Intelligent energy full-process management and control platform fusing AI decision

Abstract

The invention provides a smart energy full-process management and control platform fused with AI decision, and the platform comprises an energy data collection module which is used for collecting energy full-process data through an Internet of Things device, and integrating meteorological data, market data and user behavior data at the same time; the AI decision center module is used for training a neural network model through historical energy full-process data, dynamically predicting energy load fluctuation, new energy output and market price trend in a future time period, constructing a multi-agent game framework, training a strategy network through a near-end strategy optimization algorithm, generating a dynamic scheduling strategy of source network load storage, and sending the dynamic scheduling strategy to a cloud server; establishing an energy virtual model, simulating the actual operation state of energy in real time, verifying the feasibility of the dynamic scheduling strategy of the source network load storage, and early warning potential risks in advance; and the multi-energy collaborative module is used for performing multi-energy cascade collaborative utilization through the combined cooling heating and power system, simulating a scheduling scheme execution effect in a virtual environment, discovering potential risks in advance, and avoiding economic loss in actual operation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy management, and particularly relates to a smart energy whole-process management and control platform fusing AI decision. BACKGROUND

[0002] Energy management refers to monitoring and optimizing the production, distribution and use of energy through a series of technical means and management strategies to improve energy efficiency, reduce energy consumption and lower energy costs, while reducing the impact on the environment. Energy management technology is not only applicable to industrial enterprises, but also widely used in commercial buildings, public facilities and home energy management. Through the comprehensive application of these technologies, energy saving and emission reduction, cost reduction and energy utilization efficiency can be effectively realized.

[0003] The traditional energy management adopts a fixed threshold or empirical rule based on a fixed scheduling plan of historical load, which cannot respond to new energy fluctuations, market price changes and user demand adjustments in real time. Since the output of renewable energy such as wind power and photovoltaic is significantly affected by weather conditions, the power grid load fluctuates frequently, making it difficult for traditional static scheduling strategies to adapt to dynamic changes.

[0004] Therefore, it is necessary to provide a new smart energy whole-process management and control platform fusing AI decision to solve the above technical problems. SUMMARY

[0005] To solve the above technical problems, the present application provides a smart energy whole-process management and control platform fusing AI decision.

[0006] The smart energy whole-process management and control platform fusing AI decision provided by the present application comprises:

[0007] An energy data acquisition module is used to acquire energy whole-process data through Internet of Things devices, integrate meteorological data, market data and user behavior data, construct a unified multi-dimensional energy data set, and perform data denoising, anomaly detection and standardization processing through edge computing nodes;

[0008] An AI decision hub module is used to train a neural network model through historical energy whole-process data, dynamically predict energy load fluctuations, new energy output and market price trends in future time periods, construct a multi-agent game framework based on the dynamically predicted energy load fluctuations, new energy output and market price trends in future time periods, take power grid safety, cost minimization and carbon emission control as outputs, train a strategy network through a proximal policy optimization algorithm, generate dynamic scheduling strategies of sources, grids, loads and storages, establish an energy virtual model, real-time simulate the actual operation state of energy, verify the feasibility of the dynamic scheduling strategies of sources, grids, loads and storages and give early warning of potential risks, and perform fault isolation, standby line switching and dynamic adjustment of operating parameters through an adaptive control algorithm;

[0009] A multi-energy coordination module is used for multi-energy cascade utilization through a combined cooling, heating and power system.

[0010] Further, the combined cooling, heating and power system performs energy cascade utilization by setting a target function in combination with energy balance constraints and energy storage capacity limits, solves optimal operation parameters through mixed integer linear programming, establishes an electric-thermal-gas coupling model, and optimizes multi-energy flow distribution.

[0011] Further, the Internet of Things device includes multi-modal sensors deployed on the energy production side, transmission side, storage side and consumption side.

[0012] Further, the establishment of the energy virtual model includes integrating multi-energy flow data of device states into the virtual model, multi-modal sensors uploading device state data in real time, and updating model parameters in real time through edge computing nodes.

