Park integrated energy system model based on multi-energy network and multi-energy equipment

Abstract

The invention provides a park integrated energy system model based on a multi-energy network and multi-energy equipment, and relates to the technical field of integrated energy management, and the system model comprises a quantum calculation optimization module, a self-adaptive energy scheduling module, an equipment cooperation and self-healing module and a multi-level energy coupling and conversion module. A global optimal scheme of energy scheduling is realized, stable operation of the system in a complex environment is ensured by using a self-adaptive mechanism of deep reinforcement learning, the system can make adjustment and ensure stability before equipment failure through a real-time monitoring and self-healing mechanism for the health state of the equipment, and the system can flexibly convert among multiple energy forms, so that the energy scheduling efficiency is improved. Through optimal scheduling, self-adaptive adjustment and multi-energy conversion, the utilization efficiency of energy and the reliability of the system are improved to the maximum extent, meanwhile, energy waste and carbon emission are greatly reduced, and the advantages provide powerful support for comprehensive energy management of a modern intelligent park.

Description

Technical Field

[0001] The present invention relates to the technical field of integrated energy management, and in particular to a park integrated energy system model based on multi-energy networks and multi-energy devices. Background Art

[0002] According to a method for optimizing the configuration of a comprehensive energy system for a park disclosed in Chinese patent number "CN112418479A", the present invention first analyzes the cooling, heating and electricity loads of the park according to a clustering method and obtains a typical scenario; an optimization configuration model considering the cooling, heating and electricity loads is established with the goal of minimizing the annual operating cost of the comprehensive energy system for the park; the model considers multiple factors such as supply and demand energy balance, cooling / heating / electricity coupling, and equipment installation and operation, and the optimization variables are the capacity of each device and the optimized scheduling value of the equipment in a typical scenario. Example analysis shows that the method proposed in the present invention is reasonable and feasible, and has obvious economy on the basis of ensuring the safety and reliability of system operation.

[0003] The above patent documents and prior art have the following technical problems when used:

[0004] Problem 1: In the prior art of the above patent technology, most park integrated energy systems adopt traditional scheduling and management methods, relying on empirical rules and static algorithms. These methods have certain limitations when dealing with complex multivariable energy scheduling problems. In particular, when facing dynamically changing energy demand and supply, they may not be able to make quick adjustments and optimizations. When dealing with complex multivariable energy scheduling problems, the calculation efficiency is low, and it is difficult to meet the needs of real-time optimization, resulting in energy waste and insufficient utilization.

[0005] Problem 2: Energy scheduling optimization in the above documents and existing technologies usually relies on classical calculation methods such as linear programming and integer programming. These methods are computationally intensive and inefficient when dealing with complex energy system problems. Especially when the system is large in scale and has many variables, the optimization calculation time increases significantly.

[0006] Question three: The energy systems in the above-mentioned documents and existing technologies mostly rely on manual intervention to adjust equipment operation and handle faults, and lack an automated intelligent scheduling mechanism. In terms of equipment failure warning and self-healing mechanisms, existing technologies have not yet achieved comprehensive coverage, resulting in equipment failures that may affect the stability of the entire system. Summary of the invention

[0007] In view of the shortcomings of the prior art, the present invention provides a campus integrated energy system model based on multi-energy networks and multi-energy devices, which solves the problems of low computing efficiency, slow response speed and high demand for manual intervention.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a comprehensive energy system model for a park based on a multi-energy network and multi-energy equipment, the system model includes a quantum computing optimization module, an adaptive energy scheduling module, an equipment coordination and self-healing module, and a multi-level energy coupling and conversion module, wherein:

[0009] The quantum computing optimization module is responsible for handling complex energy scheduling problems and is used to find the optimal scheduling solution in a large-scale energy network. The quantum annealing algorithm is used to find the global optimal solution. The quantum annealing algorithm is represented by the Ising model:

[0010]

[0011] in,

[0012] H(σ) represents the minimum optimization energy:

[0013] σ i Indicates the status of energy equipment, including on and off, power supply and shutdown;

[0014] h i Indicates the local energy consumption of the equipment;

[0015] J ij Represents the interactions between devices, including transmission losses;

[0016] The adaptive energy scheduling module combines deep reinforcement learning (DRL) and fuzzy logic to handle uncertainty factors, dynamically adapt to changes in energy demand, perform adaptive scheduling and optimization control, and make optimal decisions. The formula for fuzzy logic to handle uncertainty is as follows:

[0017]

[0018] in;

[0019] x represents input variables, including temperature and energy demand;

[0020] μ A (x) represents fuzzy membership;

[0021] c represents the expected value of the input;

[0022] b indicates the fuzzy range;

[0023] p represents the control ambiguity;

[0024] On this basis, energy scheduling is performed through deep reinforcement learning, and the formula is as follows:

[0025]

[0026] in,

[0027] Q(s,a) represents the expected return of taking action a in state s;

[0028] r represents immediate reward;

[0029] γ represents the discount factor;

[0030] The equipment coordination and self-healing module: uses machine learning technology to perform predictive maintenance on multi-functional equipment. When an abnormality in equipment operation is detected, the system will take measures in advance to adjust the working status of other equipment and perform self-healing and adjustment of the system. The LSTM long short-term memory network is used to predict the equipment operation status. The core equation of LSTM is:

[0031] h t =o t tanh(C t )

[0032] in,

[0033] h t Indicates hidden state;

[0034] o t represents the output gate;

[0035] C t Represents a memory unit;

[0036] The multi-level energy coupling and conversion module: This module is responsible for the flexible coupling and conversion of various energy forms. Through intelligent management, it performs the conversion and utilization of different energy sources. The multi-level coupling and conversion process is described by a linear programming model:

[0037]

[0038] The constraints are:

[0039]

[0040] in,

[0041] C i Indicate the costs of different forms of energy;

[0042] x i Indicates the quantity supplied;

[0043] a ij Indicates the conversion efficiency between different energy forms;

[0044] b j Represents energy demand.

