Intelligent coordination control method and system of multi-source heat pump coupling dehumidification system of power distribution station building

Abstract

The application relates to the field of power system control, and provides an intelligent coordination control method and system for a multi-source heat pump coupling dehumidification system of a power distribution station house, which comprises the following steps: performing multi-source heat pump characteristic analysis through a heat source quality evaluation model to obtain heat source characteristic parameters; dynamically reconstructing a heat exchange network topology according to the heat source characteristic parameters to obtain optimized heat exchange network configuration parameters; performing gradient dehumidification working medium circulation on the optimized heat exchange network configuration parameters through a multi-effect evaporator to obtain dehumidification working medium circulation parameters; performing neuromorphic control on the dehumidification working medium circulation parameters through a bionic pulse neural network to obtain humidity control instructions; performing dynamic load distribution on the multi-source heat pump system according to the humidity control instructions to obtain an energy efficiency coordination optimization strategy, and coordinating and controlling the multi-source heat pump through the energy efficiency coordination optimization strategy. The application improves the evaluation accuracy of heat source performance parameters of the multi-source heat pump and reduces the overall energy consumption level of the system.

Description

Technical Field

[0001] This invention relates to the field of power system control technology, and in particular to an intelligent coordinated control method and system for a multi-source heat pump coupled dehumidification system in a power distribution station. Background Technology

[0002] As a crucial component of the power system, power distribution substations have extremely stringent requirements for environmental humidity control. Traditional dehumidification systems in power distribution substations primarily employ dehumidification equipment driven by a single heat source, such as electric heating dehumidifiers and condensing dehumidifiers. These systems are relatively independent in terms of heat source utilization. Existing technologies include ground source heat pump systems that utilize the constant temperature characteristics of the underground to provide a stable heat source, air source heat pumps that exchange heat with ambient air, and waste heat recovery systems for power distribution equipment that recover waste heat generated by transformers, switchgear, and other equipment. However, these heat source systems typically operate independently and fail to form an organically coupled system.

[0003] In terms of control strategies, existing systems mostly employ traditional PID controllers or simple logic control, adjusting the operating parameters of dehumidifiers based on feedback signals from temperature and humidity sensors. For existing control strategies, multi-source heat pump systems lack effective coordinated control mechanisms. Independent operation of each heat source leads to low overall energy efficiency and an inability to optimize configuration based on the performance characteristics of each heat source under different operating conditions. Secondly, traditional control strategies suffer from slow response speed and low accuracy, making it difficult to cope with the complex and variable humidity control needs of power distribution stations, especially prone to control lag when the load changes rapidly. Thirdly, the fixed topology of the heat exchange network prevents dynamic reconfiguration based on real-time operating status, limiting the system's flexibility and adaptability. Finally, existing systems lack predictive maintenance and intelligent operation and maintenance capabilities; equipment failures often require manual inspection, resulting in high maintenance costs and potential disruptions to continuous system operation. These technological limitations severely restrict the overall performance and economy of power distribution station dehumidification systems. Summary of the Invention

[0004] This invention provides an intelligent coordinated control method and system for a multi-source heat pump coupled dehumidification system in a power distribution room, in order to overcome the shortcomings of the prior art.

[0005] This invention provides an intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station, comprising:

[0006] S1: Analyze the characteristics of multi-source heat pumps using a heat source quality assessment model to obtain heat source characteristic parameters;

[0007] S2: Dynamically reconstruct the heat exchange network topology based on the heat source characteristic parameters to obtain optimized heat exchange network configuration parameters;

[0008] S3: The optimized heat exchange network configuration parameters are circulated in stages using a multi-effect evaporator to obtain the dehumidifying working fluid circulation parameters.

[0009] S4: Obtain humidity control instructions by performing neuromorphic control on the dehumidifying working fluid circulation parameters through a biomimetic pulse neural network;

[0010] S5: Dynamically allocate the load of the multi-source heat pump system according to the humidity control command to obtain an energy efficiency coordination optimization strategy, and coordinate the control of the multi-source heat pump through the energy efficiency coordination optimization strategy.

[0011] According to the present invention, a smart coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station is provided, wherein the multi-source heat pump in step S1 includes a ground source heat pump, an air source heat pump, and a waste heat recovery heat pump from power distribution equipment.

[0012] According to the present invention, a smart coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution room is provided, wherein step S1 further includes:

[0013] S11: Collect basic heat source parameters for the ground source heat pump;

[0014] S12: Collect the heat source response of the air source heat pump within the preset ambient temperature range to obtain the peak heat source parameters of the air source heat pump;

[0015] S13: Waste heat recovery from transformer and switchgear operation is carried out using micro heat pump technology to obtain waste heat recovery parameters of power distribution equipment;

[0016] S14: The heat source quality is evaluated by using artificial intelligence algorithms to assess the basic heat source parameters, the peak heat source parameters, and the waste heat recovery parameters of the power distribution equipment, thereby obtaining the heat source characteristic parameters.

[0017] According to the present invention, a smart coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution room is provided, wherein step S2 further includes:

[0018] S21: Based on the heat source characteristic parameters, the preset plate heat exchanger array is switched between series, parallel, and series-parallel hybrid modes to obtain the heat exchange network connection method;

[0019] S22: Based on the heat exchange network connection method, the mixing ratio of heat media at different temperature levels is calculated by the heat medium mixing optimization algorithm to generate target temperature adjustment parameters and phase change triggering commands.

[0020] S23: In response to the phase change trigger command, microwave radiation is applied to the embedded phase change material capsule to maintain heat flow continuity during topology switching and obtain heat flow buffer queue parameters;

[0021] S24: Using the highest overall system efficiency and lowest energy consumption as objective functions, the particle swarm optimization algorithm is used to dynamically solve the parameters of the intelligent three-way valve and the electric regulating valve to obtain the optimized heat exchange network configuration parameters.

[0022] According to the present invention, a smart coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution room is provided, wherein step S23 further includes:

[0023] S231: Embed composite phase change material capsules into the gaps of the flow channels in a plate heat exchanger;

[0024] S232: Selective excitation of phase change material capsules in the target region using a microwave radiation array;

[0025] S223: Real-time monitoring of optical transmittance changes in phase change material capsules, determining complete phase change based on transmittance decrease, and generating a heat flow continuity maintenance confirmation signal;

[0026] S234: Update the heat flow buffer queue parameters according to the heat flow continuity maintenance confirmation signal to obtain the heat flow buffer queue parameters.

[0027] According to the present invention, a smart coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution room is provided, wherein step S3 further includes:

[0028] S31: By using a triple-effect evaporator series structure, the environmentally friendly working fluid R1234yf is evaporated in stages at multiple temperature levels to obtain multi-effect evaporation parameters;

[0029] S32: Based on the multi-effect evaporation parameters, high-temperature steam, condensate and low-temperature steam are used for cascaded energy utilization to obtain parameters for maximizing energy utilization.

[0030] S33: The plate heat exchanger recovers the latent heat of the exhaust steam and preheats the working fluid entering the evaporator to obtain the dehumidifying working fluid circulation parameters.

[0031] According to the present invention, a smart coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution room is provided, wherein step S4 further includes:

[0032] S41: By using a biomimetic spiking neural network to perform neuromorphic modeling of the temperature, humidity and equipment operating status of the substation environment, a spiking neuron control architecture is obtained.

