost of battery discharge, M represents the battery discharge power, Δt is one scheduling cycle; DC(t,d(Δt)) represents the aging cost of a battery within the time interval Δt, d(Δt) represents the depth of charge and discharge of the battery within the time interval Δt, and M... AC P(t) represents the average cost per unit of energy lost during a single charge / discharge event, and P(t) represents the charge / discharge power of the battery.

[0021] In one possible implementation, the power constraint equation includes a power balance constraint equation and a power imbalance constraint equation. The power balance constraint equation consists of the power of the flexible load, the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system, constraining the sum of the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system to equal the power of the flexible load. The power imbalance constraint equation consists of upper and lower limits of the power transmitted from the distribution network and the power generated by the battery, constraining the power transmitted from the distribution network and the power generated by the energy storage device to be within a preset range.

[0022] The battery differential equation constrains the battery energy at the current moment by using the battery energy at the previous moment, the battery charging and discharging efficiency, and the battery power output at the current moment.

[0023] The battery SOC constraint equation constrains the ratio of the battery's rated capacity to its actual capacity at the current moment by setting upper and lower limits for the battery's SOC.

[0024] This invention further defines the various constraint equations. The power balance constraint equation ensures the dynamic balance between the power output from the distribution network, the power generated by the battery, and the power of photovoltaic power generation and flexible load power. The power imbalance constraint equation is used to ensure that the power output from the distribution network and the power generated by the battery are within a preset range. However, since photovoltaic power generation has significant uncertainties, no imbalance constraint is applied. Instead, the fluctuations in photovoltaic power generation are balanced by adjusting the power output from the distribution network and the power generated by the battery. Furthermore, the power generated by the battery is affected by its own electrical energy and state of charge (SOC). Therefore, a battery differential equation and a battery SOC limit equation are set for the battery to ensure stable operation and extend its service life.

[0025] In one possible implementation, the inner-layer energy management model is constructed based on a droop control strategy, including:

[0026] The droop control strategy is constructed based on the droop characteristics of the photovoltaic-storage-DC-flexible power distribution network. The droop control strategy is used to control the voltage level of the photovoltaic-storage-DC-flexible power distribution network within a preset range.

[0027] Using the first battery power output and the first photovoltaic power output in the first energy dispatch scheme as reference values, and combining load forecasting data and photovoltaic power generation data, the inner objective equation is constructed with the optimization objective of minimizing load forecasting error and photovoltaic fluctuations.

[0028] The constraint equations of the inner energy management model are the same as those of the outer energy management model.

[0029] This invention provides a method for constructing an inner-layer energy management model. Based on the droop characteristics of the photovoltaic-storage-DC-flexible power distribution network, a droop control strategy is constructed to control the voltage level of the network within a preset range, thereby improving its operational stability. In constructing the inner-layer objective equation, the optimization results of the outer-layer energy management model are utilized to reduce the impact of photovoltaic power generation fluctuations and load forecasting errors on energy management, thus improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible power distribution network. Simultaneously, for the same reasons as the outer-layer energy management model, the same constraint equations are used to ensure the normal operation of the power grid system.

[0030] Furthermore, the specific formula for the drooping characteristic is as follows:

[0031] f i =f ni -m i (P i -P ni )

[0032] U i =U ni -n i (Q i -Q nj )

[0033] Among them, f i and U i These represent the frequency and amplitude of the photovoltaic-storage-DC-flexible distribution network voltage, respectively. i and Q i These are active power and reactive power, respectively; f ni U ni P ni Q nj These are the rated frequency, voltage, active power, and reactive power, respectively; m i and n i These are the active and reactive power droop coefficients, respectively.

[0034] The specific formula for controlling the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range is as follows:

[0035] P n1 / P1=P n2 / P2

[0036] Δf=m1(P1-P n1 )=m2(P2-P n2 )

[0037] Where P1 and P2 are the actual power at the beginning and end of the transmission line, P n1 and P n2 denoted as , where m1 and m2 are the droop coefficients at the beginning and end of the transmission line, and Δf is the frequency variation.

[0038] Furthermore, the specific formula for the inner objective equation is as follows:

[0039]

[0040]

[0041]

[0042] in, For electricity load costs, These are the weighting coefficients of the inner energy management model. The deviation between the power output of the battery in the inner energy management model and the reference value. This represents the deviation between the photovoltaic power generation in the inner energy management model and the reference value. The power output of the first battery in the first energy dispatch scheme. To provide power to the batteries in the inner energy management model, The first photovoltaic power generation capacity in the first energy dispatch scheme. This refers to the photovoltaic power generation in the inner energy management model.

[0043] In one possible implementation, before inputting the load forecast data and photovoltaic power generation data into a preset outer energy management model, the flexible loads in the load forecast data are divided into shiftable loads, reduceable loads, and interruptible loads according to different load characteristics.

[0044] In this embodiment of the invention, flexible loads in the load forecast data are divided into three categories based on different load characteristics: shiftable loads, reduceable loads, and interruptible loads. Shiftable loads are continuously operating loads with constant power and fixed operating durations. Their operating time periods can be shifted, but they are constrained by the range of operating duration and start-up time, and must be prioritized during scheduling. Reduceable loads maintain a constant total power consumption within the scheduling cycle, but their power consumption time can be flexibly adjusted. Reduceable loads can be prioritized for reduction during peak power consumption periods. Interruptible loads are loads with time-segmented operation, constant power, and fixed total operating durations. Interruptible loads can be temporarily disconnected during peak power consumption periods to prioritize other types of loads. Implementing effective scheduling control based on the characteristics of different load types allows for the development of more reasonable energy scheduling schemes, further improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible power distribution network.

[0045] Secondly, correspondingly, the present invention provides an energy dispatching system for photovoltaic-storage-DC-flexible distribution networks based on load forecasting, including an acquisition module, an outer energy management module, an inner energy management module, and a dispatching module;

[0046] The acquisition module is used to acquire load forecast data and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network within the current scheduling cycle.