[0045] Preferably, the quantum computing optimization module uses the Ising model to seek the optimal solution by minimizing the energy function, and the solution process is as follows:

[0046] Sp1: Determine the device state. Each device i is in the on or off state, corresponding to σ i = +1 or σ i = -1;

[0047] Sp2: Define local energy consumption h i , the energy consumption of each device in different states, including the energy consumption of power equipment when it is running;

[0048] Sp3: Interactions between computing devicesJ ij , describes the interaction between energy devices, some devices working together will bring additional energy consumption, which is represented by the interaction term;

[0049] Sp4: Annealing process. Through quantum annealing, the system will continuously try to reduce the total energy H(σ) until the optimal scheduling combination is found.

[0050] Preferably, the fuzzy logic module in the adaptive energy scheduling module is used to handle unpredictable factors, including provocations, equipment load fluctuations, wherein the fuzzy logic formula further includes the following steps:

[0051] Sp1: Input fuzzification, which transforms input variables such as energy demand, including weather and temperature, into fuzzy sets through membership functions to reflect their uncertainty;

[0052] Sp2: Fuzzy reasoning: Generate fuzzy control instructions for energy equipment through fuzzy reasoning rules;

[0053] Sp3: Defuzzification: Convert fuzzy output into actual equipment operation, including start and stop control of equipment.

[0054] Preferably, the deep reinforcement learning in the adaptive energy scheduling module is used for long-term optimization and continuous learning, and the deep reinforcement learning further includes the following contents:

[0055] The state space s represents the current state of the system, including the current supply of energy equipment and the state of the equipment;

[0056] The action space a represents the actions that the system can take, including dispatching more power and shutting down specified equipment;

[0057] Reward r, instant feedback for each dispatch, including data on energy efficiency and cost reduction;

[0058] The optimal strategy is updated. Through continuous iteration, the system will learn which scheduling decisions can bring higher long-term returns. By handling uncertainty through fuzzy logic, deep reinforcement learning is continuously optimized in the long term, further enabling the system to dynamically adapt to changing energy needs.

[0059] Preferably, the device coordination and self-healing module predicts the failure trend of the device, performs scheduling adjustments before the failure occurs, and maintains system stability. This is achieved through the LSTM formula, and the prediction steps are as follows:

[0060] Sp1: Data input: The operating data of the multi-energy equipment, including power, temperature, and pressure, are used as the input of LSTM;

[0061] Sp2: Health prediction: LSTM predicts the future health status of the device by analyzing the time series;

[0062] Sp3: Self-healing decision: Based on the health status of the device, the system schedules other devices in advance to balance the load and avoid power supply interruption caused by failure of a single device.

[0063] Preferably, the multi-level energy coupling and conversion module efficiently converts various forms of energy, constrained by a linear programming formula, and further includes the following contents:

[0064] Sp1: Energy supply and demand matrix: define the supply and demand matrix and conversion efficiency matrix of different energy forms;

[0065] Sp2: Solve the optimal solution: Solve the optimal energy supply strategy through linear programming to minimize the energy supply cost and meet the multi-energy needs of the park.

[0066] Preferably, the operation process of the system is as follows:

[0067] Sp1: Data collection: Collect operating data and external conditions of various energy equipment in the park through sensors, including temperature and user needs;

[0068] Sp2: Quantum optimization: The quantum computing optimization module quickly calculates the optimal energy scheduling plan for the multi-energy network based on the data and puts forward preliminary scheduling suggestions;

[0069] Sp3: Adaptive Scheduling: The adaptive energy scheduling module uses deep reinforcement learning and fuzzy logic modules to further adjust energy scheduling based on the quantum optimization solution according to real-time data changes, including uncertain factors such as weather and user demand, so as to adapt to real-time demand changes;

[0070] Sp4: Equipment collaboration and self-healing: The equipment collaboration and self-healing module monitors the multi-energy equipment during operation and predicts the health status of the equipment through the LSTM model. Once an equipment abnormality occurs, the system automatically adjusts the operating status of other equipment to ensure the continuous energy supply of the system.

[0071] Sp5: Energy coupling and conversion: The multi-level energy coupling and conversion module optimizes the coupling and conversion between multi-energy devices, and the system realizes efficient conversion between different energy forms such as electricity, heat, cold, hydrogen, etc. to meet the diverse energy needs in the park.

[0072] Preferably, the data collection is performed in real time through various sensors installed in the park, and the sensors include power metering devices, temperature sensors, and weather stations. The collected data include energy equipment operation data, environmental data, and user demand data.

