Hydrotherapy equipment energy consumption optimization method based on reinforcement learning

Abstract

The invention discloses a reinforcement learning-based hydrotherapy equipment energy consumption optimization method, and relates to the technical field of hydrotherapy equipment energy consumption control, and the method comprises the steps: processing real-time energy consumption data through a wavelet decomposition algorithm, extracting transient energy consumption characteristics, and querying an equipment operation mode database in combination with a current harmonic distortion rate to determine an initial energy consumption optimization strategy; according to the method, the energy consumption optimization strategy is dynamically adjusted based on reinforcement learning, corresponding strategy parameters and an energy consumption income estimation model are obtained, income estimation analysis is carried out on all candidate strategies, meanwhile, macro-micro action hierarchical data and component corrosion degree data of equipment are considered, decision analysis is carried out, and the optimal energy consumption optimization strategy is screened out. The energy utilization efficiency is improved and the operation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of energy consumption control of hydrotherapy equipment, and more specifically, to a method for optimizing energy consumption of hydrotherapy equipment based on reinforcement learning. Background Art

[0002] In today's society, spa equipment is widely used in medical rehabilitation, sports training, leisure and entertainment and other fields. During its operation, it consumes a lot of energy to maintain multiple functions such as heating, circulation, and massage. For example, a large spa bathtub may need to continuously heat and circulate and filter a large amount of water, which makes energy consumption an important part of the operating cost of spa equipment. The energy consumption of spa equipment is closely related to multiple factors such as the operating status of the equipment, water temperature, water flow rate, and user usage habits. These factors interact with each other, making the energy consumption of spa equipment present complex dynamic characteristics. Traditional spa equipment energy consumption control methods mainly rely on fixed programs and preset parameters. For example, the equipment may turn on or off the heating system according to a preset schedule without considering actual water temperature changes and usage needs. This fixed mode control method cannot be dynamically adjusted according to the real-time operating status of the equipment, resulting in energy consumption. The phenomenon of energy waste is quite serious. In addition, traditional energy monitoring methods can only obtain relatively rough energy consumption data, and it is difficult to capture subtle changes and transient characteristics of energy consumption. This makes the formulation of optimization strategies lack accurate data support. With the continuous advancement of science and technology, energy consumption optimization technology has gradually been paid attention to in all walks of life. In the field of spa equipment, the development of methods that can monitor equipment status in real time, accurately analyze energy consumption characteristics, and dynamically adjust energy consumption optimization strategies based on reinforcement learning algorithms has become an inevitable demand for the development of the industry. Reinforcement learning, as an advanced machine learning technology, can continuously optimize strategies to achieve goal optimization through interactive learning between intelligent agents and the environment. It has significant advantages in solving optimization problems of complex systems. Applying reinforcement learning to the energy consumption optimization of spa equipment can not only improve energy utilization efficiency and reduce operating costs, but also reduce the impact on the environment and achieve sustainable development.

[0003] Existing technologies are unable to dynamically adjust energy consumption based on real-time status. Energy consumption monitoring is rough and the optimization strategy is single. The impact of corrosion of equipment components on energy consumption is ignored, resulting in inaccurate and inefficient energy consumption optimization and serious energy waste. Summary of the Invention

[0004] In order to overcome the problems that the existing technology cannot dynamically adjust energy consumption according to real-time status, energy consumption monitoring is rough and the optimization strategy is single, and the impact of corrosion of equipment components on energy consumption is ignored, resulting in inaccurate and inefficient energy consumption optimization and serious energy waste, the present invention discloses a spa equipment energy consumption optimization method based on reinforcement learning that can effectively solve the above technical problems.

[0005] In order to solve the above technical problems, the technical solutions of the present invention are as follows:

[0006] A method for optimizing energy consumption of hydrotherapy equipment based on reinforcement learning, characterized in that the method comprises:

[0007] Acquire real-time energy consumption data and device status data of the hydrotherapy equipment during operation, process the real-time energy consumption data based on a wavelet decomposition algorithm, extract transient energy consumption characteristics, and calculate the current harmonic distortion rate of the equipment as an indicator of the equipment operation status;

[0008] Determining an initial energy consumption optimization strategy based on the transient energy consumption characteristics, the current harmonic distortion rate, and a preset device operation mode library, wherein the initial energy consumption optimization strategy includes multiple candidate energy consumption optimization strategies;

[0009] Obtain the strategy parameters corresponding to each candidate energy consumption optimization strategy and the energy consumption benefit estimation model built based on reinforcement learning;

[0010] Inputting the policy parameters corresponding to each candidate energy consumption optimization strategy into the energy consumption benefit estimation model for benefit estimation processing to obtain estimated energy consumption benefit data for each candidate energy consumption optimization strategy;

[0011] Obtaining macro-micro motion layered data and corrosion data of the spa equipment components, performing decision analysis based on the estimated energy consumption benefit data, macro-micro motion layered data, and corrosion data of each candidate energy consumption optimization strategy, and selecting a target energy consumption optimization strategy from the multiple candidate energy consumption optimization strategies;

[0012] Performing resource allocation on the target energy consumption optimization strategy to obtain a resource allocation result;

[0013] When the resource configuration result meets the preset energy consumption optimization condition, the hydrotherapy equipment is controlled according to the target energy consumption optimization strategy.