[0047] The outer energy management module is used to input the load forecast data and photovoltaic power generation data into the preset outer energy management model, so that the outer energy management model generates a first energy dispatch scheme according to the outer objective equation and decision variables, updates the decision variables according to the first energy dispatch scheme, and then transmits the decision variables to the preset inner energy management model.

[0048] The inner energy management module is used to input the first energy scheduling scheme into the inner energy management model, so that the inner energy management model optimizes the first energy scheduling scheme according to the inner objective equation and the decision variables, generates a second energy scheduling scheme, updates the decision variables according to the second energy scheduling scheme, and then passes the decision variables to the outer energy management model.

[0049] The scheduling module is used to perform energy scheduling on the photovoltaic-storage-DC-flexible power distribution network within the current scheduling cycle according to the second energy scheduling scheme.

[0050] In one possible implementation, the outer energy management model is constructed based on a Min-max robust model predictive controller, comprising:

[0051] With minimizing operating costs as the optimization objective, and in conjunction with the Min-max robust model predictive controller, the outer objective equation is established, wherein the operating costs include power loss costs, battery aging costs, and user comfort costs.

[0052] Based on the power supply structure, load forecast data, and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, multiple constraint equations are established, including power constraint equations, battery differential equations, and battery SOC limiting equations.

[0053] Furthermore, the specific formula for the outer objective equation is as follows:

[0054]

[0055]

[0056]

[0057] in, For power loss costs, For the cost of battery aging, A * Cost for user comfort; m m The unit cost of photovoltaic power generation, For photovoltaic power generation, m B The unit cost of battery discharge, M represents the battery discharge power, Δt is one scheduling cycle; DC (t,d(Δt)) represents the aging cost of a battery within the time interval Δt, d(Δt) represents the depth of charge and discharge of the battery within the time interval Δt, and M... AC P(t) represents the average cost per unit of energy lost during a single charge / discharge event, and P(t) represents the charge / discharge power of the battery.

[0058] In one possible implementation, the power constraint equation includes a power balance constraint equation and a power imbalance constraint equation. The power balance constraint equation consists of the power of the flexible load, the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system, constraining the sum of the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system to equal the power of the flexible load. The power imbalance constraint equation consists of upper and lower limits of the power transmitted from the distribution network and the power generated by the battery, constraining the power transmitted from the distribution network and the power generated by the energy storage device to be within a preset range.

[0059] The battery differential equation constrains the battery energy at the current moment by using the battery energy at the previous moment, the battery charging and discharging efficiency, and the battery power output at the current moment.

[0060] The battery SOC constraint equation constrains the ratio of the battery's rated capacity to its actual capacity at the current moment by setting upper and lower limits for the battery's SOC.

[0061] In one possible implementation, the inner-layer energy management model is constructed based on a droop control strategy, including:

[0062] The droop control strategy is constructed based on the droop characteristics of the photovoltaic-storage-DC-flexible power distribution network. The droop control strategy is used to control the voltage level of the photovoltaic-storage-DC-flexible power distribution network within a preset range.

[0063] Using the first battery power output and the first photovoltaic power output in the first energy dispatch scheme as reference values, and combining load forecasting data and photovoltaic power generation data, the inner objective equation is constructed with the optimization objective of minimizing load forecasting error and photovoltaic fluctuations.

[0064] The constraint equations of the inner energy management model are the same as those of the outer energy management model.

[0065] Furthermore, the specific formula for the drooping characteristic is as follows:

[0066] f i =f ni -m i (P i -P ni )

[0067] U i =U ni -n i (Q i -Q nj )

[0068] Among them, f i and U i These represent the frequency and amplitude of the photovoltaic-storage-DC-flexible distribution network voltage, respectively. i and Q i These are active power and reactive power, respectively; f ni U ni P ni Q nj These are the rated frequency, voltage, active power, and reactive power, respectively; m i and n i These are the active and reactive power droop coefficients, respectively.

[0069] The specific formula for controlling the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range is as follows:

[0070] P n1 / P1=P n2 / P2

[0071] Δf=m1(P1-P n1 )=m2(P2-P n2 )

[0072] Where P1 and P2 are the actual power at the beginning and end of the transmission line, P n1 and P n2 denoted as , where m1 and m2 are the droop coefficients at the beginning and end of the transmission line, and Δf is the frequency variation.

[0073] Furthermore, the specific formula for the inner objective equation is as follows:

[0074]

[0075]

[0076]

[0077] in, For electricity load costs, These are the weighting coefficients of the inner energy management model. The deviation between the power output of the battery in the inner energy management model and the reference value. This represents the deviation between the photovoltaic power generation in the inner energy management model and the reference value. The power output of the first battery in the first energy dispatch scheme. To provide power to the batteries in the inner energy management model, The first photovoltaic power generation capacity in the first energy dispatch scheme. This refers to the photovoltaic power generation in the inner energy management model.

[0078] In one possible implementation, before the outer energy management module inputs the load forecast data and photovoltaic power generation data into the preset outer energy management model, the flexible loads in the load forecast data are divided into shiftable loads, reduceable loads, and interruptible loads according to different load characteristics. Attached Figure Description

[0079] Figure 1 This is a flowchart illustrating an embodiment of an energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting provided by the present invention.

[0080] Figure 2 This is a schematic diagram of the algorithm flow of an embodiment of the energy dispatching method for photovoltaic-storage-DC-flexible distribution networks based on load forecasting provided by the present invention.

[0081] Figure 3This is a graph showing the changes in photovoltaic output and base load power in an embodiment of a photovoltaic-storage-DC-flexible power dispatching method for a distribution network based on load forecasting, provided by the present invention.

[0082] Figure 4 This is a graph showing the change of battery SOC value over time in an embodiment of a photovoltaic-storage-DC-flexible power grid energy dispatching method based on load forecasting provided by the present invention.

[0083] Figure 5 The graph shows the change in battery charging and discharging power over time in an embodiment of a photovoltaic-storage-DC-flexible power dispatching method for power distribution networks based on load forecasting, provided by the present invention.

[0084] Figure 6 This is a comparison chart of battery SOC value and charging / discharging power changes in an embodiment of a photovoltaic-storage-DC-flexible power distribution network energy dispatching method based on load forecasting provided by the present invention.