[0073] Preferably, the states of multi-energy devices are encoded by the Ising model in the quantum optimization, and the quantum computing optimization module finds the device combination with the lowest energy consumption according to the energy consumption and interaction of each device, and generates preliminary scheduling suggestions.

[0074] Beneficial Effects

[0075] The present invention provides a park integrated energy system model based on multi-energy network and multi-energy equipment. It has the following beneficial effects:

[0076] 1. The present invention uses the synergy of quantum computing and deep learning. The system can find the global optimal solution for energy scheduling in real time and improve the overall energy utilization efficiency. Compared with traditional computing methods, quantum computing is faster and more accurate in dealing with complex multivariable problems, greatly reducing the optimization calculation time, especially in the energy equipment scheduling scenario. It can respond in real time. Whether it is the electricity demand or the heat demand in the park, the system can make efficient and intelligent scheduling, and can flexibly respond to complex energy demands, maximize the energy utilization efficiency and system reliability, and provide strong support for the comprehensive energy management of modern smart parks.

[0077] 2. The system used in the present invention can automatically adjust the operating status of the equipment according to real-time data, reducing the need for manual intervention. The continuous learning mechanism of the deep reinforcement learning model enables the system to adapt to future complex environmental changes. The adaptive scheduling of the system ensures smooth operation under different environmental conditions without being overly affected by external uncertain factors. Since it can accurately predict changes in demand, the system avoids excessive energy supply or energy waste, thereby further improving energy utilization efficiency.

[0078] 3. The present invention uses the system's intelligent monitoring and self-healing functions for multi-energy equipment to greatly improve the reliability of the system. Through real-time prediction of the health status of the equipment, the system can make adjustments before the equipment fails to avoid energy supply terminals. The equipment self-healing mechanism is activated when a certain device fails to ensure that other devices in the system can quickly fill in the gaps and avoid energy supply interruptions. Ensure the stable operation of the entire energy system. This fault warning and automatic self-healing capabilities have significant advantages in large-scale park integrated energy systems.

[0079] 4. The system used in the present invention can flexibly convert between various energy forms such as electric energy, thermal energy, cold energy and hydrogen energy, and realize dynamic scheduling of different energy sources. Whether there is a shortage of electricity or an increase in heat demand, the system can flexibly adjust the proportion of various energy sources to maximize energy utilization. The system significantly reduces energy consumption by optimizing energy scheduling, reducing waste, and improving energy conversion efficiency, thereby reducing carbon emissions and negative impacts on the environment. It provides important support for the development of green and low-carbon parks. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 is a system structure diagram of the present invention;

[0081] Figure 2 It is a flow chart of the system operation of the present invention;

[0082] Figure 3 A visualization diagram of the quantum computing optimization module of the present invention;

[0083] Figure 4 A fuzzy logic visualization diagram of the adaptive energy scheduling module of the present invention;

[0084] Figure 5 A reinforcement learning visualization chart for the deep reinforcement learning module of the present invention;

[0085] Figure 6 A visualization chart of the LSTM prediction module of the present invention;

[0086] Figure 7 A visualization diagram of the linear programming module of the present invention;

[0087] Figure 8 It is a visualization chart of the comprehensive operation process of the present invention. DETAILED DESCRIPTION

[0088] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Specific embodiment one:

[0090] like Figure 1-8 As shown in the figure, the integrated energy system model of the park based on multi-energy network and multi-energy equipment includes quantum computing optimization module, adaptive energy scheduling module, equipment coordination and self-healing module and multi-level energy coupling and conversion module, among which:

[0091] Quantum computing optimization module: responsible for handling complex energy scheduling problems, used to find the best scheduling solution in large-scale energy networks, and to find the global optimal solution through the quantum annealing algorithm. The quantum annealing algorithm is represented by the Ising model:

[0092]

[0093] in,

[0094] H(σ) represents the minimum optimization energy:

[0095] σ i Indicates the status of energy equipment, including on and off, power supply and shutdown;

[0096] h i Indicates the local energy consumption of the equipment;

[0097] J ij Represents the interactions between devices, including transmission losses;

[0098] Adaptive energy scheduling module: It combines deep reinforcement learning (DRL) and fuzzy logic to handle uncertainty factors, dynamically adapt to changes in energy demand, perform adaptive scheduling and optimization control, and make optimal decisions. The formula for fuzzy logic to handle uncertainty is as follows:

[0099]

[0100] in;

[0101] x represents input variables, including temperature and energy demand;

[0102] μ A (x) represents fuzzy membership;

[0103] c represents the expected value of the input;

[0104] b indicates the fuzzy range;

[0105] p represents the control ambiguity;

[0106] On this basis, energy scheduling is performed through deep reinforcement learning, and the formula is as follows:

[0107]

[0108] in,

[0109] Q(s,a) represents the expected return of taking action a in state s;

[0110] r represents immediate reward;

[0111] γ represents the discount factor;

[0112] Equipment collaboration and self-healing module: Use machine learning technology to perform predictive maintenance on multi-functional equipment. When an abnormality in equipment operation is detected, the system will take measures in advance to adjust the working status of other equipment and perform self-healing and adjustment of the system. Use LSTM long short-term memory network to predict the equipment operation status. The core equation of LSTM is:

[0113] h t =o t tanh(C t )

[0114] in,

[0115] h t Indicates hidden state;

[0116] o t represents the output gate;

[0117] C t Represents a memory unit;

[0118] Multi-level energy coupling and conversion module: This module is responsible for the flexible coupling and conversion of various energy forms. Through intelligent management, it converts and utilizes different energy sources. The multi-level coupling and conversion process is described by a linear programming model:

[0119]

[0120] The constraints are:

[0121]

[0122] in,

[0123] C i Indicate the costs of different forms of energy;

[0124] x i Indicates the quantity supplied;

[0125] a ij Indicates the conversion efficiency between different energy forms;

[0126] b j Represents energy demand.