[0014] Preferably, determining the initial energy consumption optimization strategy includes:

[0015] Obtaining a device operation mode database, wherein the device operation mode database includes a preset transient energy consumption characteristic range, a current harmonic distortion rate threshold, and a corresponding energy consumption optimization strategy;

[0016] Based on the transient energy consumption characteristics and current harmonic distortion rate, the equipment operation mode database is queried to obtain a preset energy consumption optimization strategy that matches the current equipment operation status, and the matching preset energy consumption optimization strategy is determined to be the initial energy consumption optimization strategy; the equipment operation mode database is constructed based on historical equipment operation status data and corresponding effective energy consumption optimization strategies.

[0017] Preferably, the method for constructing the equipment operation mode database includes:

[0018] Acquire a sample set of equipment operating status information, wherein the sample set includes transient energy consumption characteristic samples and current harmonic distortion rate samples under different working conditions;

[0019] Setting at least one corresponding energy consumption optimization strategy based on each device operation status information sample;

[0020] The device operation mode database is constructed based on each device operation status information sample and the energy consumption optimization strategy corresponding to each sample.

[0021] Preferably, the decision analysis based on the estimated energy consumption benefit data, macro-micro action hierarchical data and corrosion degree data of each candidate energy consumption optimization strategy to select a target energy consumption optimization strategy from the multiple candidate energy consumption optimization strategies includes:

[0022] Performing benefit analysis on the estimated energy consumption benefit data of each candidate energy consumption optimization strategy to obtain the estimated energy consumption benefit of each candidate energy consumption optimization strategy;

[0023] In the case where the estimated energy consumption benefit is greater than the preset energy consumption benefit threshold, determining the candidate energy consumption optimization strategy corresponding to the estimated energy consumption benefit greater than the preset energy consumption benefit threshold as the intermediate energy consumption optimization strategy;

[0024] Performing feature evaluation on the macro-micro motion hierarchical data and combining it with corrosion data of equipment components to obtain equipment operation feature data;

[0025] A matching analysis is performed on the intermediate energy consumption optimization strategy and the equipment operation characteristic data, and based on a probabilistic action shielding rule driven by the corrosion degree, the intermediate energy consumption optimization strategy whose matching analysis result meets the preset matching conditions is determined as the target energy consumption optimization strategy.

[0026] Preferably, obtaining the equipment operation characteristic data includes:

[0027] Calculate the energy consumption contribution of equipment motion based on macro-micro motion hierarchical data, and calculate the corrosion influence coefficient based on the corrosion degree of equipment components;

[0028] Calculate the comprehensive evaluation value based on the equipment's energy consumption contribution value and corrosion influence coefficient;

[0029] The comprehensive evaluation value is compared with a preset device operation characteristic threshold to obtain a comparison result, and the device operation characteristic data is determined based on the comparison result.

[0030] Preferably, the preset device operation characteristic threshold includes a first threshold and a second threshold, the first threshold is greater than the second threshold, the device operation characteristic data includes first operation characteristic data, second operation characteristic data, and third operation characteristic data, and comparing the comprehensive evaluation value with the preset device operation characteristic threshold to obtain a comparison result, and determining the device operation characteristic data based on the comparison result includes:

[0031] When the comprehensive evaluation value is greater than or equal to the first threshold, determining that the equipment operation characteristic data is the first operation characteristic data;

[0032] When the comprehensive evaluation value is less than the first threshold and greater than or equal to the second threshold, determining that the device operation characteristic data is the second operation characteristic data;

[0033] When the comprehensive evaluation value is less than the second threshold, the device operation characteristic data is determined to be the third operation characteristic data.

[0034] Preferably, the intermediate energy consumption optimization strategy includes a first strategy, a second strategy, and a third strategy, and the equipment operation characteristic data includes first operation characteristic data, second operation characteristic data, and third operation characteristic data. The matching analysis of the intermediate energy consumption optimization strategy and the equipment operation characteristic data is performed, and based on a probabilistic action shielding rule driven by the corrosion degree, the intermediate energy consumption optimization strategy whose matching analysis result satisfies a preset matching condition is determined as the target energy consumption optimization strategy, including:

[0035] When the device operation characteristic data is the first operation characteristic data, determining the first strategy as the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree;

[0036] When the device operation characteristic data is the second operation characteristic data, determining the second strategy as the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree;

[0037] When the equipment operation characteristic data is the third operation characteristic data, the third strategy is determined to be the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree.

[0038] Preferably, the training method of the energy consumption benefit estimation model includes:

[0039] Obtain an initial energy consumption benefit estimation model based on an Actor network constrained by thermodynamic equations and a lightweight spiking neural network;

[0040] Inputting the strategy parameters of the sample energy consumption optimization strategy into the initial energy consumption benefit estimation model to perform benefit estimation processing to obtain estimated energy consumption benefit data corresponding to the sample energy consumption optimization strategy;

[0041] Matching the energy consumption benefit label data of the sample energy consumption optimization strategy with the estimated energy consumption benefit data of the corresponding sample energy consumption optimization strategy to obtain a matching result;

[0042] If the matching result does not meet the preset condition, determining the difference information between the energy consumption benefit label data of the sample energy consumption optimization strategy and the estimated energy consumption benefit data of the corresponding sample energy consumption optimization strategy;

[0043] Based on the difference information, the network parameters of the initial energy consumption benefit estimation model are updated, and the network parameters are constrained and adjusted in combination with domain knowledge until the matching result meets the preset conditions, and the initial energy consumption benefit estimation model corresponding to the matching result that meets the preset conditions is determined to be the energy consumption benefit estimation model.