[0033] S42: By using the STDP learning rule, the spatiotemporal pulse sequence of the spiking neuron control architecture is adaptively learned to obtain distributed decision parameters;

[0034] S43: The distributed decision parameters are encoded by a sparse pulse event stream using an FPGA-accelerated pulse encoder to obtain a sub-second response control signal;

[0035] S44: Perform neuromorphic adjustment processing on the operating status of each dehumidification module according to the sub-second response control signal to obtain the humidity control command.

[0036] According to the present invention, a smart coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution room is provided, wherein step S41 further includes:

[0037] S411: Map the ground source heat pump, air source heat pump and power distribution equipment waste heat recovery heat pump to independent pulse neurons, and obtain the topology of the three-source neuron network.

[0038] S412: Set the connection strength of the three-source neuron network topology through the synaptic weight matrix to obtain the coupling parameters between neurons;

[0039] S413: Optimize the pulse propagation path according to the interneuron coupling parameters to obtain the neuromorphic control architecture.

[0040] According to the intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution room provided by the present invention, step S5 further includes:

[0041] S51: Through a cloud-edge collaborative architecture, the COP (Coefficient of Performance) and SEER (Seasonal Energy Efficiency Ratio) of multiple heat pump modules are calculated in real time to obtain energy efficiency evaluation parameters.

[0042] S52: Based on the energy efficiency assessment parameters and the humidity control command, intelligently allocate the dehumidification load to obtain a load allocation scheme;

[0043] S53: The power grid demand response function accesses electricity market information and participates in power grid peak shaving and frequency regulation services to obtain the energy efficiency coordination and optimization strategy.

[0044] This invention also provides an intelligent coordinated control system for a multi-source heat pump coupled dehumidification system in a power distribution station, comprising:

[0045] Analysis module: Used to perform multi-source heat pump characteristic analysis using a heat source quality assessment model to obtain heat source characteristic parameters;

[0046] Reconstruction module: used to dynamically reconstruct the heat exchange network topology based on the heat source characteristic parameters to obtain optimized heat exchange network configuration parameters;

[0047] Circulation module: used to circulate the dehumidifying working fluid in stages through the multi-effect evaporator to obtain the dehumidifying working fluid circulation parameters;

[0048] Control module: used to perform neuromorphic control on the circulation parameters of the dehumidifying working fluid through a biomimetic pulse neural network to obtain humidity control commands;

[0049] Distribution module: used to dynamically distribute the load of the multi-source heat pump system according to the humidity control command, obtain an energy efficiency coordination optimization strategy, and coordinate the control of the multi-source heat pump through the energy efficiency coordination optimization strategy.

[0050] This invention provides an intelligent coordinated control method and system for a multi-source heat pump coupled dehumidification system in a power distribution station. By employing a heat source quality assessment model to analyze the characteristics of ground source heat pumps, air source heat pumps, and waste heat recovery heat pumps from power distribution equipment, it achieves precise quantification and dynamic evaluation of the performance parameters of each heat source, effectively solving the problem of low energy utilization efficiency in traditional single-heat source systems and significantly reducing the overall energy consumption level of the system. Secondly, by replacing the traditional controller with a biomimetic spiking neural network for neuromorphic control processing, the system possesses the low power consumption and high parallel processing capabilities of brain-like computing, significantly shortening the control response time. Simultaneously, the introduction of STDP learning rules enables the system to autonomously learn and adapt to complex humidity change patterns, significantly improving the accuracy and stability of humidity control and solving the technical problems of lag and insufficient control precision in traditional control systems. Furthermore, the phase change material capsule-assisted topology reconstruction caching mechanism effectively solves the problem of heat loss during dynamic switching of the heat exchange network through microwave radiation-triggered instantaneous heat storage and release processing, enabling the system to maintain the continuity of heat transfer while rapidly reconstructing the topology, avoiding... This invention eliminates the response lag and energy waste inherent in traditional fixed topology systems during load changes, significantly improving system flexibility and adaptability. Furthermore, the multi-effect evaporator series structure combined with cascaded energy utilization maximizes heat recovery and utilization, fully exploiting the value of heat sources at various temperature levels and effectively reducing system operating costs. Simultaneously, the FPGA-accelerated pulse encoder algorithm, which converts analog signals into sparse pulse event streams, not only significantly reduces data transmission bandwidth requirements and computational complexity but also greatly enhances the processing power and real-time performance of edge computing nodes, providing efficient computational support for complex multi-source coordinated control. This invention also employs a cloud-edge collaborative architecture combined with a dynamic load allocation algorithm for intelligent allocation processing, automatically optimizing operating strategies based on grid demand response and real-time operating conditions. This achieves the dual goals of minimizing energy consumption and maximizing economic benefits. Furthermore, the combined application of digital twin technology and predictive maintenance algorithms enables a shift from passive maintenance to proactive prevention, significantly reducing equipment failure rates and maintenance costs, extending equipment lifespan, and providing reliable assurance for the long-term stable operation of substations. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0052] Figure 1 This invention provides a schematic flowchart of an intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station.

[0053] Figure 2 This invention provides a schematic diagram of the intelligent coordinated control system structure for a multi-source heat pump coupled dehumidification system in a power distribution station.

[0054] Figure reference numerals: 100, Analysis module; 200, Reconstruction module; 300, Loop module; 400, Control module; 500, Allocation module. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, embodiments of this invention, and should not be construed as limiting the invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. In the description of this invention, it should be understood that the terminology used is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0056] The embodiments of the present invention are described below with reference to the figures.

[0057] like Figure 1 As shown, this invention provides an intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station, comprising:

[0058] S1: Analyze the characteristics of multi-source heat pumps using a heat source quality assessment model to obtain heat source characteristic parameters.

[0059] The multi-source heat pump in step S1 includes a ground source heat pump, an air source heat pump, and a waste heat recovery heat pump from power distribution equipment.

[0060] Step S1 further includes:

[0061] S11: Collect basic heat source parameters for the ground source heat pump.

[0062] Furthermore, the process of acquiring the basic heat source parameters of the ground source heat pump is completed through multi-dimensional data acquisition using temperature sensor arrays, flow meters, and pressure sensors distributed in the vertical buried pipe system. In a specific embodiment, the temperature sensor uses a PT1000 platinum resistance thermometer, with one measuring point each at depths of 80 meters, 100 meters, and 120 meters underground, collecting underground soil temperature data once per minute to form a temperature time series; the flow meter is installed on the ground source heat pump circulating water pipeline to measure the volumetric flow rate of the circulating medium in real time; the pressure sensor monitors the pressure difference between the inlet and outlet of the circulating water; and temperature sensors are installed on the inlet and outlet water pipelines respectively to measure the inlet and outlet water temperatures; the data acquisition controller transmits these raw sensor data to the edge computing node via a CAN bus. The edge computing node first preprocesses the received data, specifically smoothing the temperature data. The preprocessed data is then input into the heat source quality assessment model, which calculates the instantaneous heating power of the ground source heat pump, as well as its circulation resistance coefficient and heat exchange efficiency. Subsequently, these calculation results are compared and analyzed with historical data to generate a basic heat source parameter matrix for the ground source heat pump. This matrix contains feature values ​​for four dimensions: heating power, heat exchange efficiency, circulation resistance, and temperature stability.

[0063] S12: Collect the heat source response of the air source heat pump within the preset ambient temperature range to obtain the peak heat source parameters of the air source heat pump.