[0127] The above system can flexibly respond to complex energy needs. Through optimized scheduling, adaptive adjustment and multi-energy conversion, it maximizes energy utilization efficiency and system reliability, while also significantly reducing energy waste and carbon emissions. These advantages will provide strong support for the comprehensive energy management of modern smart parks. Specific embodiment 2:

[0129] like Figure 1-8 As shown, according to the contents in the above specific embodiments, the following technical contents are further disclosed:

[0130] Quantum computing has significant advantages in dealing with combinatorial optimization problems and energy system scheduling because it can process multiple states in parallel and find the global optimal solution. The energy scheduling problem of the park involves the coordinated operation of multiple energy devices, including equipment in various energy forms such as electricity, thermal energy, and hydrogen energy. The quantum annealing algorithm can map the scheduling problem of these complex devices into an Ising model and find the optimal solution by minimizing the energy function. The quantum computing optimization module uses the Ising model to seek the optimal solution by minimizing the energy function. The solution process is as follows:

[0131] Sp1: Determine the device state. Each device i is in the on or off state, corresponding to σ i = +1 or σ i = -1;

[0132] Sp2: Define local energy consumption h i , the energy consumption of each device in different states, including the energy consumption of power equipment when it is running;

[0133] Sp3: Interactions between computing devicesJ ij , describes the interaction between energy devices, some devices working together will bring additional energy consumption, which is represented by the interaction term;

[0134] Sp4: Annealing process. Through quantum annealing, the system will continuously try to reduce the total energy H(σ) until the optimal scheduling combination is found.

[0135] Quantum optimization can not only significantly reduce computing time, but also ensure that the globally optimal scheduling solution is found. It is particularly suitable for highly complex system processing. Through the synergy of quantum computing and deep learning, the system can find the global optimal solution for energy scheduling in real time and improve the overall energy utilization efficiency. Compared with traditional computing methods, quantum computing is faster and more accurate when dealing with complex multivariable problems, greatly reducing the optimization calculation time, especially in energy equipment scheduling scenarios. It can respond in real time. Regardless of the power demand or heat demand in the park, the system can make efficient and intelligent scheduling, and can flexibly respond to complex energy needs, maximizing energy utilization efficiency and system reliability, and providing strong support for the comprehensive energy management of modern smart parks.

[0136] The adaptive energy scheduling module handles uncertain energy demand by combining deep reinforcement learning and fuzzy logic. Fuzzy logic is used to deal with unpredictable factors such as weather and equipment load fluctuations, while deep reinforcement learning is used for long-term optimization and continuous learning. Deep reinforcement learning (DQN) can continuously optimize decisions based on historical data, allowing the system to adapt to changing environments.

[0137] The fuzzy logic module in the adaptive energy scheduling module is used to handle unpredictable factors, including provocations, equipment load fluctuations, wherein the fuzzy logic formula further includes the following steps:

[0138] Sp1: Input fuzzification, which transforms input variables such as energy demand, including weather and temperature, into fuzzy sets through membership functions to reflect their uncertainty;

[0139] Sp2: Fuzzy reasoning: Generate fuzzy control instructions for energy equipment through fuzzy reasoning rules;

[0140] Sp3: Defuzzification: Convert fuzzy output into actual equipment operation, including start and stop control of equipment.

[0141] Deep reinforcement learning further includes the following:

[0142] The state space s represents the current state of the system, including the current supply of energy equipment and the state of the equipment;

[0143] The action space a represents the actions that the system can take, including dispatching more power and shutting down specified equipment;

[0144] Reward r, instant feedback for each dispatch, including data on energy efficiency and cost reduction;

[0145] Optimal strategy update,Through continuous iterations, the system will learn which scheduling decisions can bring higher long-term returns, and handle uncertainty through fuzzy logic.

[0146] Deep reinforcement learning is continuously optimized in the long term, further enabling the system to dynamically adapt to changing energy demands. The system can automatically adjust the operating status of the equipment according to real-time data, reducing the need for human intervention. The continuous learning mechanism of the deep reinforcement learning model enables the system to adapt to complex environmental changes in the future. The system's adaptive scheduling ensures smooth operation under different environmental conditions without being overly affected by external uncertain factors. Since it can accurately predict changes in demand, the system avoids excessive energy supply or energy waste, thereby further improving energy utilization efficiency.