[0044] An electronic device includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or at least one program is loaded and executed by the processor to implement the energy consumption optimization method of spa equipment based on reinforcement learning as described above.

[0045] A computer-readable storage medium stores computer instructions. When the computer instructions are executed, the above-mentioned method for optimizing energy consumption of spa equipment based on reinforcement learning is executed.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention processes real-time energy consumption data through a wavelet decomposition algorithm, accurately extracts transient energy consumption features, and comprehensively evaluates the equipment operating status in combination with the current harmonic distortion rate. Traditional methods can only perform rough energy consumption monitoring and cannot capture these key features, resulting in inaccurate and incomplete energy consumption data. Based on these detailed data, the present invention determines the initial energy consumption optimization strategy based on a preset equipment operation mode library, covering multiple candidate strategies, breaking through the limitations of traditional single optimization strategies; further, the present invention uses an energy consumption benefit prediction model constructed by reinforcement learning to estimate the benefits of each candidate energy consumption optimization strategy. The reinforcement learning algorithm continuously adjusts and optimizes the strategy through interactive learning between the intelligent agent and the environment, and can explore better energy consumption control solutions. This is in contrast to the traditional fixed-mode control method. The traditional method cannot dynamically adjust the strategy according to the real-time status, resulting in poor energy consumption optimization effect; in the decision analysis stage, the macro-micro action hierarchical data and component corrosion data of the equipment are taken into consideration, and the equipment action is calculated by The energy consumption contribution value and corrosion influence coefficient are combined with the preset threshold to determine the equipment operation characteristic data, so that the decision-making process fully considers the actual operation characteristics of the equipment and the wear of components. The traditional method ignores the impact of the corrosion degree of equipment components on energy consumption. The present invention performs matching analysis based on the probabilistic action shielding rule driven by corrosion degree, thereby screening out the optimal target energy consumption optimization strategy, and realizing more accurate and efficient energy consumption optimization; the present invention allocates resources for the target energy consumption optimization strategy, and controls the spa equipment when the preset conditions are met. This series of operations forms a closed-loop optimization control system, which can dynamically adjust the energy consumption optimization strategy in real time to ensure that the equipment can achieve energy consumption optimization under different operating conditions. Compared with the existing technology, the present invention not only improves the accuracy and comprehensiveness of energy consumption monitoring, but also realizes dynamic adjustment and precise matching of energy consumption optimization strategies through reinforcement learning and multi-dimensional data analysis, effectively solving the shortcomings of the existing technology in energy consumption optimization, reducing the energy consumption of spa equipment, improving energy utilization efficiency, and reducing energy waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the implementation methods or the description of the prior art. Obviously, the drawings described below are merely exemplary. For ordinary technicians in this field, other implementation drawings can be derived based on the provided drawings without any creative work.

[0048] Figure 1 It is a step diagram of the method of the present invention. DETAILED DESCRIPTION

[0049] The accompanying drawings are for illustrative purposes only and are not to be construed as limiting this patent;

[0050] In order to better illustrate this embodiment, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product size;

[0051] It is understandable to those skilled in the art that some well-known structures and descriptions thereof may be omitted in the drawings.

[0052] The technical solution of the present invention is further described below with reference to the accompanying drawings and embodiments.

[0053] Example

[0054] A method for optimizing energy consumption of spa equipment based on reinforcement learning, the method comprising:

[0055] Acquire real-time energy consumption data and device status data of the hydrotherapy equipment during operation, process the real-time energy consumption data based on a wavelet decomposition algorithm, extract transient energy consumption characteristics, and calculate the current harmonic distortion rate of the equipment as an indicator of the equipment operation status;

[0056] Determining an initial energy consumption optimization strategy based on the transient energy consumption characteristics, the current harmonic distortion rate, and a preset device operation mode library, wherein the initial energy consumption optimization strategy includes multiple candidate energy consumption optimization strategies;

[0057] Obtain the strategy parameters corresponding to each candidate energy consumption optimization strategy and the energy consumption benefit estimation model built based on reinforcement learning;

[0058] Inputting the policy parameters corresponding to each candidate energy consumption optimization strategy into the energy consumption benefit estimation model for benefit estimation processing to obtain estimated energy consumption benefit data for each candidate energy consumption optimization strategy;

[0059] Obtaining macro-micro motion layered data and corrosion data of the spa equipment components, performing decision analysis based on the estimated energy consumption benefit data, macro-micro motion layered data, and corrosion data of each candidate energy consumption optimization strategy, and selecting a target energy consumption optimization strategy from the multiple candidate energy consumption optimization strategies;

[0060] Performing resource allocation on the target energy consumption optimization strategy to obtain a resource allocation result;

[0061] When the resource configuration result meets the preset energy consumption optimization condition, the hydrotherapy equipment is controlled according to the target energy consumption optimization strategy.