[0085] Figure 7 This is a schematic diagram of an embodiment of a photovoltaic-storage-DC-flexible power grid energy dispatching system based on load forecasting provided by the present invention.

[0086] Figure 8 : A comparison table of different battery cost models and corresponding algorithms.

[0087] Figure 9 : A comparison table of energy management results for photovoltaic-storage-DC-flexible power distribution networks using different battery cost models. Detailed Implementation

[0088] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0089] It should be noted that the step numbers in this document are only for the convenience of explaining the specific embodiments and are not intended to limit the order in which the steps are performed.

[0090] Throughout this specification, the Min-max robust model predictive controller described herein is a predictive controller incorporating a robust regulation loop, thus forming a robust predictive controller. Robust predictive control is a method that handles model uncertainties within a predictive control framework, enabling the controlled system to reach asymptotic stability under feasible conditions. The structural design based on robust predictive control constitutes the robust predictive controller. The droop control described in this specification can be applied to inverter parallel systems and is a type of distributed control. Compared to centralized control, which relies heavily on a single module in the parallel system, and master-slave control, which limits the distance between power sources, droop control only requires information from the power sources themselves and eliminates the need for interconnecting signal lines. By collecting the output of each inverter and applying a given control strategy, multiple inverters can operate in parallel. It offers excellent redundancy, a simple structure, low cost, and system reliability. Droop control simulates the droop characteristics of synchronous generators in traditional power systems. Its working principle is as follows: the inverter power supply detects the magnitude of its own output power and decouples the active and reactive power for control. Based on the droop characteristics, reference values ​​for output frequency and voltage amplitude are obtained, thereby rationally allocating the active and reactive power of the system and maintaining the system voltage within the specified range.

[0091] Example 1:

[0092] like Figure 1 As shown, Embodiment 1 provides an energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting, including steps S1-S4:

[0093] Step S1: Within the current scheduling cycle, obtain load forecast data and photovoltaic power generation data for the photovoltaic-storage-DC-flexible distribution network;

[0094] Step S2: Input the load forecast data and photovoltaic power generation data into the preset outer energy management model so that the outer energy management model generates a first energy dispatch scheme according to the outer objective equation and decision variables, updates the decision variables according to the first energy dispatch scheme, and then transmits the decision variables to the preset inner energy management model.

[0095] Step S3: Input the first energy scheduling scheme into the inner layer energy management model so that the inner layer energy management model optimizes the first energy scheduling scheme according to the inner layer objective equation and the decision variables, generates a second energy scheduling scheme, updates the decision variables according to the second energy scheduling scheme, and then passes the decision variables to the outer layer energy management model;

[0096] Step S4: Perform energy dispatching on the photovoltaic-storage-DC-flexible power distribution network within the current dispatching cycle according to the second energy dispatching scheme.

[0097] The algorithm flowchart of this invention embodiment is as follows: Figure 2 As shown.

[0098] This invention provides an energy dispatch method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting. Within each dispatch cycle, based on load forecasting data and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, an outer energy management model generates a corresponding first energy dispatch scheme. This first energy dispatch scheme maximizes the satisfaction of the preset outer objective equation, performing preliminary optimization and energy management of the photovoltaic-storage-DC-flexible distribution network. Furthermore, considering the potential error between load forecasting data and actual load data, and the uncertainty of photovoltaic power generation, the first energy dispatch scheme generated by the outer energy management model can be further optimized. Therefore, the first energy dispatch scheme is input into an inner energy management model and optimized through the inner objective equation to generate a second energy dispatch scheme. This reduces the impact of photovoltaic power generation fluctuations and load forecasting errors on energy management, improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible distribution network. In addition, in this embodiment of the invention, decision variables are set to record the calculation results of certain key parameters after each run of the outer and inner energy management models. These variables are used as set values ​​during the operation of the outer and inner energy management models to guide the models in calculations, thereby further reducing the impact of uncertainties in photovoltaic power generation and load forecasting errors on energy dispatch.

[0099] In one possible implementation, in step S2, the outer energy management model is constructed based on a Min-max robust model predictive controller, including:

[0100] With minimizing operating costs as the optimization objective, and in conjunction with the Min-max robust model predictive controller, the outer objective equation is established, wherein the operating costs include power loss costs, battery aging costs, and user comfort costs.

[0101] Based on the power supply structure, load forecast data, and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, multiple constraint equations are established, including power constraint equations, battery differential equations, and battery SOC limiting equations.

[0102] Furthermore, the specific formula for the outer objective equation is as follows:

[0103]

[0104]

[0105]

[0106] in, For power loss costs, For the cost of battery aging, A *Cost for user comfort; m m The unit cost of photovoltaic power generation, For photovoltaic power generation, m B The unit cost of battery discharge, M represents the battery discharge power, Δt is one scheduling cycle; DC (t,d(Δt)) represents the aging cost of a battery within the time interval Δt, d(Δt) represents the depth of charge and discharge of the battery within the time interval Δt, and M... AC P(t) represents the average cost per unit of energy lost during a single charge / discharge event, and P(t) represents the charge / discharge power of the battery.

[0107] In a preferred embodiment, the user comfort cost is constructed by combining the indoor temperature adaptability index and the indoor humidity adaptability index. The indoor temperature adaptability index describes the user's satisfaction with the indoor temperature, and the indoor humidity adaptability index describes the user's satisfaction with the indoor humidity. To achieve satisfactory indoor humidity and indoor temperature for the user, the corresponding equipment needs to be activated, thereby generating the corresponding load cost.

[0108] The expression for the indoor temperature adaptability index is:

[0109]

[0110] t ac (t+1)=e -Δt / RC t ac (t)+R(e -Δt / RC -1)P ac (t)+(1-e -Δt / RC )T o (t)

[0111] in, t is the indoor temperature adaptability index. ac Indoor temperature, t ac (t+1) represents the indoor temperature at time t+1, which is determined by the indoor temperature t at the previous time. ac (t), air conditioning power P ac (t) and outdoor temperature T o (t) are jointly determined, where R is the indoor thermal resistance and C is the room heat capacity. When the indoor temperature is 26℃, The lowest temperature setting ensures the highest user comfort.