[0064] Furthermore, the acquisition of peak heat source parameters for the air source heat pump involves processing performance test data of the variable frequency scroll compressor under different ambient temperatures and frequency operating conditions. An ambient temperature sensor collects outdoor temperature data every 2 minutes, the compressor controller records the current operating frequency, a power meter measures the actual power consumption of the compressor, and a refrigerant temperature sensor monitors the inlet and outlet temperatures of the evaporator. The data acquisition process covers all operating condition combinations within an ambient temperature range of -20℃ to 45℃ and compressor frequencies from 30Hz to 90Hz. Data is recorded after each operating condition point has been run for 30 minutes to reach a steady state. The raw data is transmitted to the data processing unit via a serial communication interface. The processing unit first establishes a three-dimensional data matrix of temperature, frequency, and performance, including the heating capacity calculated based on the temperature difference between the evaporator inlet and outlet and the refrigerant flow rate, the refrigerant mass flow rate, and the latent heat of vaporization. Subsequently, polynomial fitting is performed on the performance data within each temperature range to establish a functional relationship between frequency and heating capacity. The fitting coefficients are solved using the least squares method. Then, the entire data matrix is ​​traversed to find the frequency point corresponding to the maximum COP value at each temperature point. These peak points are connected to form the optimal operating curve. At the same time, the response time of each peak point is calculated, that is, the time required from the issuance of the frequency adjustment command to the heating capacity reaching a stable value. This is determined by analyzing the moment when the first derivative of the heating capacity change curve is zero. Finally, the peak heat source parameter matrix of the air source heat pump is output, integrating four key characteristic parameters: maximum heating capacity, optimal COP, corresponding frequency, and response time.

[0065] S13: Waste heat recovery from transformers and switchgear is achieved using micro heat pump technology, and waste heat recovery parameters for power distribution equipment are obtained.

[0066] Furthermore, in step S13, the present invention uses micro heat pump technology to collect and convert waste heat from transformers and switchgear to obtain waste heat recovery parameters for power distribution equipment. The surface temperature of the transformer is monitored by an infrared temperature sensor array, with the sensors arranged in a grid pattern, each grid measuring 10cm × 10cm, to collect the surface temperature distribution in real time; the internal temperature of the switchgear is monitored by a thermistor sensor, which is installed near the main heat-generating components, including circuit breakers, contactors, and busbar connections; the micro heat pump system uses thermoelectric cooler (TEC) technology to convert waste heat from the equipment into electrical energy or low-grade heat energy through the Peltier effect; the hot end of the thermoelectric cooler is in close contact with the heat-generating surface of the equipment, and the cold end is connected to the radiator, with the temperature difference driving the thermoelectric conversion process.

[0067] During the data acquisition process, the data acquisition controller records the temperature difference, output voltage, and output current across the thermoelectric cooler, and calculates the conversion power. Subsequently, the waste heat quality assessment algorithm calculates the total waste heat based on the temperature distribution data. The algorithm also analyzes the time stability of the waste heat temperature, assesses the stability of the heat source and the conversion efficiency of the micro heat pump by calculating the temperature variance, and then establishes a mapping relationship between the conversion efficiency data and the temperature difference data to form a conversion characteristic curve. Ultimately, the waste heat recovery parameters of the power distribution equipment include four core parameters: total waste heat, conversion power, conversion efficiency, and heat source stability. These parameters constitute the waste heat recovery feature vector, which is then input into the heat source quality assessment model.

[0068] S14: The heat source quality is evaluated by using artificial intelligence algorithms to assess the basic heat source parameters, the peak heat source parameters, and the waste heat recovery parameters of the power distribution equipment, thereby obtaining the heat source characteristic parameters.

[0069] In step S14 of this invention, a multilayer perceptron neural network structure is used to evaluate the quality of heat sources. The input layer of the neural network contains 12 nodes, corresponding to the characteristic parameters of three types of heat sources: the heating power, heat exchange efficiency, circulation resistance, and temperature stability of ground source heat pumps; the maximum heating capacity, optimal COP, response time, and frequency range of air source heat pumps; and the total waste heat, conversion power, conversion efficiency, and heat source stability of waste heat recovery from power distribution equipment. The hidden layer of the neural network adopts a two-layer structure, with the first layer containing 8 neurons and the second layer containing 6 neurons, and the ReLU function is selected as the activation function. The output layer contains 4 nodes, which output the quality score, economic score, reliability score, and comprehensive score for each type of heat source.

[0070] After training using historical operation records and simulation experiments as training datasets, the generated quality assessment model first normalizes the input data to ensure that all parameters are of the same magnitude. Then, it performs inference, calculates the output values ​​of each layer through forward propagation, and finally obtains four scoring results. The final output is the heat source characteristic parameters, including the results of combining the scoring results output by the neural network with real-time monitoring data, forming a comprehensive characteristic parameter data structure that includes heat source type, current state, performance parameters, quality rating, and recommended operation strategy.

[0071] S2: Dynamically reconstruct the heat exchange network topology based on the heat source characteristic parameters to obtain optimized heat exchange network configuration parameters.

[0072] Step S2 further includes:

[0073] S21: Based on the heat source characteristic parameters, the preset plate heat exchanger array is switched between series, parallel, and series-parallel hybrid modes to obtain the heat exchange network connection method.

[0074] In step S21, the plate heat exchanger array mode switching is dynamically configured based on the heat source quality rating and current load demand in the heat source characteristic parameters. The heat exchanger array consists of 6 plate heat exchangers, each connected by a smart three-way valve and an electric regulating valve to form a flexible piping network. The smart three-way valve is driven by a stepper motor, and the electric regulating valve uses proportional-integral control. The valve opening is automatically adjusted according to the flow demand. The inlet and outlet temperatures of the heat exchangers are monitored in real time by thermocouples, pressure sensors monitor changes in pipeline pressure, and flow meters measure the flow distribution in each branch.

[0075] Specifically, this invention first analyzes the real-time performance data of each heat source in the heat source characteristic parameters through a mode switching control algorithm, including the heating power of the ground source heat pump, the current COP value of the air source heat pump, and the available heat of the waste heat of the power distribution equipment. Then, based on the load forecast results and the current state of each heat source, it calculates the total heating capacity and energy consumption under different modes, selects the connection method with the highest overall efficiency, and the heat exchange network connection method data structure includes four core parameters: current mode type, valve opening degree, flow distribution ratio, and expected heat exchange effect.

[0076] Among them, the series mode is suitable for working conditions with large temperature differences and the need for step-by-step heating. The controller first passes the 55℃ heat medium output by the ground source heat pump through heat exchanger 1, and then exchanges heat with the 45℃ heat medium of the air source heat pump in heat exchanger 2. Finally, the 35℃ waste heat from the power distribution equipment is added in heat exchanger 3 to form a temperature cascade utilization link. The parallel mode is suitable for working conditions with large loads that require multiple heat sources to supply heat at the same time. The three heat sources supply heat to the load side at the same time through independent heat exchangers, and the flow rate of each branch is controlled by an electric regulating valve. The series-parallel hybrid mode combines the advantages of the two modes. The ground source heat pump and the air source heat pump are connected in series to obtain a higher outlet water temperature, and the waste heat from the power distribution equipment is connected in parallel to supplement the total heat.

[0077] S22: Based on the heat exchange network connection method, the mixing ratio of heat media at different temperature levels is calculated by using a heat medium mixing optimization algorithm to generate target temperature adjustment parameters and phase change trigger commands.