[0062] Determining the initial energy consumption optimization strategy includes:

[0063] Obtaining a device operation mode database, wherein the device operation mode database includes a preset transient energy consumption characteristic range, a current harmonic distortion rate threshold, and a corresponding energy consumption optimization strategy;

[0064] Based on the transient energy consumption characteristics and current harmonic distortion rate, the equipment operation mode database is queried to obtain a preset energy consumption optimization strategy that matches the current equipment operation status, and the matching preset energy consumption optimization strategy is determined to be the initial energy consumption optimization strategy; the equipment operation mode database is constructed based on historical equipment operation status data and corresponding effective energy consumption optimization strategies.

[0065] The method for constructing the equipment operation mode database includes:

[0066] Acquire a sample set of equipment operating status information, wherein the sample set includes transient energy consumption characteristic samples and current harmonic distortion rate samples under different working conditions;

[0067] Setting at least one corresponding energy consumption optimization strategy based on each device operation status information sample;

[0068] The device operation mode database is constructed based on each device operation status information sample and the energy consumption optimization strategy corresponding to each sample.

[0069] The decision analysis based on the estimated energy consumption benefit data, macro-micro action hierarchical data and corrosion degree data of each candidate energy consumption optimization strategy, and screening out a target energy consumption optimization strategy from the multiple candidate energy consumption optimization strategies includes:

[0070] Performing benefit analysis on the estimated energy consumption benefit data of each candidate energy consumption optimization strategy to obtain the estimated energy consumption benefit of each candidate energy consumption optimization strategy;

[0071] In the case where the estimated energy consumption benefit is greater than the preset energy consumption benefit threshold, determining the candidate energy consumption optimization strategy corresponding to the estimated energy consumption benefit greater than the preset energy consumption benefit threshold as the intermediate energy consumption optimization strategy;

[0072] Performing feature evaluation on the macro-micro motion hierarchical data and combining it with corrosion data of equipment components to obtain equipment operation feature data;

[0073] A matching analysis is performed on the intermediate energy consumption optimization strategy and the equipment operation characteristic data, and based on a probabilistic action shielding rule driven by the corrosion degree, the intermediate energy consumption optimization strategy whose matching analysis result meets the preset matching conditions is determined as the target energy consumption optimization strategy.

[0074] Obtaining the equipment operation characteristic data includes:

[0075] Calculate the energy consumption contribution of equipment motion based on macro-micro motion hierarchical data, and calculate the corrosion influence coefficient based on the corrosion degree of equipment components;

[0076] Calculate the comprehensive evaluation value based on the equipment's energy consumption contribution value and corrosion influence coefficient;

[0077] The comprehensive evaluation value is compared with a preset device operation characteristic threshold to obtain a comparison result, and the device operation characteristic data is determined based on the comparison result.

[0078] The preset device operation characteristic threshold includes a first threshold and a second threshold, the first threshold is greater than the second threshold, the device operation characteristic data includes first operation characteristic data, second operation characteristic data, and third operation characteristic data, and comparing the comprehensive evaluation value with the preset device operation characteristic threshold to obtain a comparison result, and determining the device operation characteristic data based on the comparison result includes:

[0079] When the comprehensive evaluation value is greater than or equal to the first threshold, determining that the equipment operation characteristic data is the first operation characteristic data;

[0080] When the comprehensive evaluation value is less than the first threshold and greater than or equal to the second threshold, determining that the device operation characteristic data is the second operation characteristic data;

[0081] When the comprehensive evaluation value is less than the second threshold, the device operation characteristic data is determined to be the third operation characteristic data.

[0082] The intermediate energy consumption optimization strategy includes a first strategy, a second strategy, and a third strategy; the equipment operation characteristic data includes first operation characteristic data, second operation characteristic data, and third operation characteristic data; performing matching analysis on the intermediate energy consumption optimization strategy and the equipment operation characteristic data, and determining, based on a probabilistic action shielding rule driven by the corrosion degree, the intermediate energy consumption optimization strategy whose matching analysis result satisfies a preset matching condition as the target energy consumption optimization strategy, includes:

[0083] When the device operation characteristic data is the first operation characteristic data, determining the first strategy as the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree;

[0084] When the device operation characteristic data is the second operation characteristic data, determining the second strategy as the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree;

[0085] When the equipment operation characteristic data is the third operation characteristic data, the third strategy is determined to be the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree.

[0086] The training method of the energy consumption benefit estimation model includes:

[0087] Obtain an initial energy consumption benefit estimation model based on an Actor network constrained by thermodynamic equations and a lightweight spiking neural network;

[0088] Inputting the strategy parameters of the sample energy consumption optimization strategy into the initial energy consumption benefit estimation model to perform benefit estimation processing to obtain estimated energy consumption benefit data corresponding to the sample energy consumption optimization strategy;

[0089] Matching the energy consumption benefit label data of the sample energy consumption optimization strategy with the estimated energy consumption benefit data of the corresponding sample energy consumption optimization strategy to obtain a matching result;

[0090] If the matching result does not meet the preset condition, determining the difference information between the energy consumption benefit label data of the sample energy consumption optimization strategy and the estimated energy consumption benefit data of the corresponding sample energy consumption optimization strategy;

[0091] Based on the difference information, the network parameters of the initial energy consumption benefit estimation model are updated, and the network parameters are constrained and adjusted in combination with domain knowledge until the matching result meets the preset conditions, and the initial energy consumption benefit estimation model corresponding to the matching result that meets the preset conditions is determined to be the energy consumption benefit estimation model.

