Distributed energy management method and system based on transformer area energy storage

Abstract

The invention discloses a transformer area energy storage-based distributed energy management method and system, and the method comprises the steps: analyzing distributed energy operation data through a prediction model preset in a cloud master station, and generating transformer area load prediction; according to the method, the real-time distributed energy operation data collected by the area edge computing terminal is combined to dynamically adjust the energy management of the area, hierarchical control is enhanced, energy balance analysis and safe operation are realized, the new energy consumption rate of the area is improved, and the method is a multi-level and self-adaptive distributed energy management method.

Description

Technical Field

[0001] The present invention relates to the field of power distribution automation, and in particular to a distributed energy management method and system based on substation energy storage. Background Art

[0002] In recent years, the growth rate of distributed energy has been extremely rapid, and many hidden dangers of existing energy management technologies in distributed energy management have also been exposed. Most existing technologies are single-layer regulation, which can easily lead to uneven regional distribution, low-voltage dominance, source-load mismatch and other hidden dangers, which in turn cause problems such as multi-level reverse transmission and reverse overload. The management of distributed energy in the substation area lacks pertinence, which brings a series of negative impacts on the safe use of energy.

[0003] Therefore, there is an urgent need for a highly targeted multi-layer distributed energy management method. Summary of the invention

[0004] Purpose of the invention: The present invention provides a distributed energy management method and system based on substation energy storage, in order to solve the problem that the existing single-layer energy management method has insufficient control over distributed energy, weak pertinence, and cannot meet the needs of safe energy use.

[0005] Technical solution: The present invention discloses a distributed energy management method based on regional energy storage, comprising the following steps:

[0006] The mobile terminal imports the basic information of distributed energy, equipment information and substation topology information into the IoT platform for storage, and the cloud master station uses the information stored in the IoT platform to establish a prediction model;

[0007] Based on the cloud-edge interaction protocol, the edge computing terminal collects distributed energy operation data and sends it to the IoT platform;

[0008] The cloud master station uses the prediction model to predict the load of the substation based on the received distributed energy operation data, obtains the prediction results, and sends the prediction results to the edge computing terminal;

[0009] The edge computing terminal receives the prediction results sent by the cloud master station and applies them as operation strategies to the distributed energy management of the substation.

[0010] Furthermore, the distributed energy basic information includes at least one of the whole site model file and the cloud-edge interaction configuration file; the device information includes at least one of the device name, ESN, manufacturer, and model of the sub-device.

[0011] Furthermore, through a handheld mobile terminal, the communication file containing the basic information of distributed energy, equipment information and substation topology information is imported into the edge computing terminal, and then sent to the IoT platform after being parsed by the edge computing terminal.

[0012] Furthermore, after receiving the uploaded device information, the IoT platform registers the distributed energy collection terminal to improve the substation topology information.

[0013] The distributed energy operation data includes at least one of the following: energy storage cost, interaction cost with other substations, load revenue, total working time, output power at different times, photovoltaic output power, reducible load working power, shiftable load and status, shiftable load rated working power, uncontrollable load working power, output power, charging and discharging power, and charge state.

[0014] Furthermore, the cloud master station further trains the prediction model using a neural network algorithm based on the distributed energy operation data sent to the cloud master station by the edge computing terminal;

[0015] The energy optimization goal is based on a time scale of 1 hour, with the goal of maximizing the economy and energy efficiency of the substation within the dispatch day, expressed as:

[0016]

[0017] Among them, f i (P i b ) is the energy storage cost, f2(P i pcc ) is the interaction cost with other areas, f3(P i load ) is the load benefit, which is 0 here; P i b is the output power of the energy storage at time i, P i load1 is the working power that can be reduced at load i; i is the working state of the movable load at time i, 0 means working and 1 means non-working; P i PV is the photovoltaic output power, P i load2 is the rated working power of the translatable load, P i load3 is the working power of the uncontrollable load, N is the number of time scales, W is the cost of the uncontrollable load, F1 is the total energy storage cost, F2 is the total cost of interaction with other substations, F3 is the total load benefit, and F is the cost of the controllable load. The control quantity u is obtained i =[P i b ,P i load1 ,δ i ].