[0112] The expression for the indoor temperature adaptability index is:

[0113]

[0114] Where, m wh and m vh0The indoor air humidity before and after the air conditioner is turned on, Δm mh The maximum acceptable humidity variation for the user, l wh (t) represents the start / stop state of the air conditioner at time t.

[0115] For electrical appliances with flexible loads, different appliances exhibit variations in comfort metrics across multiple dimensions, including temperature, power consumption, and time. A normalization approach is used to address this. User comfort cost lies within the interval [0,1], and its magnitude is negatively correlated with user comfort. The expression for user comfort cost is established as follows:

[0116]

[0117] Furthermore, regarding the construction of the mathematical model for the battery, the lifespan expression is first established based on the two main determinants of battery life: actual full capacity and depth of discharge.

[0118] L(d) = a × d -b ×e -cd

[0119] Where L(d) is the battery life under the condition of battery charge-discharge depth d, a, b, c are curve fitting coefficients, all of which are greater than 0, d represents the battery charge-discharge depth, and e is the natural base.

[0120] Then, an expression for the battery charge / discharge depth is established, where the battery charge / discharge depth is the ratio of charge / discharge energy to total capacity during a single charge / discharge process, specifically:

[0121]

[0122]

[0123] Where d(Δt) is the depth of charge / discharge of the battery during the time interval Δt, P(t) is the average discharge power of the battery, and E A (t) is the actual capacity of the battery at time t, E A (t+Δt) is the actual capacity of the battery at time t+Δt, E rated This indicates the rated capacity of the battery.

[0124] Finally, the expression for the average unit energy loss cost of the battery is established:

[0125]

[0126] Among them, M AC Let M be the average cost per unit of energy lost during a single charge-discharge event, and η be the battery replacement cost. Bc and η BdThis is the battery charge / discharge efficiency coefficient.

[0127] This invention provides a method for constructing an outer-layer energy management model. First, with minimizing operating costs as the optimization objective, an outer-layer objective equation is established using a Min-max robust model predictive controller. During the optimization process, the power loss cost of the distribution network, the aging cost of batteries, and the user comfort cost are comprehensively considered, thereby improving user comfort while reducing operating costs. Furthermore, by constructing power constraint equations, battery differential equations, and battery SOC constraint equations, the optimization boundary of the model is established to prevent over-optimization in pursuit of the optimization objective, which could affect the normal operation of the power grid. The power constraint equations ensure that all equipment in the power grid operates within its normal power range, while the battery differential equations and battery SOC constraint equations ensure that the battery's electrical energy and state of charge are within preset ranges. This invention, by constructing an outer-layer energy management model based on a Min-max robust model predictive controller, performs preliminary optimization configuration and energy management for a photovoltaic-storage-DC-flexible distribution network, improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible distribution network.

[0128] In one possible implementation, in step S2, the power constraint equation includes a power balance constraint equation and a power imbalance constraint equation. The power balance constraint equation is composed of the power of the flexible load, the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system, constraining the sum of the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system to be equal to the power of the flexible load. The power imbalance constraint equation is composed of the upper and lower limits of the power transmitted from the distribution network and the upper and lower limits of the power generated by the battery, constraining the power transmitted from the distribution network and the power generated by the energy storage device to be within a preset range.

[0129] The battery differential equation constrains the battery energy at the current moment by using the battery energy at the previous moment, the battery charging and discharging efficiency, and the battery power output at the current moment.

[0130] The battery SOC constraint equation constrains the ratio of the battery's rated capacity to its actual capacity at the current moment by setting upper and lower limits for the battery's SOC.

[0131] The specific formula for the power balance constraint equation is as follows:

[0132] P L (t)=P M (t)+P B (t)+P PV (t)

[0133] Among them, P L (t) represents the power of the flexible load, P M(t) represents the power output from the distribution network, P B (t) represents the power generated by energy storage devices such as batteries, P PV (t) represents the photovoltaic power generation.

[0134] The specific formula for the power imbalance constraint equation is as follows:

[0135]

[0136]

[0137] in, and These are the upper and lower limits of the power output from the distribution network. and These are the upper and lower limits of the power output of the battery.

[0138] The specific formula for the battery difference equation is as follows:

[0139]

[0140] Among them, E B (t) and E B (t-1) represent the electrical energy of the battery at times t and t-1, respectively; η Bc and η Bd These represent the battery charging and discharging efficiency, respectively.

[0141] The specific formula for the battery SOC limiting equation is as follows:

[0142]

[0143] in, and These represent the upper and lower limits of the battery's SOC constraint, respectively; E B (t) and E BA (t) represents the rated capacity and actual capacity of the battery at time t, respectively.

[0144] This invention further defines the various constraint equations. The power balance constraint equation ensures the dynamic balance between the power output from the distribution network, the power generated by the battery, and the power of photovoltaic power generation and flexible load power. The power imbalance constraint equation is used to ensure that the power output from the distribution network and the power generated by the battery are within a preset range. However, since photovoltaic power generation has significant uncertainties, no imbalance constraint is applied. Instead, the fluctuations in photovoltaic power generation are balanced by adjusting the power output from the distribution network and the power generated by the battery. Furthermore, the power generated by the battery is affected by its own electrical energy and state of charge (SOC). Therefore, a battery differential equation and a battery SOC limit equation are set for the battery to ensure stable operation and extend its service life.

[0145] In one possible implementation, in step S3, the inner-layer energy management model is constructed based on a droop control strategy, including:

[0146] The droop control strategy is constructed based on the droop characteristics of the photovoltaic-storage-DC-flexible power distribution network. The droop control strategy is used to control the voltage level of the photovoltaic-storage-DC-flexible power distribution network within a preset range.

[0147] Using the first battery power output and the first photovoltaic power output in the first energy dispatch scheme as reference values, and combining load forecasting data and photovoltaic power generation data, the inner objective equation is constructed with the optimization objective of minimizing load forecasting error and photovoltaic fluctuations.

[0148] The constraint equations of the inner energy management model are the same as those of the outer energy management model.