[0092] An electronic device includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or at least one program is loaded and executed by the processor to implement the energy consumption optimization method of spa equipment based on reinforcement learning as described above.

[0093] A computer-readable storage medium stores computer instructions. When the computer instructions are executed, the above-mentioned method for optimizing energy consumption of spa equipment based on reinforcement learning is executed.

[0094] See also Figure 1 , install a variety of smart sensors on the key components and systems of spa equipment to ensure that the real-time energy consumption data and equipment status data of the equipment during operation can be fully and accurately obtained, including:

[0095] Power sensors and current sensors are installed on the water pump, motor and other components of the water massage system to collect the power consumption and current waveform data of the water pump and motor in real time.

[0096] Temperature sensors and power sensors are installed near the heater, cooler, and water temperature sensor of the temperature control system to obtain water temperature changes and energy consumption data of the heater and cooler.

[0097] Flow sensors and pressure sensors are installed at the air pump, solenoid valve and other parts of the bubble generation system to monitor the gas flow and pressure changes during the bubble generation process, and at the same time collect the energy consumption data of the air pump.

[0098] Light intensity sensors and power sensors are installed near each lighting module of the lighting atmosphere system to detect changes in light brightness and energy consumption of the lighting module.

[0099] Install audio signal sensors and power sensors near the power amplifier and speakers of the audio playback system to collect signal strength and energy consumption data during audio playback.

[0100] The above sensors collect data in real time at a frequency of 10 times per second and transmit the data to the central processing unit via communication methods such as RS485 bus, Wi-Fi or Bluetooth.

[0101] After receiving the real-time energy consumption data, the central processing unit calls the wavelet decomposition algorithm to preprocess the data, selects the Daubechies wavelet as the wavelet basis function, and sets the decomposition scale to 3. By performing wavelet decomposition on the real-time energy consumption data, the data is decomposed into approximate components and detail components in different frequency bands. The approximate components reflect the overall trend of energy consumption, while the detail components reflect the transient fluctuation characteristics of energy consumption.

[0102] The specific steps for extracting transient energy consumption features are as follows:

[0103] Calculate the energy value of each detail component, that is, calculate the square sum of the data sequence of the component.

[0104] According to the energy proportion of each detail component, the main transient energy consumption characteristic components are determined.

[0105] Calculate the statistical characteristic values ​​such as mean, variance, peak value of the main transient energy consumption characteristic components to form the transient energy consumption characteristic vector.

[0106] At the same time, the Fourier transform is used to analyze the equipment current waveform data and calculate the current harmonic distortion rate. According to the formula:

[0107]

[0108] Among them, I1 is the effective value of the fundamental current, I i is the effective value of the i-th harmonic current, n is the highest harmonic order, and the calculated current harmonic distortion rate is used as an important indicator of the equipment operating status in the energy consumption optimization strategy determination process.

[0109] The equipment operation mode database is constructed based on the historical operation data of the spa equipment. It stores the equipment operation status information under different working conditions and the corresponding effective energy consumption optimization strategies. The specific construction steps are as follows:

[0110] The operating data of the equipment in different time periods, different user usage patterns, and different environmental conditions are collected to form a sample set of equipment operating status information. Each sample contains a transient energy consumption feature sample, such as the mean, variance, peak value and other characteristic values ​​obtained by wavelet decomposition, a current harmonic distortion rate sample, and the corresponding operating parameters of each component of the equipment, such as water flow rate, water temperature, bubble generation frequency, light brightness, audio playback volume, etc.

[0111] For each sample, engineers and energy-saving experts jointly set at least one corresponding energy consumption optimization strategy based on the actual operating performance and energy consumption of the equipment at that time. For example, when the equipment is running at high load, an energy consumption optimization strategy is set to reduce the power of the water massage system and reduce the frequency of bubble generation; when the equipment is running at low load, an energy consumption optimization strategy is set to appropriately lower the water temperature set value and dim the light brightness, etc.

[0112] Each sample and its corresponding energy consumption optimization strategy are stored in the database to form an equipment operation mode database. The database adopts a relational database structure, uses the transient energy consumption characteristic range, current harmonic distortion rate threshold, etc. as query keywords, and establishes an index relationship with the corresponding energy consumption optimization strategy for fast query.

[0113] During actual operation, when it is necessary to determine the initial energy consumption optimization strategy, the currently acquired transient energy consumption characteristics and current harmonic distortion rate are input as query conditions into the equipment operation mode database query module. The query module searches the database for a preset energy consumption optimization strategy that matches the current characteristics and indicators based on a preset matching algorithm.

[0114] The specific implementation steps of the matching algorithm are as follows:

[0115] Calculate the Euclidean distance between the current transient energy consumption feature vector and the transient energy consumption feature vector of each sample in the database.

[0116] The samples whose Euclidean distance is less than a set threshold, such as 0.5, are regarded as samples matching the current transient energy consumption feature.