[0018] Furthermore, the constraint condition is substituted into the control quantity u i, input the prediction model to obtain the distributed energy prediction result; the control quantity u i The constraints include:

[0019] Translational load constraint δ i :

[0020] δ i+1 -δ i ≤δ τ τ=i+2,i+3,...,i+T i up -1

[0021] Among them, T i up The total working time of the translatable load i;

[0022] Reducible load constraint P i load1 :

[0023] P min ≤P i load1 ≤P max

[0024] Among them, P min , P max is the minimum and maximum output power that can reduce the load;

[0025] Energy storage constraint P i b :

[0026] -P cmax ≤P i b ≤P dcmax

[0027] SOC min ≤SOC i ≤SOC max

[0028] SOC last =SOC first

[0029] Among them, P cmax , P dcmax They are the maximum charging and discharging power of energy storage; SOC min , SOC max They are the minimum and maximum state of charge respectively; SOC first , SOC last are the initial and final states of charge of energy storage, respectively.

[0030] The present invention also discloses a distributed energy management system based on regional energy storage, comprising:

[0031] Edge computing terminals are used to collect distributed energy operation data and send it to the cloud master station; receive the prediction results sent by the cloud master station and apply them to distributed energy management;

[0032] IoT platform, used to store basic information of distributed energy, equipment information and area topology information; connected with edge computing terminals and cloud master stations;

[0033] The cloud master station is used to call the information stored in the IoT platform to enter the prediction model; based on the distributed energy operation data sent by the edge computing terminal, the prediction model is used to predict the load of the substation, obtain the prediction results, and send the prediction results to the edge computing terminal.

[0034] Furthermore, the distributed energy basic information includes at least one of the whole site model file and the cloud-edge interaction configuration file; the device information includes at least one of the device name, ESN, manufacturer, and model of the sub-device.

[0035] Furthermore, through a handheld mobile terminal, the communication file containing the basic information of distributed energy, equipment information and substation topology information is imported into the edge computing terminal, and then sent to the IoT platform after being parsed by the edge computing terminal.

[0036] Furthermore, after receiving the uploaded device information, the IoT platform registers the edge computing terminal to improve the area topology information.

[0037] Furthermore, the distributed energy operation source data includes at least one of the following: energy storage cost, interaction cost with other substations, load benefit, total working time, output power at different times, photovoltaic output power, reducible load working power, shiftable load and status, shiftable load rated working power, uncontrollable load working power, output power, charging and discharging power, and charge state.

[0038] Furthermore, the prediction model generates prediction results according to the energy optimization target. The energy optimization target takes 1 hour as the time scale and aims to maximize the economy and energy efficiency of the substation within the dispatch day, which is expressed as:

[0039]

[0040] Among them, f1(P i b ) is the energy storage cost, f2(P i pcc ) is the interaction cost with other areas, f3(P i load ) is the load benefit, which is 0 here; P i b is the output power of the energy storage at time i, P iload1 is the working power that can be reduced at load i; i is the working state of the movable load at time i, 0 means working and 1 means non-working; P i PV is the photovoltaic output power, P i load2 is the rated working power of the translatable load, P i load3 is the working power of uncontrollable load, N is the number of time scales, W is the cost of uncontrollable load, F1 is the total energy storage cost, F2 is the total cost of interaction with other substations, F3 is the total load benefit, F is the cost of controllable load, and the control amount u i =[P i b ,P i load1 ,δ i ].

[0041] Furthermore, the cloud master uses a neural network algorithm to train the prediction model based on the distributed energy operation data sent by the edge computing terminal.