[0149] This invention provides a method for constructing an inner-layer energy management model. Based on the droop characteristics of the photovoltaic-storage-DC-flexible power distribution network, a droop control strategy is constructed to control the voltage level of the network within a preset range, thereby improving its operational stability. In constructing the inner-layer objective equation, the optimization results of the outer-layer energy management model are utilized to reduce the impact of photovoltaic power generation fluctuations and load forecasting errors on energy management, thus improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible power distribution network. Simultaneously, for the same reasons as the outer-layer energy management model, the same constraint equations are used to ensure the normal operation of the power grid system.

[0150] Furthermore, the specific formula for the drooping characteristic is as follows:

[0151] f i =f ni -m i (P i -P ni )

[0152] U i =U ni -n i (Q i -Q nj )

[0153] Among them, f i and U i These represent the frequency and amplitude of the photovoltaic-storage-DC-flexible distribution network voltage, respectively. i and Q i These are active power and reactive power, respectively; f ni U ni P ni Q njThese are the rated frequency, voltage, active power, and reactive power, respectively; m i and n i These are the active and reactive power droop coefficients, respectively.

[0154] The specific formula for controlling the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range is as follows:

[0155] P n1 / P1=P n2 / P2

[0156] Δf=m1(P1-P n1 )=m2(P2-P n2 )

[0157] Where P1 and P2 are the actual power at the beginning and end of the transmission line, P n1 and P n2 denoted as , where m1 and m2 are the droop coefficients at the beginning and end of the transmission line, and Δf is the frequency variation.

[0158] Furthermore, the specific formula for the inner objective equation is as follows:

[0159]

[0160]

[0161]

[0162] in, For electricity load costs, These are the weighting coefficients of the inner energy management model. The deviation between the power output of the battery in the inner energy management model and the reference value. This represents the deviation between the photovoltaic power generation in the inner energy management model and the reference value. The power output of the first battery in the first energy dispatch scheme. To provide power to the batteries in the inner energy management model, The first photovoltaic power generation capacity in the first energy dispatch scheme. This refers to the photovoltaic power generation in the inner energy management model.

[0163] In one possible implementation, before inputting the load forecast data and photovoltaic power generation data into a preset outer energy management model, the flexible loads in the load forecast data are divided into shiftable loads, reduceable loads, and interruptible loads according to different load characteristics.

[0164] The expression for the movable load is as follows:

[0165]

[0166] In the formula, This represents the upper and lower limits of the startup time range. and d i These represent the start-up time, shutdown time, and operating duration of load i, respectively.

[0167] The expression for the interruptible load is:

[0168]

[0169] The expression for the load that can be reduced is:

[0170] T i (t+1)=e -Δt / RC T i (t)+R(e -Δt / RC -1)P ac (t)+(1-e -Δt / RC )T o (t)

[0171] In the formula, T i R is the indoor temperature, in °C; C is the indoor thermal resistance, in °C / kW; P is the room heat capacity, in kWh / °C; ac Air conditioner power, unit is kW; T o This is the outdoor temperature, in °C.

[0172] In a preferred embodiment, different pricing models are adopted for different types of loads. These different pricing models affect the operating costs in the objective function, making the first energy dispatch scheme generated by the outer energy management model more accurate and better reflect actual electricity consumption. Based on the interrelationship between dynamic electricity prices and user response, two different pricing methods can be used: time-of-use pricing and real-time pricing. Time-of-use pricing is generally formulated and published at an earlier time scale, giving users ample time to plan their electricity consumption. Real-time pricing reflects the supply and demand changes at various times, allowing suppliers and users to reasonably share market risks, and only applies to price-sensitive and rapidly responsive flexible loads.

[0173] In this embodiment of the invention, flexible loads in the load forecast data are divided into three categories based on different load characteristics: shiftable loads, reduceable loads, and interruptible loads. Shiftable loads are continuously operating loads with constant power and fixed operating durations. Their operating time periods can be shifted, but they are constrained by the range of operating duration and start-up time, and must be prioritized during scheduling. Reduceable loads maintain a constant total power consumption within the scheduling cycle, but their power consumption time can be flexibly adjusted. Reduceable loads can be prioritized for reduction during peak power consumption periods. Interruptible loads are loads with time-segmented operation, constant power, and fixed total operating durations. Interruptible loads can be temporarily disconnected during peak power consumption periods to prioritize other types of loads. Implementing effective scheduling control based on the characteristics of different load types allows for the development of more reasonable energy scheduling schemes, further improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible power distribution network.

[0174] In a preferred embodiment, the specific implementation method of the energy dispatching method for photovoltaic-storage-DC-flexible distribution networks based on load forecasting provided by the present invention includes:

[0175] Step 1: Import the load, photovoltaic power generation, and energy storage data of a building's photovoltaic-storage-DC-flexible power distribution system from February 1st to February 2nd, 2024 into the MATLAB software workspace. The curves showing the changes in photovoltaic power generation and load over time are shown below. Figure 3 The diagram illustrates the uncertainties of photovoltaic power generation. Based on different load characteristics, the flexible loads of the photovoltaic-storage-DC-flexible distribution system are divided into shiftable loads, interruptible loads, and reduceable loads. A power pricing mechanism is determined based on the interrelationship between dynamic electricity prices and user response.

[0176] Step 2: Establish a battery aging cost model and battery constraint equations based on the charging and discharging frequency, speed, and state of charge (SOC) of the batteries in the energy storage device.

[0177] Step 3: Based on the flexible load power P of the photovoltaic-storage-DC-flexible distribution network L (t), Power output P from the distribution network M (t), the power P generated by energy storage devices such as batteries B (t) and photovoltaic power generation P PV (t), establish the power constraint equation.