[0078] Furthermore, in step S22, the following are first inputs: the heat medium flow rate of the ground source heat pump at 55°C, the heat medium flow rate of the air source heat pump at 45°C, the heat medium flow rate of the waste heat from the power distribution equipment at 35°C, and the target outlet water temperature. Then, the mixing ratio is calculated and solved using a set of heat balance equations. The algorithm considers both pipeline resistance loss and heat exchanger heat transfer loss, and establishes a modified equation. The objective function for optimization is set to minimize the total flow rate and the temperature deviation. The optimal mixing ratio is solved using the Lagrange multiplier method.

[0079] The phase change trigger command generation is based on the topology switching requirements of the heat exchange network. When a load change exceeding a set threshold or a heat source state switch is detected, the algorithm calculates the heat loss risk during the topology switching process. The phase change material capsule triggering algorithm calculates the required heat storage based on the current heat medium temperature and the target temperature. The generated trigger command includes the target area coordinates, microwave power setpoint, radiation time, and expected phase change completion time. The final target temperature regulation parameter data structure integrates the mixing ratio matrix, flow distribution scheme, temperature setpoint, and control accuracy requirements. The phase change trigger command includes spatial coordinates, power parameters, time parameters, and status feedback requirements.

[0080] S23: In response to the phase change trigger command, microwave radiation is applied to the embedded phase change material capsule to maintain heat flow continuity during topology switching and obtain heat flow buffer queue parameters.

[0081] Step S23 further includes:

[0082] S231: Embed composite phase change material capsules into the gaps of the flow channels in a plate heat exchanger.

[0083] In one specific embodiment, the flow channel gap of the plate heat exchanger is 3-5mm, so the diameter of the phase change material capsule of the present invention is designed to be 50-100μm to ensure that the capsule will not affect normal flow. The capsule material is a paraffin-based composite phase change material with a melting point designed in the range of 45-55℃ and a latent heat value of about 200kJ / kg.

[0084] During the embedding process, microfluidic injection technology is used to inject the capsule suspension into the flow channel gaps through micro-injection orifices. The capsules in the flow channel gaps are distributed in a grid pattern, with each grid measuring 5mm × 5mm. The capsule concentration is controlled at 1000-2000 capsules per milliliter. The capsules are preferentially placed in areas with lower flow rates and larger temperature gradients to avoid capsule breakage in areas with high shear stress. Capsule fixation is achieved through surface modification technology, coating the capsule surface with a hydrophilic polymer layer to enhance adhesion to the heat exchanger surface. After embedding, the uniformity of capsule distribution is checked using an endoscope to ensure that the capsule density in each area meets the design requirements.

[0085] S232: Selective excitation of phase change material capsules in the target region using a microwave radiation array.

[0086] In step S232, the present invention selectively excites the phase change material capsule using a 2.45GHz industrial frequency microwave source and achieves spatial selective heating based on a phased array antenna. The microwave radiation array contains 16 microwave transmitting units, each with a power of 50W, and beam focusing is achieved through phase modulation.

[0087] In the selective excitation process, the present invention first determines the coordinates of the target heating area according to the phase change trigger command, and then calculates the phase and power distribution of each transmitting unit. Since the phase change material capsule has good dielectric loss characteristics for 2.45GHz microwaves, with a dielectric constant of 3.2 and a dielectric loss factor of 0.8, the phase change material heats up rapidly under the action of microwaves. During the excitation process, it is necessary to monitor the temperature change of the target area in real time. When the temperature reaches the phase change point, the microwave power is automatically adjusted to maintain a constant phase change rate, and finally the selective excitation is completed.

[0088] S223: Real-time monitoring of changes in the optical transmittance of the phase change material capsule, determining complete phase change based on the decrease in transmittance, and generating a signal to confirm the continuity of heat flow.

[0089] Furthermore, optical transmittance monitoring achieves real-time detection of the phase transition state based on Lambert-Beer's law, utilizing the significant changes in the optical properties of the phase change material during solid-liquid conversion to determine its state. The optical monitoring device includes an LED light source and a photodiode receiver. The LED emits a beam of light of a specific wavelength that penetrates the phase change material capsule, and the receiver measures the intensity of the transmitted light. During monitoring, the monitoring algorithm performs real-time analysis of the transmittance data, calculates the transmittance change gradient, and determines the start of the phase transition when the gradient exceeds a preset threshold. Then, based on transmittance stability analysis, a complete phase transition is determined. When the transmittance change is less than 2% within 10 consecutive seconds, the phase transition is determined to be complete. Finally, the output heat flow continuity maintenance confirmation signal is generated based on the comprehensive analysis of multi-point transmittance monitoring results. A confirmation signal is generated when more than 80% of the monitoring points in the target area have completed the phase transition. The signal format includes four data fields: timestamp, area coordinates, phase transition completion degree, and expected heat storage.

[0090] S234: Update the heat flow buffer queue parameters according to the heat flow continuity maintenance confirmation signal to obtain the heat flow buffer queue parameters.

[0091] The heat flow buffer queue parameters are dynamically adjusted based on the heat flow continuity confirmation signal. The queue parameters include four core elements: heat storage capacity, heat release rate, temperature distribution, and available time. The queue update algorithm first reads the phase change completion data from the confirmation signal, calculates the number of capsules that have actually completed phase change, then calculates the heat storage capacity based on the mass of the capsules that have completed phase change and the latent heat of phase change. Subsequently, it considers the convective heat transfer coefficient and temperature difference within the heat exchanger channels to calculate the heat release rate. After obtaining the above parameters, the temperature distribution is updated, and the temperature field distribution of the entire area is extrapolated based on the temperature data from the monitoring points. The queue management algorithm adopts a first-in, first-out (FIFO) principle, prioritizing the use of phase change material capsules with the longest heat storage time to ensure effective heat utilization. The parameter update process simultaneously records historical data, establishing a phase change performance database. The final output heat flow buffer queue parameter data structure integrates real-time heat storage status, available heat, expected release time, and control priority information.

[0092] S24: Using the highest overall system efficiency and lowest energy consumption as objective functions, the particle swarm optimization algorithm is used to dynamically solve the parameters of the intelligent three-way valve and the electric regulating valve to obtain the optimized heat exchange network configuration parameters.

[0093] The Particle Swarm Optimization (PSO) algorithm solves for the parameters of a smart three-way valve and an electric regulating valve with the objective functions of maximizing overall system efficiency and minimizing energy consumption. The algorithm initializes a population of 50 particles, each representing a combination of valve openings. The particle dimension includes 6 three-way valve openings and 12 regulating valve openings, totaling 18 decision variables. The objective function is designed as a multi-objective optimization problem, including overall system efficiency and total power consumption. Constraints include flow balance, pressure limits, and temperature range. The global optimum is searched by comparing the fitness values ​​of all particles. The algorithm terminates when there is no improvement in the global optimum for 20 consecutive generations. The final output is the optimized heat exchange network configuration parameters, including the optimal opening setpoints for each valve, expected flow distribution, system efficiency indicators, and predicted energy consumption.

[0094] S3: The dehumidifying working fluid is circulated in stages through a multi-effect evaporator to obtain the dehumidifying working fluid circulation parameters.

[0095] Step S3 further includes:

[0096] S31: By using a triple-effect evaporator series structure, the environmentally friendly working fluid R1234yf is evaporated in stages at multiple temperature levels to obtain multi-effect evaporation parameters.