[0042] Furthermore, the constraint condition is substituted into the control quantity u i , and input the control amount into the prediction model to obtain the distributed energy prediction result; the constraint conditions include:

[0043] Translational load constraint δ i :

[0044] δ i+1 -δ i ≤δ τ τ=i+2,i+3,...,i+T i up -1

[0045] Among them, T i up The total working time of the translatable load i;

[0046] Reducible load constraint P i load1 :

[0047] P min ≤P i load1 ≤P max

[0048] Among them, P min , P max is the minimum and maximum output power that can reduce the load;

[0049] Energy storage constraint P ib :

[0050] -P cmax ≤P i b ≤P dcmax

[0051] SOC min ≤SOC i ≤SOC max

[0052] SOC last =SOC first

[0053] Among them, P cmax , P dcmax They are the maximum charging and discharging power of energy storage; SOC min , SOC max They are the minimum and maximum state of charge respectively; SOC first , SOC last are the initial and final states of charge of energy storage, respectively.

[0054] Beneficial effects: The present invention discloses a distributed energy management method and system for substations based on cloud-edge collaborative technology. The method analyzes distributed energy operation data through a prediction model preset in the cloud master station to generate a substation load forecast. The method strengthens hierarchical control, realizes energy balance analysis and safe operation, and improves the new energy consumption rate of the substation. It is a highly targeted multi-layer distributed energy management method for substations. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a schematic diagram of the principle of the present invention;

[0056] Figure 2 The figure is a flow chart of the interaction method of the present invention. DETAILED DESCRIPTION

[0057] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent forms of modifications to the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0058] like Figure 1 , Figure 2 As shown, a distributed energy management method and system for a substation based on cloud-edge collaboration technology in this embodiment. The method of this embodiment specifically includes the following steps:

[0059] The establishment of the substation topology information and distributed energy basic information and model. The substation topology information and distributed energy basic information include: the whole station model file, cloud-edge-end interaction configuration file, and device information, substation topology information imported into at least one of the IoT platform; the cloud master station calls the information stored in the IoT platform to enter the device model and at least one of the device name, ESN, manufacturer, and model of the sub-device.

[0060] Specifically, in this embodiment, the staff uses a handheld mobile maintenance terminal to pre-set data models and substation topology information to the cloud master station and the substation edge computing terminal. The edge computing terminal is regarded as an end device. The staff transcribes the information of the distribution terminal and the low-voltage equipment, registers the distribution terminal and the low-voltage equipment on the PMS, imports the device model on the cloud master station and the IoT platform, generates the platform access information of the distribution terminal, and imports the platform access information into the distribution terminal, and adapts the communication protocol between the end device access App and the end device. The communication file containing the substation topology information and the basic information of distributed energy is imported into the substation smart terminal through a handheld mobile terminal, and the cloud-edge interaction APP reads and parses the file. After the edge device goes online on the cloud, the IoT platform sends the model file and the communication file to the cloud-edge interaction APP of the substation smart terminal, and the cloud-edge interaction APP reads and parses the file.

[0061] Plug-and-play access for distributed energy. According to the adapted terminal device protocol, the terminal device access APP senses the downstream terminal device, automatically identifies the terminal device identity, and registers it with the cloud-edge interaction APP; the cloud master station records the uploaded terminal device registration information based on the topology information and station area information, and improves the station area topology information.

[0062] Based on the cloud-edge interaction protocol, the edge computing terminal collects distributed energy operation data and sends it to the cloud master through a standardized interface. The edge terminals use periodic active reporting or protocol response methods. The collection APP obtains the terminal device data and sends it to the data center backup of the edge computing terminal. The cloud-edge interaction APP periodically obtains data from the data center based on the registered devices and reports it to the cloud master.

[0063] Distributed energy operation data includes at least one of the following: energy storage cost, interaction cost with other substations, load revenue, total working time, output power at different times, photovoltaic output power, reduceable load working power, shiftable load and status, shiftable load rated working power, uncontrollable load working power, output power, charging and discharging power, charge state, etc.