[0178] Step 4: Establish a two-layer energy management model based on the Min-max robust model predictive controller. Based on the electricity cost of the photovoltaic-storage-DC-flexible distribution network and the loss cost of the batteries, establish the objective functions for the optimization problems of the inner and outer energy management models. The outer energy management model generates a scheduling scheme within the time range Δt based on load forecasting results, photovoltaic power generation, and electricity prices in the distribution network, reducing the operating cost of the photovoltaic-storage-DC-flexible distribution network during energy scheduling, and updating decision variables as reference values ​​for the inner layer's optimized scheduling. The inner energy management model optimizes the scheduling scheme based on the decision variables. Due to the feedback mechanism of the robust model predictive controller, the uncertainty of the load forecasting results can be compensated.

[0179] Because flexible loads are diverse and possess different mechanistic characteristics, this invention can implement effective scheduling control and rationally plan operation based on the characteristics of different load types. When energy storage scheduling is performed using two different electricity pricing mechanisms—time-of-use pricing and real-time pricing—the changes in the battery's SOC value and charging / discharging power are as follows: Figure 4 and Figure 5 As shown. Furthermore, as... Figure 6 As shown, after using the energy dispatching method for photovoltaic-storage-DC-flexible distribution networks based on load forecasting proposed in this invention, the charging and discharging power of the battery can be more accurately matched with the battery's SOC value, thereby improving the battery's service life and charging and discharging efficiency.

[0180] Regarding battery cost models, besides the battery aging cost model proposed in this invention, existing technologies mainly use fixed aging cost models, piecewise linear models, and quadratic models. The corresponding methods for these models are shown in Table 1. To verify the advantages of the battery aging cost model compared to other models, this embodiment used three other battery cost models and their corresponding methods for comparative experiments. The experimental results are shown in Table 2. Table 2 shows that the dual-layer energy management method proposed in this invention, combined with the battery aging cost model, can effectively reduce the operating costs of energy management in photovoltaic-storage-DC-flexible power distribution networks and the degradation costs of batteries. Furthermore, it achieves the lowest possible discharge rate during energy dispatch, effectively improving battery lifespan.

[0181] Example 2:

[0182] like Figure 7 As shown, correspondingly, Embodiment 2 provides an energy dispatching system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting, including an acquisition module 10, an outer energy management module 20, an inner energy management module 30, and a dispatching module 40.

[0183] The acquisition module 10 is used to acquire load forecast data and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network within the current scheduling cycle.

[0184] The outer energy management module 20 is used to input the load forecast data and photovoltaic power generation data into the preset outer energy management model, so that the outer energy management model generates a first energy dispatch scheme according to the outer objective equation and decision variables, updates the decision variables according to the first energy dispatch scheme, and then transmits the decision variables to the preset inner energy management model.

[0185] The inner energy management module 30 is used to input the first energy scheduling scheme into the inner energy management model, so that the inner energy management model optimizes the first energy scheduling scheme according to the inner objective equation and the decision variables, generates a second energy scheduling scheme, updates the decision variables according to the second energy scheduling scheme, and then passes the decision variables to the outer energy management model.

[0186] The scheduling module 40 is used to perform energy scheduling on the photovoltaic-storage-DC-flexible power distribution network within the current scheduling cycle according to the second energy scheduling scheme.

[0187] In one possible implementation, the outer energy management model is constructed based on a Min-max robust model predictive controller, comprising:

[0188] With minimizing operating costs as the optimization objective, and in conjunction with the Min-max robust model predictive controller, the outer objective equation is established, wherein the operating costs include power loss costs, battery aging costs, and user comfort costs.

[0189] Based on the power supply structure, load forecast data, and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, multiple constraint equations are established, including power constraint equations, battery differential equations, and battery SOC limiting equations.

[0190] Furthermore, the specific formula for the outer objective equation is as follows:

[0191]

[0192]

[0193]

[0194] in, For power loss costs, For the cost of battery aging, A * Cost for user comfort; m m The unit cost of photovoltaic power generation, For photovoltaic power generation, m B The unit cost of battery discharge, M represents the battery discharge power, Δt is one scheduling cycle; DC (t,d(Δt)) represents the aging cost of a battery within the time interval Δt, d(Δt) represents the depth of charge and discharge of the battery within the time interval Δt, and M... AC P(t) represents the average cost per unit of energy lost during a single charge / discharge event, and P(t) represents the charge / discharge power of the battery.

[0195] In one possible implementation, the power constraint equation includes a power balance constraint equation and a power imbalance constraint equation. The power balance constraint equation consists of the power of the flexible load, the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system, constraining the sum of the power transmitted from the distribution network, the power generated by the battery, and the power generated by the photovoltaic system to equal the power of the flexible load. The power imbalance constraint equation consists of upper and lower limits of the power transmitted from the distribution network and the power generated by the battery, constraining the power transmitted from the distribution network and the power generated by the energy storage device to be within a preset range.

[0196] The battery differential equation constrains the battery energy at the current moment by using the battery energy at the previous moment, the battery charging and discharging efficiency, and the battery power output at the current moment.

[0197] The battery SOC constraint equation constrains the ratio of the battery's rated capacity to its actual capacity at the current moment by setting upper and lower limits for the battery's SOC.

[0198] In one possible implementation, the inner-layer energy management model is constructed based on a droop control strategy, including:

[0199] The droop control strategy is constructed based on the droop characteristics of the photovoltaic-storage-DC-flexible power distribution network. The droop control strategy is used to control the voltage level of the photovoltaic-storage-DC-flexible power distribution network within a preset range.

[0200] Using the first battery power output and the first photovoltaic power output in the first energy dispatch scheme as reference values, and combining load forecasting data and photovoltaic power generation data, the inner objective equation is constructed with the optimization objective of minimizing load forecasting error and photovoltaic fluctuations.

[0201] The constraint equations of the inner energy management model are the same as those of the outer energy management model.

[0202] Furthermore, the specific formula for the drooping characteristic is as follows:

[0203] f i =f ni -m i (P i -P ni )

[0204] U i =U ni -n i (Q i -Q nj )

[0205] Among them, f i and U i These represent the frequency and amplitude of the photovoltaic-storage-DC-flexible distribution network voltage, respectively. i and Q i These are active power and reactive power, respectively; f ni U ni P ni Q nj These are the rated frequency, voltage, active power, and reactive power, respectively; m i and n i These are the active and reactive power droop coefficients, respectively.