[0013] Further, verifying the feasibility of the dynamic scheduling strategy includes the following steps:

[0014] Step one, inject the generated source network load storage dynamic scheduling strategy into the virtual model;

[0015] Step two, simulate transformer load rate changes and predict whether short-circuit capacity is out of limits;

[0016] Step three, verify whether the SOC fluctuation is within a safe range;

[0017] Step four, if the simulation shows that the current of a certain feeder is close to the threshold, immediately trigger an alarm and suggest adjusting the scheduling strategy;

[0018] Step five, determine whether new energy fluctuations will cause the power grid frequency deviation to exceed the limit through power flow analysis.

[0019] Further, the power flow analysis includes the following steps:

[0020] Step one, divide the power grid into multiple nodes and clearly define the electrical connection relationship of each node;

[0021] Step two, input historical output data of wind power and photovoltaic, set the rated power, response time and regulation capacity of thermal power units and energy storage systems, and define the baseline value and flexibility of electricity load;

[0022] Step three, use power flow calculation software to perform power flow simulation, minimize active power imbalance to ensure that node voltages and power flows are within a safe range.

[0023] Further, the multi-agent game framework includes:

[0024] State space: energy load, new energy output market price trend and energy storage SOC in the future time period are taken as inputs;

[0025] Action space: grid power purchase and energy storage charging and discharging power are defined as executable operations;

[0026] Reward function: a multi-objective optimization goal is constructed by weighting cost and renewable energy utilization rate.

[0027] Further, the management and control platform further comprises a chain-up module for recording carbon footprints of energy production and consumption of new energy output by using a blockchain, generating an auditable carbon asset report, predicting a carbon price fluctuation trend through an AI decision hub module, and formulating a carbon trading strategy.

[0028] Further, the carbon footprint calculation is based on an emission factor method to calculate carbon emissions corresponding to energy consumption.

[0029] Further, the predicted carbon price fluctuation trend is predicted by a time series analysis model to formulate a buy-sell strategy to reduce carbon cost.

[0030] Compared with the related art, the intelligent energy whole-process management and control platform integrated with AI decision provided by the present application has the following beneficial effects:

[0031] 1. The neural network model is trained by historical energy whole-process data, the energy load fluctuation, new energy output and market price trend in the future time period are dynamically predicted, a multi-agent game framework is constructed based on the dynamically predicted energy load fluctuation, new energy output and market price trend in the future time period, the grid safety, cost minimization and carbon emission control are taken as outputs, a strategy network is trained by a proximal policy optimization algorithm, a dynamic scheduling strategy of source-grid-load-storage is generated, an energy virtual model is established, the real-time energy actual operation state is simulated, the feasibility of the dynamic scheduling strategy of source-grid-load-storage is verified and potential risks are warned in advance, fault isolation, standby line switching and dynamic adjustment of operation parameters are performed through an adaptive control algorithm, the scheduling scheme execution effect is simulated in a virtual environment, potential risks are found in advance, and economic losses in actual operation are avoided.

[0032] 2. The operation parameters of a combined cooling heating and power system are optimized by a mixed integer linear programming, the energy cascade utilization efficiency is improved, an electric-thermal-gas coupling model is established, multi-energy flow resources are dynamically allocated, the system redundancy capacity demand is reduced, and infrastructure investment is saved. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 The structure diagram of the intelligent energy whole-process management and control platform integrated with AI decision provided by the present application.

[0034] Figure 2 a flow chart for verifying the feasibility of the dynamic scheduling strategy provided by the present application;

[0035] Figure 3 a flow chart for the power flow analysis and judgment provided by the present application;

[0036] Figure 4 a structural diagram of the multi-agent game framework provided by the present application. DETAILED DESCRIPTION

[0037] The present application will be further described below in combination with the drawings and embodiments.

[0038] Please refer to Figure 1 , Figure 2 , Figure 3 and Figure 4 , among which, Figure 1 a structural diagram of the intelligent energy whole-process management and control platform integrating AI decision provided by the present application; Figure 2 a flow chart for verifying the feasibility of the dynamic scheduling strategy provided by the present application; Figure 3 a flow chart for the power flow analysis and judgment provided by the present application; Figure 4 a structural diagram of the multi-agent game framework provided by the present application.