[0064] Based on the received distributed energy operation data, the cloud master station predicts the load of the substation and sends the prediction results to the edge computing terminal.

[0065] The following describes how to perform load forecasting in conjunction with an embodiment. Considering the day-ahead power generation plan, with a time scale of 1 hour, comprehensively considering the power situation of renewable energy in the distribution network, the short-term load forecast of power generation, and the real-time electricity price, with the goal of maximizing the economy and energy efficiency of the distribution network operation within the dispatching day, the distributed generation, various loads, and energy storage devices in the distribution network are optimally dispatched to obtain a translatable load power consumption plan, that is, the changes in the energy storage SOC throughout the day.

[0066] The power generation optimization target is based on a time scale of 1 hour, with the goal of maximizing the economy and energy efficiency of the substation within the dispatch day, expressed as:

[0067]

[0068] Among them, f i (P i b ) is the energy storage cost, f2(P i pcc ) is the interaction cost with other areas, f3(P i load ) is the load benefit, which is 0 here; P i b is the output power of the energy storage at time i, P i load1 is the working power that can be reduced at load i; i is the working state of the movable load at time i, 0 means working and 1 means non-working; P i PV is the photovoltaic output power, P i load2 is the rated working power of the translatable load, P i load3 is the working power of the uncontrollable load, N is the number of time scales, W is the cost of the uncontrollable load, F1 is the total energy storage cost, F2 is the total cost of interaction with other substations, F3 is the total load benefit, and F is the cost of the controllable load. The control quantity u is obtained i =[P i b ,P i load1 ,δ i ].

[0069] Substitute the constraints into the control quantity u i , input the prediction model of the cloud master station to obtain the prediction result; the constraints include:

[0070] Translational load constraint δ i :

[0071] δ i+1 -δ i ≤δτ τ=i+2,i+3,...,i+T i up -1

[0072] Among them, T i up The total working time of the translatable load i;

[0073] Reducible load constraint P i load1 :

[0074] P min ≤P i load1 ≤P max

[0075] Among them, P min , P max is the minimum and maximum output power that can reduce the load;

[0076] Energy storage constraint P i b :

[0077] -P cmax ≤P i b ≤P dcmax

[0078] SOC min ≤SOC i ≤SOC max

[0079] SOC last =SOC first

[0080] Among them, P cmax , P dcmax They are the maximum charging and discharging power of energy storage; SOC min , SOC max They are the minimum and maximum state of charge respectively; SOC first , SOC last are the initial and final states of charge of energy storage, respectively.

[0081] Load forecasting can also be expanded to divide forecasting units according to different functions. Based on historical load, date type, meteorological factors, distributed photovoltaics, maintenance plan and other data, corresponding forecasting models can be established through similar day method, time series method and various artificial intelligence algorithms to realize load forecasting in the substation.

[0082] The edge computing terminal receives the prediction results sent by the cloud master station and applies them as operation strategies to the distributed energy management of the substation.

[0083] By using the prediction results, the power generation of distributed energy in the substation can be predicted and the power generation plan can be formulated. Furthermore, it is also possible to combine the prediction results issued by the cloud master station with the distributed energy power data in the substation obtained by real-time monitoring of the edge computing terminal, and through the coordinated regulation and flexible control of the distributed photovoltaic, energy storage, charging piles and other adjustable resources in the substation, the prediction results issued by the cloud master station are used as constraints to formulate the substation's own energy adaptive management strategy, ensure the balanced operation of distributed resources in the substation, and respond to the adjustment of the substation energy management strategy in real time.

[0084] The present invention also discloses a distributed energy management system based on regional energy storage, the system comprising:

[0085] Edge computing terminals are used to collect distributed energy operation data and send it to the cloud master station; the prediction results sent by the cloud master station are applied to distributed energy management;

[0086] IoT platform, used to store basic information of distributed energy, equipment information and area topology information; connected with edge computing terminals and cloud master stations;

[0087] The cloud master station is used to call the information stored in the IoT platform to enter the prediction model; based on the received distributed energy operation data, the prediction model is used to predict the load of the substation, obtain the prediction results, and send the prediction results to the edge computing terminal.