[0206] The specific formula for controlling the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range is as follows:

[0207] P n1 / P1=P n2 / P2

[0208] Δf=m1(P1-P n1 )=m2(P2-P n2 )

[0209] Where P1 and P2 are the actual power at the beginning and end of the transmission line, P n1 and P n2 denoted as , where m1 and m2 are the droop coefficients at the beginning and end of the transmission line, and Δf is the frequency variation.

[0210] Furthermore, the specific formula for the inner objective equation is as follows:

[0211]

[0212]

[0213]

[0214] in, For electricity load costs, These are the weighting coefficients of the inner energy management model. The deviation between the power output of the battery in the inner energy management model and the reference value. This represents the deviation between the photovoltaic power generation in the inner energy management model and the reference value. The power output of the first battery in the first energy dispatch scheme. To provide power to the batteries in the inner energy management model, The first photovoltaic power generation capacity in the first energy dispatch scheme. This refers to the photovoltaic power generation in the inner energy management model.

[0215] In one possible implementation, before the outer energy management module 20 inputs the load forecast data and photovoltaic power generation data into the preset outer energy management model, the flexible loads in the load forecast data are divided into shiftable loads, reduceable loads, and interruptible loads according to different load characteristics.

[0216] This invention provides an energy dispatch system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting. Within each dispatch cycle, a first energy dispatch scheme is generated using an outer-layer energy management model based on load forecasting data and photovoltaic power generation data. This first energy dispatch scheme maximizes the fulfillment of the preset outer-layer objective equation, enabling preliminary optimization and energy management of the photovoltaic-storage-DC-flexible distribution network. Furthermore, considering the potential errors between load forecasting data and actual load data, and the uncertainty of photovoltaic power generation, the first energy dispatch scheme generated by the outer-layer energy management model can be further optimized. Therefore, the first energy dispatch scheme is input into an inner-layer energy management model and optimized using the inner-layer objective equation to generate a second energy dispatch scheme. This reduces the impact of photovoltaic power generation fluctuations and load forecasting errors on energy management, improving the operational stability and efficiency of the photovoltaic-storage-DC-flexible distribution network. In addition, in this embodiment of the invention, decision variables are set to record the calculation results of certain key parameters after each run of the outer and inner energy management models. These variables are used as set values ​​during the operation of the outer and inner energy management models to guide the models in calculations, thereby further reducing the impact of uncertainties in photovoltaic power generation and load forecasting errors on energy dispatch.

[0217] For a more detailed explanation of the working principle and procedures of this embodiment, please refer to the relevant description in Embodiment 1.

[0218] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for energy dispatching in a photovoltaic-storage-DC-flexible distribution network based on load forecasting, characterized in that, include: Within the current scheduling cycle, acquire load forecast data and photovoltaic power generation data for the photovoltaic-storage-DC-flexible distribution network; The load forecast data and photovoltaic power generation data are input into a preset outer energy management model, so that the outer energy management model generates a first energy dispatch scheme according to the outer objective equation and decision variables, updates the decision variables according to the first energy dispatch scheme, and then transmits the decision variables to a preset inner energy management model. The outer energy management model is constructed based on a Min-max robust model predictive controller, including: establishing the outer objective equation with the goal of minimizing operating costs, combined with the Min-max robust model predictive controller, wherein the operating costs include power loss costs, battery aging costs, and user comfort costs; and establishing multiple constraint equations based on the power supply structure, load forecast data, and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, wherein the constraint equations include power constraint equations, battery differential equations, and battery SOC limiting equations. The first energy scheduling scheme is input into the inner energy management model, so that the inner energy management model optimizes the first energy scheduling scheme according to the inner objective equation and the decision variables, generates a second energy scheduling scheme, updates the decision variables according to the second energy scheduling scheme, and then passes the decision variables to the outer energy management model. The inner-layer energy management model is constructed based on a droop control strategy, including: constructing the droop control strategy according to the droop characteristics of the photovoltaic-storage-DC-flexible distribution network, the droop control strategy being used to control the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range; using the first battery power output and the first photovoltaic power output in the first energy dispatch scheme as reference values, and combining load forecasting data and photovoltaic power output data, constructing the inner-layer objective equation with the optimization objective of minimizing load forecasting error and photovoltaic fluctuations; the constraint equations of the inner-layer energy management model are the same as those of the outer-layer energy management model; According to the second energy dispatch scheme, the energy dispatch of the photovoltaic-storage-DC-flexible power distribution network is carried out within the current dispatch cycle.

2. The energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 1, characterized in that, The specific formula for the outer objective equation is as follows: in, For power loss costs, Due to the cost of battery aging, Costs related to user comfort; The unit cost of photovoltaic power generation, Photovoltaic power generation capacity, The unit cost of battery discharge, This refers to the battery discharge power. One scheduling cycle; for The aging cost of a battery over a time interval. for The depth of charge and discharge of the battery within the time interval. P(t) represents the average cost per unit of energy lost during a single charge / discharge event, and P(t) represents the charge / discharge power of the battery.

3. The energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 1, characterized in that, The power constraint equations include power balance constraint equations and power imbalance constraint equations. The power balance constraint equations consist of flexible load power, power transmitted from the distribution network, power generated by the battery, and photovoltaic power, constraining the sum of the power transmitted from the distribution network, the power generated by the battery, and the photovoltaic power to equal the flexible load power. The power imbalance constraint equations consist of upper and lower limits for the power transmitted from the distribution network and the power generated by the battery, constraining the power transmitted from the distribution network and the power generated by the energy storage device to be within a preset range. The battery differential equation constrains the battery energy at the current moment by using the battery energy at the previous moment, the battery charging and discharging efficiency, and the battery power output at the current moment. The battery SOC constraint equation constrains the ratio of the battery's rated capacity to its actual capacity at the current moment by setting upper and lower limits for the battery's SOC.