[0117] At the same time, it is determined whether the current harmonic distortion rate of the current falls within the current harmonic distortion rate threshold range of the corresponding sample in the database. The current harmonic distortion rate threshold range is determined according to the statistical distribution of the harmonic distortion rate during normal operation of the equipment. It is generally set to [0%, 15%] as the normal range. If it exceeds 15%, it is considered an abnormal operating state.

[0118] The energy consumption optimization strategy corresponding to the sample that satisfies both the transient energy consumption feature matching and the current harmonic distortion rate threshold range is taken out as the initial energy consumption optimization strategy. If there are multiple matching samples, the energy consumption optimization strategies corresponding to these samples are aggregated to form an initial energy consumption optimization strategy set containing multiple candidate energy consumption optimization strategies.

[0119] The energy consumption benefit estimation model is built based on reinforcement learning. It uses a structure that combines an actor network based on thermodynamic equation constraints and a lightweight spiking neural network. The specific steps of model training are as follows:

[0120] A large amount of sample data is collected, including energy consumption optimization strategy parameters under different equipment operating conditions and the corresponding energy consumption benefit label data. The energy consumption benefit label data is obtained by calculating the energy consumption difference of the actual measurement equipment before and after executing the corresponding energy consumption optimization strategy. At the same time, changes in equipment operating performance indicators such as water flow massage effect, water temperature stability, bubble richness, lighting brightness uniformity, audio playback quality, etc. are taken into consideration to ensure that the energy consumption optimization strategy will not have a significant negative impact on equipment operating performance while reducing energy consumption.

[0121] Initialize the parameters of the Actor network and lightweight pulse neural network based on thermodynamic equation constraints. The Actor network is responsible for generating energy consumption optimization strategy actions based on the equipment operating status. Its network structure includes an input layer, a hidden layer, and an output layer. The number of nodes in the input layer corresponds to the characteristic dimensions of the equipment operating status, such as the transient energy consumption characteristic dimension, the current harmonic distortion rate dimension, and the operating parameter dimensions of each equipment component. The hidden layer adopts a multi-layer neuron structure, and the number of nodes in the output layer corresponds to the energy consumption optimization strategy parameter dimension. The lightweight pulse neural network is used to simulate the pulse emission characteristics of biological neurons and model the dynamic change process of equipment energy consumption. Its network structure includes an input pulse encoding layer, a hidden pulse layer, and an output pulse decoding layer.

[0122] The strategy parameters in the sample data are input into the initial energy consumption benefit estimation model. The model performs forward propagation calculation based on the current network parameters to obtain the estimated energy consumption benefit data of the corresponding sample energy consumption optimization strategy.

[0123] Calculate the error between the estimated energy consumption benefit data and the actual energy consumption benefit label data, and use the mean square error (MSE) as the loss function, that is:

[0124]

[0125] Where N is the number of samples, y i is the actual energy consumption benefit label data of the i-th sample, is the estimated energy consumption benefit data of the i-th sample.

[0126] Based on the loss function value, the backpropagation algorithm is used to update the network parameters of the model. During the update process, the parameters of the Actor network are constrained and adjusted in combination with the constraints of the thermodynamic equation to ensure that the generated energy consumption optimization strategy conforms to the physical operation laws of the equipment. At the same time, the parameters of the lightweight pulse neural network are optimized so that it can more accurately simulate the dynamic changes in equipment energy consumption. The above steps are repeated until the loss function value of the model converges to the set threshold range, for example, less than 0.01. At this time, the model is a trained energy consumption benefit estimation model.

[0127] After determining the initial energy consumption optimization strategy, the strategy parameters corresponding to each candidate energy consumption optimization strategy are obtained and input into the trained energy consumption benefit estimation model. The model performs benefit estimation processing on each candidate energy consumption optimization strategy based on the internal neural network structure and learned parameters, and outputs the estimated energy consumption benefit data of each candidate energy consumption optimization strategy. The estimated energy consumption benefit data is expressed in percentage form, reflecting the proportion of equipment energy consumption that is expected to be reduced by implementing the strategy.

[0128] The macro-micro motion layered data of the spa equipment reflects the different levels of motion behavior and energy consumption contribution of the equipment during operation. The specific acquisition method is as follows:

[0129] For the macro-actions of the water massage system, the water flow velocity and flow data during the water circulation process are collected through the water flow velocity sensor and the water flow flow sensor to calculate the overall energy consumption contribution of the water circulation system. For example, the higher the water flow velocity and the greater the flow, the higher the energy consumption of the water circulation system.

[0130] For the micro-movements of the water massage system, the vibration sensor and pressure sensor of the water pump are used to collect data on tiny vibrations and pressure fluctuations inside the water pump, and analyze the tiny energy consumption changes during the operation of the water pump. For example, tiny vibrations inside the water pump may cause additional energy loss.

[0131] For the macro actions of the temperature control system, the power change data of the heater and cooler are collected through the power sensors. Combined with the temperature change data of the water temperature sensor, the overall energy consumption contribution of the temperature control system in the process of regulating the water temperature is calculated. For example, when the water temperature needs to be increased or decreased quickly, the high-power operation of the heater or cooler will result in greater energy consumption.

[0132] For micro-movements of the temperature control system, the surface temperature sensors and internal structure monitoring sensors of the heater and cooler are used to collect data on their surface temperature distribution and internal structure changes, and analyze the tiny heat losses of the heater and cooler during operation. For example, the uneven distribution of the heater surface temperature may lead to increased heat loss.