[0088] The basic information of distributed energy includes at least one of the whole station model file and the cloud-edge interaction configuration file; the device information includes at least one of the device name, ESN, manufacturer, and model of the sub-device. In an embodiment, a communication file containing the basic information of distributed energy, device information, and substation topology information is imported into the edge computing terminal through a handheld mobile terminal, and is parsed by the edge computing terminal and uploaded to the Internet of Things platform. After receiving the uploaded device information, the Internet of Things platform records the edge computing terminal to improve the substation topology information.

[0089] Distributed energy operation source data include at least one of the following: energy storage cost, interaction cost with other substations, load revenue, total working time, output power at different times, photovoltaic output power, reduceable load working power, shiftable load and status, shiftable load rated working power, uncontrollable load working power, output power, charging and discharging power, and charge state.

[0090] The prediction model is used to predict the load of the substation, obtain the prediction results, and analyze the energy optimization goals of distributed energy; the cloud master station uses the prediction model combined with the energy optimization goal to predict the load of the substation and obtain the prediction results. The energy optimization goal is based on a time scale of 1 hour, with the highest economy and energy efficiency of the substation within the dispatch day as the goal, expressed as,

[0091]

[0092] Among them, f1(P i b ) is the energy storage cost, f2(P i pcc ) is the interaction cost with other areas, f3(P i load ) is the load benefit, which is 0 here; P i b is the output power of the energy storage at time i, P i load1 is the working power that can be reduced at load i; i is the working state of the movable load at time i, 0 means working and 1 means non-working; P i PV is the photovoltaic output power, P i load2 is the rated working power of the translatable load, P i load3 is the working power of uncontrollable load, N is the number of time scales, W is the cost of uncontrollable load, F1 is the total energy storage cost, F2 is the total cost of interaction with other substations, F3 is the total load benefit, F is the cost of controllable load, and the control amount u i =[P i b ,P i load1 ,δ i ].

[0093] The cloud master uses the neural network algorithm to train the prediction model based on the distributed energy operation data sent to the cloud master by the edge computing terminal. Substitute the constraints into the control quantity u i , and input the control amount into the prediction model to obtain the distributed energy prediction result; the constraint conditions include:

[0094] Translational load constraint δ i :

[0095] δ i+1 -δ i ≤δ τ τ=i+2,i+3,...,i+T i up -1

[0096] Among them, T i up The total working time of the translatable load i;

[0097] Reducible load constraint P i load1 :

[0098] P min ≤P i load1 ≤P max

[0099] Among them, P min , P max is the minimum and maximum output power that can reduce the load;

[0100] Energy storage constraint P i b :

[0101] -P cmax ≤P i b ≤P dcmax

[0102] SOC min ≤SOC i ≤SOC max

[0103] SOC last =SOC first

[0104] Among them, P cmax , P dcmax They are the maximum charging and discharging power of energy storage; SOC min , SOC max They are the minimum and maximum state of charge respectively; SOC first , SOC last are the initial and final states of charge of energy storage, respectively.

[0105] The present invention analyzes distributed energy operation data through the prediction model preset in the cloud master station, generates the load forecast of the substation, and then makes targeted adjustments to the energy management of the substation based on the prediction results. This method strengthens the hierarchical control, realizes energy balance analysis and safe operation, and improves the new energy consumption rate of the substation. It is a highly targeted multi-layer distributed energy management method.