[0147] Equipment collaboration and self-healing module, the core of this module is to perform intelligent monitoring and predictive maintenance on multi-energy equipment to ensure continuous operation of the system. By using the LSTM neural network, the system can predict the failure trend of the equipment, make scheduling adjustments before the failure occurs, and keep the system stable. The equipment collaboration and self-healing module predicts the failure trend of the equipment, makes scheduling adjustments before the failure occurs, and keeps the system stable. This is achieved through the LSTM formula. The prediction steps are as follows:

[0148] Sp1: Data input: The operating data of the multi-energy equipment, including power, temperature, and pressure, are used as the input of LSTM;

[0149] Sp2: Health prediction: LSTM predicts the future health status of the device by analyzing the time series;

[0150] Sp3: Self-healing decision: Based on the health status of the device, the system schedules other devices in advance to balance the load and avoid power supply interruption caused by failure of a single device.

[0151] This module ensures that the system can be adjusted before equipment fails, improving the stability and continuity of the system. The system's intelligent monitoring and self-healing functions for multi-energy equipment greatly improve the reliability of the system. Through real-time prediction of equipment health status, the system can make adjustments before equipment fails to avoid energy supply terminals. The equipment self-healing mechanism is activated when a certain device fails, ensuring that other equipment in the system can quickly fill in the gaps and avoid energy supply interruptions. This ensures the stable operation of the entire energy system. This fault warning and automatic self-healing capability has significant advantages in large-scale park integrated energy systems.

[0152] The multi-level energy coupling and conversion module mainly deals with the flexible coupling and conversion between various energy forms, such as the efficient conversion between electric energy, thermal energy, cold energy and hydrogen energy. Through intelligent regulation, different energy sources complement each other and improve energy utilization. The multi-level energy coupling and conversion module efficiently converts various forms of energy and constrains them through linear programming formulas. It further includes the following contents:

[0153] Sp1: Energy supply and demand matrix: define the supply and demand matrix and conversion efficiency matrix of different energy forms;

[0154] Sp2: Solve the optimal solution: Solve the optimal energy supply strategy through linear programming to minimize the energy supply cost and meet the multi-energy needs of the park.

[0155] The system can flexibly convert between various energy forms such as electricity, heat, cold energy and hydrogen energy, and realize dynamic scheduling of different energy sources. Whether there is a shortage of electricity or an increase in heat demand, the system can flexibly adjust the proportion of various energy sources to maximize energy utilization. The system significantly reduces energy consumption by optimizing energy scheduling, reducing waste and improving energy conversion efficiency, thereby reducing carbon emissions and negative impacts on the environment. It provides important support for the development of green and low-carbon parks. Specific embodiment three:

[0157] like Figure 1-8 As shown, according to the content in the above specific embodiments, the following technical contents are further disclosed

[0158] The system operation process is as follows:

[0159] Sp1: Data collection: The operation data and external conditions of each energy device in the park, including temperature and user demand, are collected through sensors installed in the park in real time. The data collection module ensures that the entire system has enough information to make intelligent scheduling decisions. The sensors include power metering devices, temperature sensors, and weather stations. The collected data include energy equipment operation data, environmental data, and user demand data. The collected data include the following:

[0160] Energy equipment operation data: real-time output power, operating status, temperature, pressure, etc. of various energy equipment;

[0161] Environmental data: external temperature, humidity, wind speed, solar radiation intensity, etc.;

[0162] User demand data: real-time energy demand, such as electricity, heat, and cooling demand during a certain period of time;

[0163] Sp2: Quantum optimization: Once enough data is collected, the quantum computing optimization module starts working. First, it calculates the preliminary scheduling plan for each energy device in the park. Through the quantum annealing algorithm, the system can quickly find the optimal combination of energy consumption and optimize the scheduling order of multi-energy devices. The quantum computing optimization module quickly calculates the optimal energy scheduling plan for the multi-energy network based on the data and makes preliminary scheduling suggestions. In quantum optimization, the state of multi-energy devices is encoded through the Ising model. The quantum computing optimization module finds the device combination with the lowest energy consumption based on the energy consumption and interaction of each device, and generates preliminary scheduling suggestions;

[0164] Sp3: Adaptive Scheduling: Based on the preliminary scheduling results provided by quantum optimization, the adaptive energy scheduling module will combine real-time environmental changes and demand fluctuations to further dynamically adjust the operation strategy of the equipment. The deep reinforcement learning module will continuously optimize the strategy based on historical data. At the same time, the fuzzy logic module handles uncertain factors, including weather, user demand and other uncertain factors. Based on the quantum optimization solution, the energy scheduling is further adjusted to adapt to real-time demand changes.

[0165] Sp4: Equipment collaboration and self-healing: The equipment collaboration and self-healing module monitors the multi-functional equipment during operation. During the operation of the equipment, the equipment collaboration and self-healing module continuously monitors the operating status of each equipment. The LSTM neural network analyzes the time series data of the equipment, predicts the failure trend of the equipment, and makes adjustments in advance. When an abnormality is detected in a certain equipment, the system automatically dispatches other equipment to share the load and realizes the self-healing of the system. The health status of the equipment is predicted by the LSTM model. Once an abnormality occurs in the equipment, the system automatically adjusts the operating status of other equipment to ensure the continuous energy supply of the system.

[0166] Sp5: Energy coupling and conversion: The multi-level energy coupling and conversion module optimizes the coupling and conversion between multi-energy devices. The system realizes efficient conversion between different energy forms such as electricity, heat, cold, and hydrogen to meet the diverse energy needs in the park. In order to maximize energy utilization, the system will adjust the conversion ratio between different energy forms according to real-time needs. For example, when the park's electricity demand decreases and the heat demand increases, the system will automatically start the electric-heat conversion equipment to convert excess electricity into thermal energy storage.