[0039] Embodiment I

[0040] In the specific implementation process, as shown in Figure 1 , the intelligent energy whole-process management and control platform integrating AI decision includes:

[0041] an energy data acquisition module, configured to acquire energy whole-process data through Internet of Things devices, integrate meteorological data, market data and user behavior data, construct a unified multi-dimensional energy data set, and perform data noise reduction, anomaly detection and standardization processing through an edge computing node, wherein the Internet of Things devices include multi-modal sensors deployed on the production side, transmission side, storage side and consumption side of energy, the meteorological data includes wind speed and light intensity, and the market data includes electricity price and carbon price;

[0042] The AI decision hub module is used to train a neural network model through historical energy whole-process data, dynamically predict energy load fluctuations, new energy output and market price trends in a future time period, construct a multi-agent game framework based on the dynamically predicted energy load fluctuations, new energy output and market price trends in the future time period, take grid safety, cost minimization and carbon emission control as outputs, train a strategy network through a proximal policy optimization algorithm to generate a dynamic scheduling strategy of source-grid-load-storage, establish an energy virtual model to simulate the actual operation state of energy in real time, verify the feasibility of the dynamic scheduling strategy of source-grid-load-storage and give early warning of potential risks, and perform fault isolation, standby line switching and dynamic adjustment of operating parameters through an adaptive control algorithm.

[0043] The parameter dynamic adjustment uses a recursive least squares method to estimate system dynamic parameters online, dynamically adjusts PID gains when an abnormality is detected, and increases an integral term to eliminate steady-state error.

[0044] The multi-energy coordination module is used for multi-energy cascade utilization. A combined energy balance constraint and energy storage capacity limit are used to perform energy cascade utilization in a cold heat and power cogeneration system through a set target function (minimizing electricity, gas and maintenance costs), and a mixed integer linear programming is used to solve optimal operating parameters, an electric-thermal-gas coupling model is established, multi-energy flow distribution is optimized, and energy balance constraints such as load and power supply matching are used.

[0045] It should be noted that the establishment of the energy virtual model includes integrating the multi-energy flow data of the device state into the virtual model, the multi-modal sensor uploads the device state data in real time, the model parameters are updated in real time through the edge computing node, and the multi-energy flow data includes energy flow data of electricity, heat, gas and cold.

[0046] It should be noted that, as shown in Figure 2 The feasibility of the dynamic scheduling strategy is verified, including the following steps:

[0047] Step 1: Inject the generated dynamic scheduling strategy of source-grid-load-storage into the virtual model.

[0048] Step 2: Simulate the transformer load rate change to predict whether the short-circuit capacity is out of limit.

[0049] Step 3: Verify whether the SOC fluctuation is within a safe range.

[0050] Step 4: If the simulation shows that the current of a certain feeder is close to the threshold, an alarm is triggered immediately and the scheduling strategy is recommended to be adjusted.

[0051] Step 5: Through power flow analysis, it is judged whether new energy fluctuations will cause the grid frequency deviation to be out of limit.

[0052] The real-time risk warning mechanism includes:

[0053] Draw control charts for key parameters (such as fan vibration amplitude, battery internal resistance), trigger early warning when exceeding upper and lower limits;

[0054] Use random forest to classify equipment status and predict potential failures;

[0055] Establish multi-level early warning, primary warning parameter slight anomaly, suggest increasing inspection frequency, senior warning parameter mutation, automatically switch to standby line and stop high-risk operation.

[0056] It should be further pointed out that, as shown in Figure 3 The power flow analysis includes the following steps:

[0057] Step 1: Divide the power grid into power generation nodes, load nodes and new energy access points, and clarify the electrical connection relationship of each node;

[0058] Step 2: Input the historical output data of wind power and photovoltaic, set the rated power, response time and regulation capacity of thermal power units and energy storage systems, define the baseline value and flexibility of electricity load;

[0059] Step 3: Use power flow calculation software for power flow simulation, minimize active power imbalance to ensure that each node voltage and power flow is within a safe range;

[0060] PSS / E or MATLAB Simulin power flow software can be used for power flow simulation;

[0061] Among them, the active power imbalance is:

[0062] ΔP(t) = P gen (t) + P renewable (t) - P load (t)

[0063] If ΔP(t)≠0, there is power imbalance in the system, which needs to be compensated by adjusting resources (such as energy storage, frequency modulation units).