Claims

1. A distributed energy management method based on regional energy storage, characterized in that: The following steps are involved: The mobile terminal imports the basic information of distributed energy, equipment information and substation topology information into the IoT platform for storage, and the cloud master station uses the information stored in the IoT platform to establish a prediction model; Based on the cloud-edge interaction protocol, the edge computing terminal collects distributed energy operation data and sends it to the cloud master station; The cloud master station uses the prediction model to predict the load of the substation based on the received distributed energy operation data, obtains the prediction results, and sends them to the edge computing terminal; The edge computing terminal receives the prediction results sent by the cloud master and applies them to distributed energy management as an operation strategy.

2. The distributed energy management method according to claim 1, characterized in that: The distributed energy basic information includes at least one of the whole site model file and the cloud-edge-end interaction configuration file; the device information includes at least one of the device name, ESN, manufacturer, and model of the sub-device.

3. The distributed energy management method according to claim 1, characterized in that: The mobile terminal imports the basic information of distributed energy, equipment information and substation topology information into the Internet of Things platform for storage, including: the mobile terminal imports the communication file containing the basic information of distributed energy, equipment information and substation topology information into the edge computing terminal, and sends it to the Internet of Things platform after being parsed by the edge computing terminal.

4. The distributed energy management method according to claim 3, characterized in that: Also includes: After receiving the uploaded device information, the IoT platform will register the edge computing terminal to complete the area topology information.

5. The distributed energy management method according to claim 1, characterized in that: The distributed energy operation data includes at least one of the following: energy storage cost, interaction cost with other substations, load revenue, total working time, output power at different times, photovoltaic output power, reducible load working power, shiftable load and status, shiftable load rated working power, uncontrollable load working power, output power, charging and discharging power, and charge state.

6. The distributed energy management method according to claim 1, characterized in that: The cloud master station uses the prediction model to predict the load of the substation based on the received distributed energy operation data and obtains the prediction results, including: The cloud master station generates prediction results based on the received distributed energy operation data using the prediction model according to the energy optimization target. The energy optimization target is based on a time scale of 1 hour and aims to maximize the economy and energy efficiency of the substation within the dispatch day, which is expressed as: Among them, f1(P i b ) is the energy storage cost, f2(P i pcc ) is the interaction cost with other areas, f3(P i load ) is the load benefit, which is 0 here; P i b is the output power of the energy storage at time i, P i load1 is the working power that can be reduced at load i; i is the working state of the movable load at time i, 0 means working and 1 means non-working; P i PV is the photovoltaic output power, P i load2 is the rated working power of the translatable load, P i load3 is the working power of uncontrollable load, N is the number of time scales, W is the cost of uncontrollable load, F1 is the total energy storage cost, F2 is the total cost of interaction with other substations, F3 is the total load benefit, F is the cost of controllable load, and the control amount u i =[P i b ,P i load1 ,δ i ].

7. The distributed energy management method according to claim 1 or 6, characterized in that: The cloud master station trains the prediction model using a neural network algorithm based on the distributed energy operation data sent to the cloud master station by the edge computing terminal.

8. The distributed energy management method according to claim 1 or 6, characterized in that: The cloud master station uses the prediction model to predict the load of the substation based on the received distributed energy operation data and obtains the prediction results, including: Substitute the constraints into the control quantity u i , and input the control amount into the prediction model to obtain the distributed energy prediction result; the constraint conditions include: Translational load constraint δ i : d i+1 -d i ≤δ τ τ=i+2,i+3,...,i+T i up -1 Among them, T i up The total working time of the translatable load i; Reducible load constraint P i load1 : P min ≤P i load1 ≤P max Among them, P min , P max is the minimum and maximum output power that can reduce the load; Energy storage constraint P i b : -P cmax ≤P i b ≤P dcmax SOC min ≤SOC i ≤SOC max SOCIETY last =SOC first Among them, P cmax , P dcmax They are the maximum charging and discharging power of energy storage; SOC min , SOC max They are the minimum and maximum state of charge respectively; SOC first , SOC last are the initial and final states of charge of energy storage, respectively.