[0167] Through real-time data collection, sensors are used to collect real-time data of various energy equipment and user needs in the park, providing basic information for intelligent scheduling and accurate data support, so that the system can make real-time decisions. Quantum computing preliminary scheduling uses quantum computing to find the global optimal solution for energy scheduling and formulates a preliminary scheduling plan to achieve the initial scheduling plan with the lowest energy consumption and improve energy utilization efficiency. Deep reinforcement learning and fuzzy logic adaptive adjustment combines fuzzy logic and deep learning to further optimize the scheduling strategy in a real-time environment to cope with uncertainty and demand fluctuations, ensure that the system can flexibly respond to external changes, and ensure the stability and efficiency of energy supply. Equipment health monitoring and self-healing predict equipment failures through equipment monitoring and LSTM models, take scheduling strategies in advance, ensure the continuity of system operation, improve system reliability, and avoid energy supply interruptions. Multi-energy conversion and optimization When energy demand changes, the system intelligently optimizes the conversion and scheduling between different energy forms, realizes multi-energy coupling and coordination, improves the overall utilization of multiple energy forms, reduces energy conversion losses, and flexibly responds to multiple energy needs. Through these five steps, the system can efficiently manage and schedule multiple energy forms in a complex park environment to achieve energy saving, consumption reduction and intelligent control. Specific embodiment four:

[0169] like Figure 1-8 As shown, according to the contents in the above specific embodiments, the following technical contents are further disclosed:

[0170] like Figure 3 As shown, a visualized picture of the quantum computing optimization module, the interaction matrix between devices: By visualizing the interaction matrix between devices through the imagesc function, you can intuitively see the mutual influence between different devices. The highlighted area indicates the strong interaction between devices. The on / off state of the device is displayed through a bar chart, so that the state of each device is clear at a glance. Visualization can help engineers understand how the quantum computing optimization module finds the optimal energy scheduling plan through different device states and interactions, thereby minimizing total energy consumption. It provides a clear way to show the interaction between devices and the state of each device, making the optimization plan more explainable.

[0171] like Figure 4 As shown in the figure, the fuzzy logic visualization processing of the adaptive energy scheduling module, fuzzy membership function: by visualizing the fuzzy membership function through the plot function, it can show how the input variables (such as temperature, energy demand) are converted into fuzzy sets and reflect the fuzziness of these variables. The visualization of fuzzy logic helps to understand how to deal with uncertain factors such as weather and equipment load fluctuations and enhance the robustness of the system. Through the fuzzy membership function, we can better understand how the system adjusts energy scheduling according to actual conditions to achieve better dynamic adaptability and optimization control.

[0172] like Figure 5 As shown in the figure, the reinforcement learning visualization of the deep reinforcement learning module. By visualizing the Q value in the Q table through the bar chart, you can intuitively understand the expected returns of different actions in the current state. Visualizing the Q value helps to understand how deep reinforcement learning evaluates the value of different actions in different states, thereby optimizing long-term strategies. It shows the relationship between immediate rewards and long-term returns, and helps engineers understand how the system can improve energy efficiency and reduce costs through continuous learning and adjustment.

[0173] like Figure 6 As shown in the figure, the visualization of the LSTM prediction module shows the comparison between the actual data and the LSTM prediction data through the plot function, which can intuitively show the effect and accuracy of the prediction model. The visualization of the LSTM prediction module can help understand how to predict the future status of the equipment, provide early warning, and avoid the impact of equipment failure. Through the prediction data, the working status of other equipment can be better adjusted to ensure the stability and sustainability of the system. Specific embodiment five:

[0175] like Figure 1-8 As shown, according to the contents in the above specific embodiments, the following technical contents are further disclosed:

[0176] According to the content of the above embodiment, an experiment of comprehensive energy in an actual park is conducted to verify the application effect of the core technology of the system model in the regional park, compare its differences with the traditional energy management system in terms of energy scheduling optimization, energy utilization efficiency, equipment fault self-healing and multi-energy conversion, and analyze the technical advantages and energy-saving effects of the system. The experimental data are as follows:

[0177] Experimental design:

[0178] Experimental location: Two similar regional parks were selected for comparative experiments, one using the traditional energy dispatching system and the other using this system model for processing;

[0179] Experimental period: 6 months;

[0180] Energy type: including electricity, heat, cold, hydrogen and other forms;

[0181] Data collection equipment: sensors, power metering devices, temperature sensors, weather stations, etc., used to collect energy demand, equipment status and external conditions such as weather and temperature;

[0182] Comparison dimensions: energy dispatch efficiency, energy utilization, energy consumption cost, equipment self-healing rate, system stability, including equipment failure rate and self-healing response time;

[0183] The experimental steps are as follows:

[0184] Sp1: Data Collection

[0185] Park A (traditional energy management system): collects the park's energy demand, environmental data and equipment status through sensors every hour, and records the actual use of different energy forms such as electricity, heat, cold, and hydrogen;

[0186] Park B (this system): uses the same sensor data collection method, but performs quantum computing optimization, adaptive scheduling, equipment health prediction, and multi-level energy conversion management in real time in the system;

[0187] Sp2: Quantum Optimization

[0188] In Park B, the quantum computing optimization module is used to run the quantum annealing algorithm once an hour based on sensor data to optimize global energy scheduling. The optimization content includes equipment switching status, energy transmission scheduling, energy conversion ratio, etc.