[0133] For the macro action of the bubble generation system, the gas flow and pressure data are collected through the flow sensor and pressure sensor of the air pump to calculate the overall energy consumption contribution of the bubble generation system in the bubble generation process. For example, the higher the bubble generation frequency and the larger the bubble volume, the higher the energy consumption of the air pump.

[0134] For the micro-movements of the bubble generation system, the bubble size and distribution data are collected through bubble size sensors and bubble distribution sensors to analyze the impact of the formation and collapse of tiny bubbles on energy consumption during the bubble generation process. For example, the formation of tiny bubbles may require higher energy input.

[0135] For the macro actions of the lighting atmosphere system, the light intensity sensor and power sensor of the lighting module collect light brightness change and power consumption data, and calculate the overall energy consumption contribution of the lighting atmosphere system in different lighting modes. For example, high-brightness lighting mode consumes more energy than low-brightness lighting mode.

[0136] For micro-movements in the lighting atmosphere system, the color temperature sensor and spectrum analysis sensor of the lighting module collect light color temperature and spectral distribution data, and analyze the slight energy consumption differences of the light during different color temperature and spectrum changes. For example, lights with certain specific spectra may require more energy consumption to produce.

[0137] For the macro actions of the audio playback system, the power sensor and audio signal strength sensor of the audio playback device are used to collect power changes and audio signal strength data during the audio playback process, and calculate the overall energy consumption contribution of the audio playback system under different volume levels and audio content. For example, high-volume playback consumes more energy than low-volume playback.

[0138] For micro-movements in the audio playback system, audio distortion and frequency response data are collected through the audio playback device's audio distortion sensor and audio frequency response sensor to analyze the impact of tiny changes in audio quality on energy consumption during audio playback. For example, high-fidelity audio playback requires more energy consumption to maintain sound quality.

[0139] The corrosion data of equipment components are measured and calculated regularly using ultrasonic thickness gauges, electrochemical corrosion monitors and other testing equipment. The specific methods are as follows:

[0140] For components of the water massage system, such as the water pump impeller and the inner wall of the pipe, an ultrasonic thickness gauge is used to measure the thickness change of the components. By comparing the thickness values ​​of the components at different time points, the average corrosion rate of the components is calculated. For example, after a water pump impeller has been running for a period of time, its thickness may decrease by 0.1mm. The annual average corrosion rate is calculated based on the operating time and thickness change.

[0141] For components of the temperature control system, such as heater tube bundles and refrigerator heat exchange tubes, an electrochemical corrosion monitor is used to measure the electrochemical corrosion current and potential on the surface of the components. The corrosion rate of the components is calculated based on the empirical relationship between electrochemical corrosion parameters and corrosion rate. For example, an increase in the electrochemical corrosion current of the heater tube bundle may lead to an accelerated corrosion rate.

[0142] For components of the bubble generation system, such as the air pump cylinder and solenoid valve core, the corrosion morphology and degree of the component surface are observed through disassembly inspection and surface analysis methods. Combined with the equipment operation time and corrosive environment factors, the corrosion degree of the components is estimated. For example, slight corrosion spots appear on the inner wall of the air pump cylinder. The corrosion degree is estimated to be 5% based on the spot area and depth.

[0143] For components of the lighting atmosphere system, such as the lighting module housing and reflector, spectral analysis and surface energy spectrum analysis methods are used to detect the composition and distribution of corrosion products on the component surface. Based on the content and distribution of corrosion products, the corrosion degree of the components is evaluated. For example, a high content of corrosion products on the surface of the lighting module housing indicates a high corrosion degree, which can reach 10%.

[0144] For components of the audio playback system, such as speaker diaphragms and amplifier circuit boards, visual inspection and functional testing methods are used to check whether the components have performance degradation caused by corrosion. For example, slight oxidation of the speaker diaphragm may affect the audio playback quality. The estimated corrosion level is 3%.

[0145] The specific steps for analyzing the estimated energy consumption benefit data of each candidate energy consumption optimization strategy are as follows:

[0146] The estimated energy consumption benefit data is combined with the current actual energy consumption level of the device to calculate the absolute value of the energy consumption that can be saved after executing the strategy. For example, if the current energy consumption of the device is 10kW and the estimated energy consumption benefit of candidate strategy A is 10%, the estimated energy consumption saving is 1kW.

[0147] An economic analysis is conducted based on the absolute value of the expected energy savings and the operating costs of the equipment (including electricity costs, equipment maintenance fees, etc.) to calculate the expected economic benefits of implementing the strategy. For example, if 1kW of energy is saved, based on the local electricity price of 0.8 yuan / kWh, and the equipment runs for 8 hours a day, the daily electricity savings can be 8×1×0.8=6.4 yuan.

[0148] The preset energy consumption benefit threshold is set to 5%, and the candidate energy consumption optimization strategies with a value greater than the threshold are determined as intermediate energy consumption optimization strategies. For example, among the three candidate strategies, the estimated energy consumption benefit of strategy A is 10%, strategy B is 6%, and strategy C is 4%, then strategies A and B are determined as intermediate energy consumption optimization strategies.