9. A distributed energy management system based on regional energy storage, characterized in that: include: Edge computing terminals are used to collect distributed energy operation data and send it to the cloud master station; Receive the prediction results sent by the cloud master and apply them to distributed energy management; IoT platform, used to store basic information of distributed energy, equipment information and substation topology information; Connect with edge computing terminals and cloud master stations; The cloud master station is used to call the information stored in the IoT platform to enter the prediction model; According to the distributed energy operation data sent by the edge computing terminal, the prediction model is used to predict the load of the substation, and the prediction results are obtained and sent to the edge computing terminal.

10. The distributed energy management system according to claim 9, characterized in that: The distributed energy basic information includes at least one of the whole site model file and the cloud-edge-end interaction configuration file; the device information includes at least one of the device name, ESN, manufacturer, and model of the sub-device.

11. The distributed energy management system according to claim 9, characterized in that: The mobile terminal imports the basic information of distributed energy, equipment information and substation topology information into the Internet of Things platform for storage, including: through the handheld mobile terminal, the communication file containing the basic information of distributed energy, equipment information and substation topology information is imported into the edge computing terminal, and then sent to the Internet of Things platform after being parsed by the edge computing terminal.

12. The distributed energy management system according to claim 11, characterized in that: Also includes: After receiving the uploaded device information, the IoT platform will register the distributed energy collection terminal to improve the substation topology information.

13. The distributed energy management system according to claim 9, characterized in that: The distributed energy operation source data includes at least one of the following: energy storage cost, interaction cost with other substations, load benefit, total working time, output power at different times, photovoltaic output power, reducible load working power, shiftable load and status, shiftable load rated working power, uncontrollable load working power, output power, charging and discharging power, and charge state.

14. The distributed energy management system according to claim 9, characterized in that: The cloud master station uses the prediction model to predict the load of the substation based on the received distributed energy operation data and obtains the prediction results, including: The cloud master station generates prediction results based on the received distributed energy operation data using the prediction model according to the energy optimization target. The energy optimization target is based on a time scale of 1 hour and aims to maximize the economy and energy efficiency of the substation within the dispatch day, which is expressed as: Among them, f1(P i b ) is the energy storage cost, f2(P i pcc ) is the interaction cost with other areas, f3(P i load ) is the load benefit, which is 0 here; P i b is the output power of the energy storage at time i, P i load1 is the working power that can be reduced at load i; i is the working state of the movable load at time i, 0 means working and 1 means non-working; P i PV is the photovoltaic output power, P i load2 is the rated working power of the translatable load, P i load3 is the working power of uncontrollable load, N is the number of time scales, W is the cost of uncontrollable load, F1 is the total energy storage cost, F2 is the total cost of interaction with other substations, F3 is the total load benefit, F is the cost of controllable load, and the control amount u i =[P i b ,P i load1 ,δ i ].

15. The distributed energy management system according to claim 9 or 14, characterized in that: The cloud master station trains the prediction model using a neural network algorithm based on the distributed energy operation data sent to the cloud master station by the edge computing terminal.

16. The distributed energy management system according to claim 9 or 14, characterized in that: Substitute the constraints into the control quantity u i , and input the control quantity into the prediction model to obtain the distributed energy prediction result; the control quantity u i The constraints include: Translational load constraint δ i : d i+1 -d i ≤δ τ τ=i+2,i+3,...,i+T i up -1 Among them, T i up The total working time of the translatable load i; Reducible load constraint P i load1 : P min ≤P i load1 ≤P max Among them, P min , P max is the minimum and maximum output power that can reduce the load; Energy storage constraint P i b : -P cmax ≤P i b ≤P dcmax SOC min ≤SOC i ≤SOC max SOCIETY last =SOC first Among them, P cmax , P dcmax They are the maximum charging and discharging power of energy storage; SOC min , SOC max They are the minimum and maximum state of charge respectively; SOC first , SOC last are the initial and final states of charge of energy storage, respectively.

Images