[0097] In step S31, the present invention aims to use a three-effect evaporator series structure to achieve staged evaporation treatment of environmentally friendly working fluid R1234yf by inputting heat sources of different temperature levels in the optimized heat exchange network configuration parameters into the first, second and third effect evaporators respectively.

[0098] The first-effect evaporator receives a 55-60℃ heat transfer medium from a ground-source heat pump as its heating source. The working fluid, R1234yf, enters the evaporator at a pressure of 8.5 bar. The heat transfer medium transfers heat to the working fluid through the tube bundle, rapidly raising its temperature from a liquid state of 25℃ to its boiling point of 48℃, initiating evaporation. During operation, a temperature sensor monitors the working fluid temperature in real time. When the temperature reaches 48℃, a level sensor detects a drop in the liquid level, and a pressure sensor detects that the evaporation pressure has risen to the set value of 8.5 bar, indicating that the first-effect evaporation process is proceeding normally. The steam generated by the first-effect evaporation enters the condenser through pipes for partial condensation. The latent heat generated during condensation directly serves as the heat source for the second-effect evaporator. The second-effect evaporator receives the condensation heat from the first effect and an auxiliary heat source of 45-50℃ provided by the air-source heat pump. The working fluid evaporates at a lower pressure of 6.2 bar, with the evaporation temperature dropping to 42℃. The third-effect evaporator utilizes the condensation heat from the second effect and the waste heat from the power distribution equipment to recover a 35-40℃ heat source, and the working fluid evaporates at 36℃ under a pressure of 4.1 bar.

[0099] Each evaporator is equipped with an independent flow meter to measure the working fluid flow rate, a temperature sensor to monitor the evaporation temperature, and a pressure sensor to detect the evaporation pressure. This sensor data is collected by a data acquisition module and transmitted to the control processor. The control processor calculates the evaporation efficiency based on the operating parameters of each evaporator effect. The first-effect evaporation efficiency is calculated as the ratio of evaporation rate to input heat; the second-effect evaporation efficiency is determined based on the condensation heat utilization rate of the first effect; and the third-effect evaporation efficiency is obtained through an assessment of waste heat utilization. The final multi-effect evaporation parameter data structure includes five core parameters for each evaporator effect: evaporation temperature, evaporation pressure, working fluid flow rate, evaporation efficiency, and total evaporation rate. These parameters constitute the basic data for the cascade dehumidification working fluid circulation.

[0100] S32: Based on the multi-effect evaporation parameters, high-temperature steam, condensate and low-temperature steam are used for cascade energy utilization to obtain parameters for maximizing energy utilization.

[0101] Furthermore, the high-temperature steam originates from the first-effect evaporator, with a temperature of 48°C and a pressure of 8.5 bar. The steam quality (dryness fraction) is determined to be 0.95 by referring to a table based on steam temperature and pressure, indicating that the steam contains 5% liquid component. This high-temperature steam first enters the first-stage condenser, releasing latent heat through heat exchange with external cooling water. During condensation, the temperature remains constant, but the pressure gradually decreases to 6.8 bar. A humidity sensor installed at the condenser outlet detects the condensate content, a temperature sensor monitors the condensation temperature, and a pressure sensor measures the condensation pressure. The condensate produced in the first-stage condenser has a temperature of 48°C and contains sensible heat energy. This sensible heat is transferred to the feed medium of the second-effect evaporator via a heat exchanger, achieving secondary utilization of thermal energy. Simultaneously, the latent heat released by the first-stage condenser is collected by a heat recovery device as a supplementary heat source for the second-effect evaporator. The medium-temperature steam produced in the second effect has a temperature of 42°C, a pressure of 6.2 bar, and a steam quality fraction of 0.92. Through a similar condensation process, it produces medium-temperature condensate and medium-quality latent heat. The low-temperature steam originates from the third-effect evaporator, with a temperature of 36°C, a pressure of 4.1 bar, and a steam quality of 0.88. Although the temperature is low, it still contains usable latent heat energy. This invention uses a cascade utilization control algorithm to calculate the total recoverable heat energy based on the enthalpy of each stage of steam. The recoverable heat energy of the first effect is calculated based on steam flow rate and latent heat of vaporization. The recoverable heat energy of the second and third effects is calculated using the same method. The obtained energy maximization utilization parameters integrate the heat energy quality of each stage of steam, sensible heat of condensate, latent heat recovery, and overall energy efficiency ratio, forming a complete cascade energy utilization data matrix.

[0102] S33: The plate heat exchanger recovers the latent heat of the exhaust steam and preheats the working fluid entering the evaporator to obtain the dehumidifying working fluid circulation parameters.

[0103] In step S33, the present invention utilizes a plate heat exchanger for heat recovery, aiming to recover the latent heat of the emitted steam through efficient counter-current heat exchange and preheat the fresh working fluid entering the evaporator. In a specific embodiment, the plate heat exchanger employs a 316L stainless steel corrugated plate structure with a plate spacing of 3mm, a heat exchange area of ​​15 square meters, and a designed heat transfer coefficient of 3200W / (m²). 2The exhaust steam, at 36°C and 4.1 bar, is drawn from the outlet of the third-effect evaporator and enters the hot-side channel of the plate heat exchanger through a pipe. Fresh working fluid R1234yf is pumped from a storage tank into the cold-side channel of the heat exchanger, with an inlet temperature of 25°C and a pressure of 12 bar. During heat exchange, the exhaust steam condenses on the plate surface, releasing latent heat. This latent heat is transferred to the fresh working fluid on the cold side through the plates, gradually increasing the working fluid temperature. At the heat exchanger outlet, the exhaust steam completely condenses into a liquid, reducing its temperature to 32°C, while the fresh working fluid temperature rises to 40°C, achieving effective heat transfer. The heat transfer calculation is based on the heat transfer rate equation. ,in The heat transfer coefficient is... For heat exchange area, The preheating effect is evaluated by the temperature rise of the working fluid, which is the logarithmic mean temperature difference. and preheating efficiency Quantization representation, in which The inlet steam temperature on the hot side of the heat exchanger. This refers to the inlet liquid temperature on the cold side of the heat exchanger. The heat exchanger outlet liquid temperature and the final dehumidifying working fluid circulation parameter data structure include five dimensions of feature data: preheating temperature, heat recovery amount, circulation efficiency, pressure distribution, and working fluid quality. These parameters directly affect the subsequent control decisions of the biomimetic pulse neural network.

[0104] S4: The dehumidifying working fluid circulation parameters are controlled by a biomimetic spiking neural network to obtain humidity control instructions.

[0105] Step S4 further includes:

[0106] S41: By using a biomimetic spiking neural network to perform neuromorphic modeling of the temperature, humidity and equipment operating status of the substation environment, a spiking neuron control architecture is obtained.

[0107] Furthermore, the modeling process in step S41 aims to establish a data structure for the spiking neuron control architecture, integrating neuron parameter matrices, synaptic weight matrices, delay matrices, and threshold configurations to form a complete neuromorphic control framework.

[0108] Step S41 further includes:

[0109] S411: Map the ground source heat pump, air source heat pump, and waste heat recovery heat pump from power distribution equipment to independent pulse neurons, and obtain the topology of the three-source neuron network.