4. The energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 1, characterized in that, The specific formula for the drooping characteristic is as follows: in, and These represent the frequency and amplitude of the voltage in the photovoltaic-storage-DC-flexible distribution network, respectively. and These are active power and reactive power, respectively. , , , These are the rated frequency, voltage, active power, and reactive power, respectively. and These are the active and reactive power droop coefficients, respectively. The specific formula for controlling the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range is as follows: in, and This represents the actual power at both ends of the transmission line. and The rated power at the beginning and end of the transmission line. and Δf represents the droop coefficient at the beginning and end of the line, and Δf represents the frequency change.

5. The energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 1, characterized in that, The specific formula for the inner layer objective equation is as follows: in, For electricity load costs, , These are the weighting coefficients of the inner energy management model. The deviation between the power output of the battery in the inner energy management model and the reference value. This represents the deviation between the photovoltaic power generation in the inner energy management model and the reference value. The power output of the first battery in the first energy dispatch scheme. To provide power to the batteries in the inner energy management model, The first photovoltaic power generation capacity in the first energy dispatch scheme. This refers to the photovoltaic power generation in the inner energy management model.

6. The energy dispatching method for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 1, characterized in that, Before inputting the load forecast data and photovoltaic power generation data into the preset outer energy management model, the flexible loads in the load forecast data are divided into shiftable loads, reduceable loads, and interruptible loads according to different load characteristics.

7. An energy dispatching system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting, characterized in that, It includes an acquisition module, an outer energy management module, an inner energy management module, and a scheduling module; The acquisition module is used to acquire load forecast data and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network within the current scheduling cycle. The outer energy management module is used to input the load forecast data and photovoltaic power generation data into the preset outer energy management model, so that the outer energy management model generates a first energy dispatch scheme according to the outer objective equation and decision variables, updates the decision variables according to the first energy dispatch scheme, and then transmits the decision variables to the preset inner energy management model. The outer energy management model is constructed based on a Min-max robust model predictive controller, including: establishing the outer objective equation with the goal of minimizing operating costs, combined with the Min-max robust model predictive controller, wherein the operating costs include power loss costs, battery aging costs, and user comfort costs; and establishing multiple constraint equations based on the power supply structure, load forecast data, and photovoltaic power generation data of the photovoltaic-storage-DC-flexible distribution network, wherein the constraint equations include power constraint equations, battery differential equations, and battery SOC limiting equations. The inner energy management module is used to input the first energy scheduling scheme into the inner energy management model, so that the inner energy management model optimizes the first energy scheduling scheme according to the inner objective equation and the decision variables, generates a second energy scheduling scheme, updates the decision variables according to the second energy scheduling scheme, and then passes the decision variables to the outer energy management model. The inner-layer energy management model is constructed based on a droop control strategy, including: constructing the droop control strategy according to the droop characteristics of the photovoltaic-storage-DC-flexible distribution network, the droop control strategy being used to control the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range; using the first battery power output and the first photovoltaic power output in the first energy dispatch scheme as reference values, and combining load forecasting data and photovoltaic power output data, constructing the inner-layer objective equation with the optimization objective of minimizing load forecasting error and photovoltaic fluctuations; the constraint equations of the inner-layer energy management model are the same as those of the outer-layer energy management model; The scheduling module is used to perform energy scheduling on the photovoltaic-storage-DC-flexible power distribution network within the current scheduling cycle according to the second energy scheduling scheme.

8. The energy dispatching system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 7, characterized in that, The specific formula for the outer objective equation is as follows: in, For power loss costs, Due to the cost of battery aging, Costs related to user comfort; The unit cost of photovoltaic power generation, Photovoltaic power generation capacity, The unit cost of battery discharge, This refers to the battery discharge power. One scheduling cycle; for The aging cost of a battery over a time interval. for The depth of charge and discharge of the battery within the time interval. P(t) represents the average cost per unit of energy lost during a single charge / discharge event, and P(t) represents the charge / discharge power of the battery.

9. The energy dispatching system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 7, characterized in that, The power constraint equations include power balance constraint equations and power imbalance constraint equations. The power balance constraint equations consist of flexible load power, power transmitted from the distribution network, power generated by the battery, and photovoltaic power, constraining the sum of the power transmitted from the distribution network, the power generated by the battery, and the photovoltaic power to equal the flexible load power. The power imbalance constraint equations consist of upper and lower limits for the power transmitted from the distribution network and the power generated by the battery, constraining the power transmitted from the distribution network and the power generated by the energy storage device to be within a preset range. The battery differential equation constrains the battery energy at the current moment by using the battery energy at the previous moment, the battery charging and discharging efficiency, and the battery power output at the current moment. The battery SOC constraint equation constrains the ratio of the battery's rated capacity to its actual capacity at the current moment by setting upper and lower limits for the battery's SOC.

10. The energy dispatching system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 7, characterized in that, The specific formula for the drooping characteristic is as follows: in, and These represent the frequency and amplitude of the voltage in the photovoltaic-storage-DC-flexible distribution network, respectively. and These are active power and reactive power, respectively. , , , These are the rated frequency, voltage, active power, and reactive power, respectively. and These are the active and reactive power droop coefficients, respectively. The specific formula for controlling the voltage level of the photovoltaic-storage-DC-flexible distribution network within a preset range is as follows: in, and This represents the actual power at both ends of the transmission line. and The rated power at the beginning and end of the transmission line. and Δf represents the droop coefficient at the beginning and end of the line, and Δf represents the frequency change.

11. The energy dispatching system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 7, characterized in that, The specific formula for the inner layer objective equation is as follows: in, For electricity load costs, , These are the weighting coefficients of the inner energy management model. The deviation between the power output of the battery in the inner energy management model and the reference value. This represents the deviation between the photovoltaic power generation in the inner energy management model and the reference value. The power output of the first battery in the first energy dispatch scheme. To provide power to the batteries in the inner energy management model, The first photovoltaic power generation capacity in the first energy dispatch scheme. This refers to the photovoltaic power generation in the inner energy management model.

12. The energy dispatching system for a photovoltaic-storage-DC-flexible distribution network based on load forecasting as described in claim 7, characterized in that, Before the outer energy management module inputs the load forecast data and photovoltaic power generation data into the preset outer energy management model, the flexible loads in the load forecast data are divided into shiftable loads, reduceable loads, and interruptible loads according to different load characteristics.

Images