[0064] Frequency deviation calculation:

[0065] The relationship between power grid frequency deviation Δf and ΔP is determined by the inertia constant H:

[0066]

[0067] Where f0=50Hz, Δt is the time step, if ΔP is compensated by frequency modulation resources (such as energy storage, thermal power), the frequency deviation decreases.

[0068] It should be noted that, as shown in Figure 4 The multi-agent game framework includes:

[0069] State space: energy load, new energy output market price trend and energy storage SOC in the future time period are taken as inputs;

[0070] Action space: grid power purchase and energy storage charging and discharging power are defined as executable operations;

[0071] Reward function: a multi-objective optimization target is constructed by the weighted cost of electricity price multiplied by power purchase and the renewable energy utilization rate of load and new energy output matching degree.

[0072] Embodiment two

[0073] In one specific implementation process, referring to Figure 1 As shown in the figure, the management and control platform further includes a chain-up module for recording the carbon footprint of new energy output energy production and consumption using blockchain to generate an auditable carbon asset report, predicting carbon price fluctuation trends through an AI decision hub module, and developing carbon trading strategies;

[0074] Carbon footprint calculation is based on emission factor method to calculate the carbon emissions corresponding to energy consumption. The emission factor method calculates the carbon emissions of each link by multiplying the energy consumption and the corresponding carbon emission factor.

[0075] Predicting carbon price fluctuation trends involves predicting carbon prices through time series analysis models and developing buy / sell strategies to reduce carbon costs.

[0076] The computing device according to the embodiments of the present application for implementing the above method includes a processor and a memory;

[0077] The processor can be a multi-core processor or can include multiple processors. In some embodiments, the processor can include a general-purpose main processor and one or more special-purpose coprocessors, such as a graphics processing unit (GPU), a digital signal processor (DSP), etc. In some embodiments, the processor can be implemented using custom circuits, such as an application-specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

[0078] The memory can include various types of storage units, such as a system memory, a read-only memory (ROM), and a permanent storage device. Among them, the ROM can store static data or instructions required by the processor or other modules of the computer. The permanent storage device can be a read and write storage device. The permanent storage device can be a non-volatile storage device that does not lose stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device uses a mass storage device (such as a magnetic or optical disk, flash memory) as a permanent storage device. In some other embodiments, the permanent storage device can be a removable storage device (such as a floppy disk, an optical drive). The system memory can be a read and write storage device or a volatile read and write storage device, such as a dynamic random access memory. The system memory can store some or all of the instructions and data required by the processor during runtime. In addition, the memory can include a combination of any computer readable storage media, including various types of semiconductor storage chips (DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), magnetic disks and / or optical disks. In some embodiments, the memory can include a read and / or write removable storage device, such as a compact disc (CD), a read-only digital versatile disc (such as DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (such as an SD card, a min SD card, a Micro-SD card, etc.), a magnetic floppy disk, etc. The computer readable storage medium does not include carrier waves and transient electronic signals transmitted through wireless or wired transmission.

[0079] It should be understood that, unless otherwise explicitly stated herein, the execution of the above steps is not strictly limited in order, and the steps can be executed in other orders. Moreover, at least part of the steps in the processes involved in the above embodiments can include multiple steps or multiple stages, which do not necessarily be executed at the same time, but can be executed at different times, and the execution order of the steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0080] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application, and any reference signs in the claims should not be regarded as limiting the claims involved.

[0081] Furthermore, it should be understood that although the description is made according to the embodiments, not every embodiment includes only one independent technical solution, and the description of the specification is only for the sake of clarity, and the skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be combined appropriately to form other embodiments that can be understood by the skilled in the art.