[0110] Furthermore, in step S411, this invention aims to establish a distributed control network architecture based on a three-source neural network topology. This is achieved by mapping the ground source heat pump, air source heat pump, and waste heat recovery heat pump from the power distribution equipment to independent pulse neuron groups. The neural group corresponding to the ground source heat pump contains six neurons, each handling parameters such as inlet water temperature, outlet water temperature, circulation flow rate, power consumption, operating frequency, and efficiency. Each neuron uses the same LIF model but has different time constants and threshold parameters. In a specific implementation, the time constant of the ground source heat pump temperature neuron is set to 15ms, and the threshold potential is -55mV, reflecting the relatively slow temperature change characteristic of the ground source heat pump. The air source heat pump neural group contains eight neurons, handling parameters such as ambient temperature, evaporation temperature, condensation temperature, compressor frequency, refrigerant pressure, power, COP, and defrost status. In a specific implementation, the time constant of the air source heat pump neuron is set to 8ms, and the threshold potential is set to -50mV, reflecting the rapid response characteristics of the air source heat pump to environmental changes. The waste heat recovery of power distribution equipment corresponds to 4 neurons, which respectively process the parameters of transformer temperature, switch cabinet temperature, waste heat power and conversion efficiency. In the specific implementation, the time constant is set to 25ms and the threshold potential is set to -60mV to reflect the relatively stable characteristics of waste heat changes.

[0111] The neural network topology adopts a fully connected structure, where each neuron establishes synaptic connections with other neurons, but the connection strength is differentiated based on physical correlation. The connection weights between neurons in the ground source heat pump and air source heat pump are higher because they have a coordinated relationship in terms of heat supply. The connection weights between neurons in the power distribution equipment waste heat source are relatively lower, reflecting their role as an auxiliary heat source. In addition, the network topology includes inhibitory connections to prevent energy waste caused by the simultaneous overactivation of multiple heat sources. The three-source neural network topology data structure output in the final step S411 records the parameter configuration, connection matrix, activation mode, and interaction rules of each neuron group, forming the basic architecture of distributed control.

[0112] S412: The connection strength of the three-source neuron network topology is set by the synaptic weight matrix to obtain the coupling parameters between neurons.

[0113] The synaptic weight matrix setting process in step S412 of this invention aims to determine the connection strength parameters between neurons by analyzing the functional correlation and control priority among neurons in the three-source neuron network topology. Specifically, the weight matrix established in this invention is an 18×18 square matrix, with the values ​​of the matrix elements ranging from -1 to 1. Positive values ​​represent excitatory connections, and negative values ​​represent inhibitory connections. The weights between the ground source heat pump temperature neurons and the air source heat pump temperature neurons are set based on the principle of temperature coordination. When the ground source heat pump temperature is low, the output of the air source heat pump needs to be increased, so a positive weight of 0.65 is set. An inhibitory connection is set between the air source heat pump power neurons and the ground source heat pump power neurons, with a weight of -0.45, to prevent energy waste caused by the simultaneous high-power operation of the two main heat sources. A small positive weight of 0.25-0.35 is set between the waste heat neurons of the power distribution equipment and the main heat source neurons to reflect the coordinating role of waste heat as a supplementary heat source.

[0114] Furthermore, the weight matrix setting also considers the reliability and response time differences of sensor data. Temperature sensor data has a higher weight than pressure sensor data, and directly measured parameters have a higher weight than calculated parameters. During weight learning, this invention employs a variant of the Hebbian learning rule, strengthening the connection weight when two neurons are frequently activated simultaneously and weakening the connection weight when activation patterns are mismatched. The final output inter-neuron coupling parameter data structure integrates the weight matrix, delay matrix, learning parameters, and connection types, providing complete connection information for pulse propagation.

[0115] S413: Optimize the pulse propagation path according to the interneuron coupling parameters to obtain the neuromorphic control architecture.

[0116] In step S413, this invention aims to optimize the propagation path of pulses in a neural network based on the coupling parameters between neurons, by analyzing network connectivity and information flow efficiency. Specifically, a network connectivity graph is first constructed, with neurons as nodes, synaptic connections as edges, and edge weights representing connection strengths. Then, path optimization is performed using an improved version of the shortest path algorithm, considering not only the shortest propagation distance but also the accuracy and timeliness of information propagation. The propagation path from the input neuron to the output neuron is calculated using Dijkstra's algorithm. During the optimization process, this invention identifies certain key neurons as bottleneck nodes in information propagation. Therefore, this invention alleviates the bottleneck effect by adding parallel propagation paths. Parallel path construction is achieved by replicating key neurons or adding backup connections, ensuring that information can still propagate normally even when some connections fail. The path optimization algorithm adjusts the propagation timing according to the refractory period characteristics of each neuron to prevent key information from arriving within the neuron's refractory period. The optimized propagation path has multiple parallel channels: the main channel propagates key control information, and the auxiliary channels propagate state feedback and anomaly detection information. The final neuromorphic control architecture data structure records the optimized propagation path, key node identifiers, parallel channel configurations, and feedback loop parameters, forming a complete pulse propagation network.

[0117] S42: Using the STDP learning rule, the spatiotemporal pulse sequence of the spiking neuron control architecture is adaptively learned to obtain distributed decision parameters.

[0118] In step S42, this invention employs the STDP (Spike-Timing-Dependent Plasticity) learning rule, based on the biological principle that the first synaptic neuron is strengthened and the later synaptic neuron is weakened. When the current synaptic neuron is activated before the subsequent synaptic neuron, the connection weight is enhanced; conversely, the weight is weakened, thus performing adaptive learning of the spatiotemporal spike sequence of the spiking neuron control architecture.

[0119] First, spatiotemporal pulse sequence analysis is performed. Specifically, the firing timestamps of each neuron are recorded to construct time-series data. Then, a double exponential decay function is used as the learning window function: when... hour: ,when 0:00 ,in For the time difference, and The magnitude of the weight change, and The time constant is used; subsequently, all connected neuron pairs are traversed, the temporal relationship is calculated, and the weights are updated incrementally. At the same time, weight boundary constraints are set to prevent infinite increase or decrease, and a synaptic competition mechanism is introduced, where the presynapse with the closest activation time receives the maximum weight enhancement. During the learning process, the characteristics of neuronal activity frequency are considered, and high-frequency activation of connections results in stronger weight enhancement. A forgetting mechanism is also included, where the weights of connections that have not been activated for a long time gradually decay, and the decay rate is inversely proportional to the activation frequency. Finally, distributed decision parameters are output, including the updated weight matrix, learning history, activation frequency statistics, and decision priority ranking.

[0120] S43: The distributed decision parameters are encoded by a sparse pulse event stream using an FPGA-accelerated pulse encoder to obtain a sub-second response control signal.