[0189] Sp3: Adaptive Energy Scheduling

[0190] Park B uses fuzzy logic to handle uncertain factors, such as weather and energy demand fluctuations, and combines deep reinforcement learning to perform dynamic energy scheduling and adjust equipment operation strategies to ensure that real-time needs are met;

[0191] Sp4: Equipment collaboration and self-healing

[0192] In the case of equipment failure in the two parks, the response time and self-healing rate of the equipment were recorded:

[0193] Park A’s traditional approach relies on manual intervention or simple equipment redundancy strategies;

[0194] Park B uses LSTM to predict the health status of equipment, predict equipment failures in advance, and perform system self-adjustment;

[0195] Sp5: Energy coupling and conversion

[0196] The multi-level energy coupling and conversion module of Park B dynamically adjusts the supply ratio of each energy form according to actual demand, optimizes the conversion efficiency of electricity, heat, cold and hydrogen energy, and ensures efficient coordination between multi-energy equipment;

[0197] Data collection and analysis:

[0198] Energy dispatch efficiency (hourly):

[0199] Park A: Equipment is manually scheduled every hour, and coordination between equipment is not precise enough;

[0200] Park B: Based on the quantum annealing algorithm, the global optimal solution can be found for each scheduling;

[0201] Energy utilization rate (%):

[0202] Park A: There is a lot of energy waste in the traditional energy dispatching system, especially the over-supply when the equipment is turned on and off;

[0203] Park B: The system reduces energy waste through quantum optimization and deep learning, and energy utilization is significantly improved;

[0204] Equipment failure rate (%):

[0205] Park A: Faulty equipment failed to provide early warning, resulting in multiple power outages;

[0206] Park B: The LSTM model predicts failures in advance and performs equipment switching, resulting in a high system self-healing rate and stable energy supply.

[0207] Energy consumption cost (unit: yuan):

[0208] Park A: High energy costs, especially during peak energy hours, and manual scheduling leads to oversupply of energy;

[0209] Park B: The system optimizes energy conversion, has high utilization rate and significantly reduces costs;

[0210] Energy conversion efficiency (%):

[0211] Park A: The conversion efficiency between various energy forms is low and cannot be adjusted dynamically according to demand;

[0212] Park B: Optimize energy conversion through linear programming to ensure that the conversion efficiency of each energy form is maximized.

[0213] Through the above experiments, the experimental data obtained are shown in Table 1 below:

[0214]

[0215] According to Table 1 above, the park using this system has significantly improved the energy scheduling efficiency, equipment coordination capability, system reliability and energy utilization rate. Compared with the traditional system, it has outstanding technical advantages and economic benefits in regional park energy management. This system quickly finds the global optimal scheduling plan through the quantum annealing algorithm, significantly improves the energy scheduling efficiency, and can adapt to changes in energy demand in real time. The energy scheduling efficiency is significantly improved. Through the combination of deep reinforcement learning and fuzzy logic, this system enables the system to adaptively adjust the energy supply to avoid excessive energy supply, and the energy utilization rate is increased to 90%. The LSTM long short-term memory network predicts equipment failures in advance, so that the park system can automatically adjust the equipment status, reduce the failure rate, and shorten the self-healing response time. Through the intelligent management of multi-level energy coupling and conversion modules, the energy supply and demand are highly matched, the overall energy consumption cost is reduced, and the energy expenditure of the park is significantly reduced. The efficient energy utilization and conversion of the system reduces energy waste and carbon emissions, and promotes the development of green and low-carbon parks.

[0216] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprising a reference structure" do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0217] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A comprehensive energy system model for a park based on multi-energy networks and multi-energy devices, characterized by: The system model includes a quantum computing optimization module, an adaptive energy scheduling module, an equipment coordination and self-healing module, and a multi-level energy coupling and conversion module, wherein: The quantum computing optimization module is responsible for handling complex energy scheduling problems and is used to find the optimal scheduling solution in a large-scale energy network. The quantum annealing algorithm is used to find the global optimal solution. The quantum annealing algorithm is represented by the Ising model: in, H(σ) represents the minimum optimization energy: σ i Indicates the status of energy equipment, including on and off, power supply and shutdown; h i Indicates the local energy consumption of the equipment; J ij Represents the interactions between devices, including transmission losses; The adaptive energy scheduling module combines deep reinforcement learning (DRL) and fuzzy logic to handle uncertainty factors, dynamically adapt to changes in energy demand, perform adaptive scheduling and optimization control, and make optimal decisions. The formula for fuzzy logic to handle uncertainty is as follows: in; x represents input variables, including temperature and energy demand; μ A (x) represents fuzzy membership; c represents the expected value of the input; b indicates the fuzzy range; p represents the control ambiguity; On this basis, energy scheduling is performed through deep reinforcement learning, and the formula is as follows: in, Q(s,a) represents the expected return of taking action a in state s; r represents immediate reward; γ represents the discount factor; The equipment coordination and self-healing module: uses machine learning technology to perform predictive maintenance on multi-functional equipment. When an abnormality in equipment operation is detected, the system will take measures in advance to adjust the working status of other equipment and perform self-healing and adjustment of the system. The LSTM long short-term memory network is used to predict the equipment operation status. The core equation of LSTM is: h t =o t ·tanh(C t ) in, h t Indicates hidden state; o t represents the output gate; C t Represents a memory unit; The multi-level energy coupling and conversion module: This module is responsible for the flexible coupling and conversion of various energy forms. Through intelligent management, it performs the conversion and utilization of different energy sources. The multi-level coupling and conversion process is described by a linear programming model: The constraints are: in, C i Indicate the costs of different forms of energy; x i Indicates the quantity supplied; a ij Indicates the conversion efficiency between different energy forms; b j Represents energy demand.

2. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 1 is characterized by: The quantum computing optimization module uses the Ising model to seek the optimal solution by minimizing the energy function. The solution process is as follows: Sp1: Determine the device state. Each device i is in the on or off state, corresponding to σ i = +1 or σ i = -1; Sp2: Define local energy consumption h i , the energy consumption of each device in different states, including the energy consumption of power equipment when it is running; Sp3: Interactions between computing devicesJ ij , describes the interaction between energy devices, where some devices working together will bring additional energy consumption, which is represented by the interaction term; Sp4: Annealing process. Through quantum annealing, the system will continuously try to reduce the total energy H(σ) until the optimal scheduling combination is found.

3. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 1 is characterized by: The fuzzy logic module in the adaptive energy scheduling module is used to handle unpredictable factors, including provocations, equipment load fluctuations, wherein the fuzzy logic formula further includes the following steps: Sp1: Input fuzzification, which transforms input variables such as energy demand, including weather and temperature, into fuzzy sets through membership functions to reflect their uncertainty; Sp2: Fuzzy reasoning: Generate fuzzy control instructions for energy equipment through fuzzy reasoning rules; Sp3: Defuzzification: Convert fuzzy output into actual equipment operation, including start and stop control of equipment.

4. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 1 is characterized by: The deep reinforcement learning in the adaptive energy scheduling module is used for long-term optimization and continuous learning. The deep reinforcement learning further includes the following: The state space s represents the current state of the system, including the current supply of energy equipment and the state of the equipment; The action space a represents the actions that the system can take, including dispatching more power and shutting down specified equipment; Reward r, instant feedback for each dispatch, including data on energy efficiency and cost reduction; The optimal strategy is updated. Through continuous iteration, the system will learn which scheduling decisions can bring higher long-term returns. By handling uncertainty through fuzzy logic, deep reinforcement learning is continuously optimized in the long term, further enabling the system to dynamically adapt to changing energy needs.

5. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 1 is characterized by: The device coordination and self-healing module predicts the failure trend of the device and makes scheduling adjustments before the failure occurs to maintain system stability. This is achieved through the LSTM formula. The prediction steps are as follows: Sp1: Data input: The operating data of the multi-energy equipment, including power, temperature, and pressure, are used as the input of LSTM; Sp2: Health prediction: LSTM predicts the future health status of the device by analyzing the time series; Sp3: Self-healing decision: Based on the health status of the device, the system schedules other devices in advance to balance the load and avoid power supply interruption caused by failure of a single device.

6. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 1 is characterized by: The multi-level energy coupling and conversion module efficiently converts various forms of energy, constrained by linear programming formulas, and further includes the following contents: Sp1: Energy supply and demand matrix: define the supply and demand matrix and conversion efficiency matrix of different energy forms; Sp2: Solve the optimal solution: Solve the optimal energy supply strategy through linear programming to minimize the energy supply cost and meet the multi-energy needs of the park.

7. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 1 is characterized by: The operation process of the system is as follows: Sp1: Data collection: Collect operating data and external conditions of various energy equipment in the park through sensors, including temperature and user needs; Sp2: Quantum optimization: The quantum computing optimization module quickly calculates the optimal energy scheduling plan for the multi-energy network based on the data and puts forward preliminary scheduling suggestions; Sp3: Adaptive Scheduling: The adaptive energy scheduling module uses deep reinforcement learning and fuzzy logic modules to further adjust energy scheduling based on the quantum optimization solution according to real-time data changes, including uncertain factors such as weather and user demand, so as to adapt to real-time demand changes; Sp4: Equipment collaboration and self-healing: The equipment collaboration and self-healing module monitors the multi-energy equipment during operation and predicts the health status of the equipment through the LSTM model. Once an equipment abnormality occurs, the system automatically adjusts the operating status of other equipment to ensure the continuous energy supply of the system. Sp5: Energy coupling and conversion: The multi-level energy coupling and conversion module optimizes the coupling and conversion between multi-energy devices, and the system realizes efficient conversion between different energy forms such as electricity, heat, cold, hydrogen, etc. to meet the diverse energy needs in the park.

8. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 7 is characterized by: In the data collection, various sensors installed in the park are used to collect data in real time, and the sensors include power metering devices, temperature sensors, and weather stations. The collected data include energy equipment operation data, environmental data, and user demand data.

9. The park integrated energy system model based on multi-energy network and multi-energy equipment according to claim 7 is characterized by: In the quantum optimization, the state of multi-energy devices is encoded through the Ising model. The quantum computing optimization module finds the device combination with the lowest energy consumption according to the energy consumption and interaction of each device and generates preliminary scheduling suggestions.

Images