Claims

1. A smart energy whole-process management and control platform integrating AI decision, characterized in that, The application relates to a multi-energy collaborative system based on an energy data collection module, an AI decision center module and a multi-energy collaborative module. The energy data collection module is used for collecting energy whole-process data through an Internet of Things device, integrating meteorological data, market data and user behavior data, constructing a unified multi-dimensional energy data set, and performing data noise reduction, abnormality detection and standardization processing through an edge computing node. The AI decision center module is used for training a neural network model through historical energy whole-process data, dynamically predicting energy load fluctuation, new energy output and market price trends in a future time period, constructing a multi-agent game framework based on the dynamically predicted energy load fluctuation, new energy output and market price trends in the future time period, taking grid safety, cost minimization and carbon emission control as outputs, training a strategy network through a proximal policy optimization algorithm, generating a dynamic scheduling strategy of a source-grid-load-storage system, establishing an energy virtual model, real-time simulating an energy actual operation state, verifying the feasibility of the dynamic scheduling strategy of the source-grid-load-storage system and early warning potential risks, and performing fault isolation, standby line switching and dynamic adjustment of operation parameters through an adaptive control algorithm. The multi-energy collaborative module is used for multi-energy cascade collaborative utilization through a combined cooling heat and power (CCHP) system. 2.The smart energy whole-process management and control platform of fusion AI decision according to claim 1, characterized in that, The CCHP system performs energy cascade collaborative utilization through a target function, energy balance constraints and energy storage capacity limits, solves optimal operation parameters through a mixed integer linear programming, establishes an electric-thermal-gas coupling model, and optimizes multi-energy flow distribution. 3.The smart energy whole-process management and control platform with fused AI decision according to claim 2, characterized in that, The Internet of Things device comprises multi-modal sensors arranged on an energy production side, a transmission side, a storage side and a consumption side. 4.The smart energy whole-process management and control platform with fused AI decision according to claim 3, characterized in that, The establishment of the energy virtual model comprises integrating multi-energy flow data of device states into the virtual model, real-time uploading of device state data by the multi-modal sensors, and real-time updating of model parameters by the edge computing node. 5.The smart energy whole-process management and control platform with fused AI decision according to claim 4, characterized in that, The verification of the feasibility of the dynamic scheduling strategy comprises the following steps: Step one: injecting the generated dynamic scheduling strategy of the source-grid-load-storage system into the virtual model; Step two: simulating transformer load rate change and predicting whether short-circuit capacity is out of limit; Step three: verifying whether SOC fluctuation is within a safe range; Step four: if simulation shows that the current of a certain feeder is close to a threshold value, an alarm is triggered and a scheduling strategy is suggested to be adjusted; Step five: judging whether new energy fluctuation will cause grid frequency deviation out of limit through power flow analysis. 6.The smart energy whole-process management and control platform with fused AI decision according to claim 5, characterized in that, The power flow analysis judgment comprises the following steps: Step one: dividing the grid into multiple nodes and clearly defining the electrical connection relationship of the nodes; Step two: inputting historical output data of wind power and photovoltaic power, setting rated power, response time and regulation capacity of thermal power units and energy storage systems, and defining baseline value and elasticity of power load; Step three: performing power flow simulation through power flow calculation software, and minimizing active power imbalance to ensure that node voltage and power flow are within a safe range. 7.The smart energy whole-process management and control platform with fused AI decision according to claim 6, characterized in that, The multi-agent game framework comprises: a state space: integrating energy load, new energy output market price trend and energy storage SOC in a future time period as input; an action space: defining grid power purchase power and energy storage charging and discharging power as executable operations; a reward function: constructing a multi-objective optimization target through weighted cost and renewable energy utilization rate. 8.The smart energy whole-process management and control platform with fused AI decision according to claim 1, characterized in that, The management platform further comprises a chain-up module for recording carbon footprints of energy production and consumption of new energy output by using a block chain, generating an auditable carbon asset report, predicting a carbon price fluctuation trend through an AI decision hub module, and formulating a carbon trading strategy. 9.The smart energy whole-process management and control platform with fused AI decision according to claim 8, characterized in that, The carbon footprint calculation is based on an emission factor method for calculating carbon emissions corresponding to energy consumption. The emission factor method calculates carbon emissions at each link by multiplying energy consumption and the corresponding carbon emission factor. 10.The smart energy whole-process management and control platform with fused AI decision according to claim 8, characterized in that, The predicted carbon price fluctuation trend is predicted by a time series analysis model to predict the carbon price and formulate a buy-sell strategy to reduce the carbon cost.

Images