[0121] Specifically, the hardware platform is first initialized. This invention uses a Xilinx Zynq-7000 series FPGA chip, which includes a dual-core ARM processor and programmable logic resources, supporting 128 parallel encoding channels. After initializing the hardware platform, data preprocessing and discretization are performed. The FPGA first preprocesses the distributed decision parameters, discretizing the continuous weight values ​​and activation intensities into pulse frequencies and amplitudes. A 1ms time window is used for discretization, and the pulse firing probability is determined based on the neuron activation intensity within each time slice. Subsequently, sparse encoding is performed. This invention implements sparse encoding through an event-driven mechanism, generating pulse events only when the neuron activation intensity exceeds a threshold. Each pulse event contains... The pulse event stream is encoded using a fixed 32-bit format, consisting of four fields: pulse ID, excitation timestamp, pulse amplitude, and duration. After encoding, parallel processing and data compression are performed. Parallel processing utilizes 128 parallel encoding channels in a pipelined manner to improve throughput. For data compression, this invention integrates a data compression algorithm that leverages the sparsity of the pulse sequence for lossless compression, employing a hybrid approach combining run-length encoding and dictionary encoding. The compressed pulse event stream is transmitted via a high-speed serial interface using the Aurora protocol, achieving a transmission rate of 10Gbps. The event stream includes synchronization markers and error detection codes to ensure data transmission accuracy and timing synchronization. Ultimately, the processing latency is reduced from milliseconds to microseconds, the data transmission volume is reduced to one-tenth of the original, and a sub-second response control signal is output.

[0122] S44: Perform neuromorphic adjustment processing on the operating status of each dehumidification module according to the sub-second response control signal to obtain the humidity control command.

[0123] In step S44, signal decoding and conversion are first performed. The sub-second response control signal is first converted into an analog control quantity by a pulse decoder, and the integrator circuit converts the pulse frequency into a voltage signal. The voltage amplitude is proportional to the power setting of the dehumidification module. Then, the dehumidification module performs hierarchical control. The dehumidification module includes a triple-effect evaporator, a circulating pump, a regulating valve, and a fan. Each device corresponds to an independent neuromorphic control channel: the first-effect dehumidification module receives the control signal from the ground source heat pump neuron group, with a power adjustment range of 20kW-80kW and an adjustment accuracy of 1kW. The second and third-effect dehumidification modules receive the control signals from the air source heat pump and waste heat recovery neuron groups, respectively, to achieve hierarchical power control. Then, the control algorithm automatically adjusts the control strategy according to changes in environmental parameters by modifying the neuron threshold and connection weight. At the same time, the sensor network collects feedback parameters such as evaporation temperature, pressure, flow rate, and humidity in real time to form a closed-loop control. Finally, a predictive control algorithm is used to predict the future system behavior based on the current state, adjust the control parameters in advance, and analyze the historical control effect based on the reinforcement learning principle, adjusting the neural network parameters according to the results. The final output humidity control command integrates the power settings, operating modes, safety limits, and performance expectations of each dehumidification module.

[0124] S5: Dynamically allocate the load of the multi-source heat pump system according to the humidity control command to obtain an energy efficiency coordination optimization strategy, and coordinate the control of the multi-source heat pump through the energy efficiency coordination optimization strategy.

[0125] Step S5 further includes:

[0126] S51: Through a cloud-edge collaborative architecture, the COP (Coefficient of Performance) and SEER (Seasonal Energy Efficiency Ratio) of multiple heat pump modules are calculated in real time to obtain energy efficiency evaluation parameters.

[0127] Furthermore, in step S51, the edge-cloud architecture is first deployed. This invention employs a collaborative approach between edge computing nodes and cloud servers. Edge nodes are deployed at the substation site, using ARM Cortex-A72 processors for data acquisition, preprocessing, and preliminary calculations. The cloud server utilizes a GPU acceleration platform for complex data analysis and global optimization. The architecture is connected via a 5G network. Edge nodes upload aggregated data every minute, while the cloud distributes optimization parameters and model updates hourly. It also includes a fault-tolerant mechanism; edge nodes operate independently during network interruptions and automatically synchronize upon recovery. After deployment, real-time COP calculation is performed. Specifically, the edge nodes collect power consumption measured by a power meter and heat generation measured by a flow meter / temperature sensor every second, calculating the coefficient of performance (COP) based on their ratio. Subsequently, the cloud server calculates the seasonal energy efficiency ratio using the formula: ,in For the heating capacity of the i-th operating condition, For the corresponding power consumption, This refers to the number of operating hours per year.

[0128] S52: Based on the energy efficiency assessment parameters and the humidity control command, intelligently allocate the dehumidification load to obtain a load allocation scheme.

[0129] In step S52, the present invention first analyzes the aforementioned humidity control command, including the target humidity value, current humidity value, trend of change, and urgency level, and calculates the dehumidification load demand. Then, combined with energy efficiency evaluation parameters, a multi-objective optimization method is adopted to establish an objective function including minimizing total energy consumption, minimizing response time, and minimizing equipment wear. Priority is determined according to the current COP value of each heat source, with higher COP values ​​bearing more load. Subsequently, a graded load allocation strategy is established. For ground source heat pumps, the COP value is stable and relatively high, bearing 60-70% of the basic load. For air source heat pumps, the allocation ratio is dynamically adjusted according to the ambient temperature. When the temperature is high, the COP decreases, and the allocation ratio decreases. For waste heat recovery of power distribution equipment, the cost is the lowest, and full-load operation is prioritized when available. After establishing the objective function, a dynamic programming method is used to solve the optimal allocation scheme. The state space includes the operating status and load level of each heat source, and the action space includes load increase / decrease and start / stop control. At the same time, it is ensured that the load allocation of each heat source does not exceed the equipment capacity limit and meets the total load demand of the system. Finally, the load allocation scheme is output, including the allocation matrix, priority ranking, constraints, and expected performance indicators.

[0130] S53: The power grid demand response function accesses electricity market information and participates in power grid peak shaving and frequency regulation services to obtain the energy efficiency coordination and optimization strategy.

[0131] In step S53 of this invention, the power dispatch center is connected via the IEC 61850 communication protocol to obtain real-time electricity price, load forecast and peak shaving and frequency regulation demand information. The information includes time-of-use electricity price, load curve, interruptible load compensation and ancillary service demand, and is updated every 15 minutes. Then, the heat pump operation strategy is adjusted according to the electricity price signal. During periods of low electricity prices, the heat pump operating power is increased, and excess heat energy is stored using phase change material heat storage devices. During periods of high electricity prices, the electricity load is reduced, and the stored heat energy is released to meet dehumidification needs. At the same time, peak shaving and frequency regulation services are also provided. When the grid needs peak shaving, the power of non-critical equipment is automatically reduced, with the peak shaving capacity reaching 30% of the total load. The heat pump frequency conversion control response time is less than 4 seconds, meeting the primary frequency regulation requirements of the grid. Settlement is conducted through the electricity market based on the provided peak shaving and frequency regulation capacity and the actual implementation effect. In addition, this invention also performs load forecasting. By analyzing historical electricity consumption patterns and weather forecast data, the ARIMA model combined with neural networks is used to predict the electricity demand curve for the next 24 hours. The final output energy efficiency coordination optimization strategy integrates the results of grid demand response, load allocation, and energy efficiency assessment, and achieves system-level optimal control through a multi-layer optimization architecture.

[0132] like Figure 2 As shown, the present invention also provides an intelligent coordination control system for a multi-source heat pump coupled dehumidification system in a power distribution station, comprising:

[0133] Analysis module 100: Used to perform multi-source heat pump characteristic analysis through heat source quality assessment model to obtain heat source characteristic parameters;

[0134] Reconstruction module 200: used to dynamically reconstruct the heat exchange network topology according to the heat source characteristic parameters, and obtain optimized heat exchange network configuration parameters;

[0135] Circulation module 300: used to perform cascade dehumidification working fluid circulation through the multi-effect evaporator to obtain dehumidification working fluid circulation parameters by circulating the optimized heat exchange network configuration parameters in stages;

[0136] Control module 400: used to perform neuromorphic control on the circulation parameters of the dehumidifying working fluid through a biomimetic pulse neural network to obtain humidity control commands;

[0137] Distribution module 500: used to dynamically distribute the load of the multi-source heat pump system according to the humidity control command, obtain an energy efficiency coordination optimization strategy, and coordinate the control of the multi-source heat pump through the energy efficiency coordination optimization strategy.

[0138] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0139] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the intelligent coordinated control method of a multi-source heat pump coupled dehumidification system in a power distribution room as described in various embodiments or some parts of embodiments.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent coordinated control of a multi-source heat pump coupled dehumidification system in a power distribution station, characterized in that, include: S1: Analyze the characteristics of multi-source heat pumps using a heat source quality assessment model to obtain heat source characteristic parameters; S2: Dynamically reconstruct the heat exchange network topology based on the heat source characteristic parameters to obtain optimized heat exchange network configuration parameters; the optimized heat exchange network configuration parameters include: optimal opening setpoints for multiple valves, expected flow distribution, system efficiency indicators, and energy consumption prediction values; S3: The optimized heat exchange network configuration parameters are circulated in stages using a multi-effect evaporator to obtain dehumidifying working fluid circulation parameters; the dehumidifying working fluid circulation parameters include: preheating temperature, heat recovery rate, circulation efficiency, pressure distribution data, and working fluid quality data; S4: Obtain humidity control instructions by performing neuromorphic control on the dehumidifying working fluid circulation parameters through a biomimetic pulse neural network; S5: Dynamically allocate the load of the multi-source heat pump system according to the humidity control command to obtain an energy efficiency coordination optimization strategy, and coordinate the control of the multi-source heat pump through the energy efficiency coordination optimization strategy.

2. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 1, characterized in that, The multi-source heat pump in step S1 includes a ground source heat pump, an air source heat pump, and a waste heat recovery heat pump from power distribution equipment.

3. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 2, characterized in that, Step S1 further includes: S11: Collect basic heat source parameters for the ground source heat pump; S12: Collect the heat source response of the air source heat pump within the preset ambient temperature range to obtain the peak heat source parameters of the air source heat pump; S13: Waste heat recovery from transformer and switchgear operation is carried out using micro heat pump technology to obtain waste heat recovery parameters of power distribution equipment; S14: The heat source quality is evaluated by using artificial intelligence algorithms to assess the basic heat source parameters, the peak heat source parameters, and the waste heat recovery parameters of the power distribution equipment, thereby obtaining the heat source characteristic parameters.

4. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 1, characterized in that, Step S2 further includes: S21: Based on the heat source characteristic parameters, the preset plate heat exchanger array is switched between series, parallel, and series-parallel hybrid modes to obtain the heat exchange network connection method; S22: Based on the heat exchange network connection method, the mixing ratio of heat media at different temperature levels is calculated by the heat medium mixing optimization algorithm to generate target temperature adjustment parameters and phase change triggering commands. S23: In response to the phase change trigger command, microwave radiation is applied to the embedded phase change material capsule to maintain heat flow continuity during topology switching and obtain heat flow buffer queue parameters; S24: Using the highest overall system efficiency and lowest energy consumption as objective functions, the particle swarm optimization algorithm is used to dynamically solve the parameters of the intelligent three-way valve and the electric regulating valve to obtain the optimized heat exchange network configuration parameters.

5. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 4, characterized in that, Step S23 further includes: S231: Embed composite phase change material capsules into the gaps of the flow channels in a plate heat exchanger; S232: Selective excitation of phase change material capsules in the target region using a microwave radiation array; S223: Real-time monitoring of optical transmittance changes in phase change material capsules, determining complete phase change based on transmittance decrease, and generating a heat flow continuity maintenance confirmation signal; S234: Update the heat flow buffer queue parameters according to the heat flow continuity maintenance confirmation signal to obtain the heat flow buffer queue parameters.

6. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 1, characterized in that, Step S3 further includes: S31: By using a triple-effect evaporator series structure, the environmentally friendly working fluid R1234yf is evaporated in stages at multiple temperature levels to obtain multi-effect evaporation parameters; S32: Based on the multi-effect evaporation parameters, high-temperature steam, condensate and low-temperature steam are used for cascaded energy utilization to obtain parameters for maximizing energy utilization. S33: The plate heat exchanger recovers the latent heat of the exhaust steam and preheats the working fluid entering the evaporator to obtain the dehumidifying working fluid circulation parameters.

7. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 1, characterized in that, Step S4 further includes: S41: By using a biomimetic spiking neural network to perform neuromorphic modeling of the temperature, humidity and equipment operating status of the substation environment, a spiking neuron control architecture is obtained. S42: By using the STDP learning rule, the spatiotemporal pulse sequence of the spiking neuron control architecture is adaptively learned to obtain distributed decision parameters; S43: The distributed decision parameters are encoded by a sparse pulse event stream using an FPGA-accelerated pulse encoder to obtain a sub-second response control signal; S44: Perform neuromorphic adjustment processing on the operating status of each dehumidification module according to the sub-second response control signal to obtain the humidity control command.

8. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 7, characterized in that, Step S41 further includes: S411: Map the ground source heat pump, air source heat pump and power distribution equipment waste heat recovery heat pump to independent pulse neurons, and obtain the topology of the three-source neuron network. S412: Set the connection strength of the three-source neuron network topology through the synaptic weight matrix to obtain the coupling parameters between neurons; S413: Optimize the pulse propagation path according to the interneuron coupling parameters to obtain the neuromorphic control architecture.

9. The intelligent coordinated control method for a multi-source heat pump coupled dehumidification system in a power distribution station according to claim 1, characterized in that, Step S5 further includes: S51: Through a cloud-edge collaborative architecture, the COP (Coefficient of Performance) and SEER (Seasonal Energy Efficiency Ratio) of multiple heat pump modules are calculated in real time to obtain energy efficiency evaluation parameters. S52: Based on the energy efficiency assessment parameters and the humidity control command, intelligently allocate the dehumidification load to obtain a load allocation scheme; S53: The power grid demand response function accesses electricity market information and participates in power grid peak shaving and frequency regulation services to obtain the energy efficiency coordination and optimization strategy.

10. An intelligent coordinated control system for a multi-source heat pump coupled dehumidification system in a power distribution room, characterized in that, include: Analysis module: Used to perform multi-source heat pump characteristic analysis using a heat source quality assessment model to obtain heat source characteristic parameters; Reconstruction module: used to dynamically reconstruct the heat exchange network topology based on the heat source characteristic parameters to obtain optimized heat exchange network configuration parameters; The optimized heat exchange network configuration parameters include: optimal valve opening settings, expected flow distribution, system efficiency indicators, and energy consumption predictions. Circulation module: used to perform cascaded dehumidification working fluid circulation through the multi-effect evaporator to obtain dehumidification working fluid circulation parameters; the dehumidification working fluid circulation parameters include: preheating temperature, heat recovery rate, circulation efficiency, pressure distribution data and working fluid quality data; Control module: used to perform neuromorphic control on the circulation parameters of the dehumidifying working fluid through a biomimetic pulse neural network to obtain humidity control commands; Distribution module: used to dynamically distribute the load of the multi-source heat pump system according to the humidity control command, obtain the energy efficiency coordination optimization strategy, and coordinate the control of the multi-source heat pump through the energy efficiency coordination optimization strategy.

Images