[0149] The specific method for evaluating the characteristics of macro-micro motion layered data and combining it with the corrosion data of equipment components to obtain the equipment operation characteristic data is as follows:

[0150] The energy consumption contribution value of the equipment motion is calculated based on the macro-micro motion hierarchical data. For each system, such as the water massage system, temperature control system, etc., the energy consumption contribution ratio of its macro motion and micro motion is calculated respectively. For example, the energy consumption contribution ratio of the macro motion of the water massage system is 60%, and the energy consumption contribution ratio of the micro motion is 40%; the energy consumption contribution ratio of the macro motion of the temperature control system is 70%, and the energy consumption contribution ratio of the micro motion is 30%, etc. Then, according to the energy consumption contribution ratio of each system, the energy consumption contribution value of the motion of the entire equipment is calculated. For example, the overall energy consumption contribution value of the water massage system is 30%, the temperature control system is 25%, the bubble generation system is 20%, the lighting atmosphere system is 15%, and the audio playback system is 10%.

[0151] The corrosion influence coefficient is calculated based on the corrosion degree of the equipment components. For each system component, the corrosion influence coefficient is calculated based on its corrosion degree and importance in the system. For example, if the corrosion degree of the water massage system components is 10% and its importance coefficient is 0.8, then the corrosion influence coefficient is 0.1×0.8=0.08; if the corrosion degree of the temperature control system components is 8% and its importance coefficient is 0.9, then the corrosion influence coefficient is 0.08×0.9=0.072, and so on. The corrosion influence coefficients of each system are then weighted and summed to obtain the corrosion influence coefficient of the entire equipment. If the weight of the water massage system is 0.3, the weight of the temperature control system is 0.25, the weight of the bubble generation system is 0.2, the weight of the lighting atmosphere system is 0.15, and the weight of the audio playback system is 0.1, then the overall corrosion influence coefficient of the equipment is 0.3×0.08+0.25×0.072+0.2×0.06+0.15×0.05+0.1×0.03=0.0648.

[0152] The comprehensive evaluation value is calculated based on the equipment's energy consumption contribution value and the corrosion influence coefficient. The calculation formula for the comprehensive evaluation value is:

[0153] Comprehensive evaluation value = equipment operation energy consumption contribution value × corrosion influence coefficient

[0154] The contribution value of equipment operation energy consumption is 0.3, and the corrosion influence coefficient is 0.0648, so the comprehensive evaluation value is 0.3×0.0648=0.01944.

[0155] Set the first threshold value of the preset device operation characteristic threshold value to 0.02 and the second threshold value to 0.01. According to the comparison result of the comprehensive evaluation value and the preset threshold value, determine the device operation characteristic data:

[0156] If the comprehensive evaluation value is greater than or equal to the first threshold value 0.02, the equipment operation characteristic data is determined to be the first operation characteristic data, which indicates that the energy consumption contribution of the equipment is high and the corrosion situation is relatively serious. A more conservative energy consumption optimization strategy needs to be adopted to give priority to ensuring the reliable operation of the equipment and avoid accelerated damage to equipment components due to excessive optimization.

[0157] If the comprehensive evaluation value is less than the first threshold value 0.02 and greater than or equal to the second threshold value 0.01, the equipment operation characteristic data is determined to be the second operation characteristic data, which means that the energy consumption contribution and corrosion condition of the equipment are at a medium level, and energy consumption optimization can be appropriately performed under the premise of ensuring the normal operation of the equipment.

[0158] If the comprehensive evaluation value is less than the second threshold value 0.01, the equipment operation characteristic data is determined to be the third operation characteristic data, which means that the energy consumption contribution of the equipment is low and the corrosion situation is light. A more radical energy consumption optimization strategy can be adopted to achieve greater energy saving effects.

[0159] The intermediate energy consumption optimization strategies include the first strategy, the second strategy, and the third strategy, which correspond to different energy consumption optimization degrees and equipment operation adjustment modes, respectively. The equipment operation characteristic data include the first operation characteristic data, the second operation characteristic data, and the third operation characteristic data, which correspond to different equipment operation status assessment results, respectively. Based on the probabilistic action shielding rule driven by the corrosion degree, the specific method for determining the intermediate energy consumption optimization strategy whose matching analysis results meet the preset matching conditions as the target energy consumption optimization strategy is as follows:

[0160] When the equipment operation characteristic data is the first operation characteristic data, due to the high energy consumption contribution of the equipment and the serious corrosion situation, a more conservative energy consumption optimization strategy needs to be adopted. At this time, according to the probabilistic action shielding rule driven by the corrosion degree, those actions that may cause the corrosion of equipment components to be aggravated are shielded with a higher probability (for example, 80%), such as reducing the power of the water massage system, reducing the frequency of bubble generation, etc. Among the remaining executable actions, the first strategy is determined as the target energy consumption optimization strategy that meets the preset matching conditions. The first strategy moderately reduces energy consumption while ensuring the reliable operation of the equipment. For example, it reduces the power of the water massage system by 10%, and optimizes the operating parameters of the temperature control system to reduce energy consumption by 5% while meeting the water temperature requirements.

[0161] When the equipment operation characteristic data is the second operation characteristic data, the energy consumption contribution and corrosion condition of the equipment are at a medium level. At this time, according to the probabilistic action shielding rule driven by the corrosion degree, some actions that may cause aggravated corrosion of equipment components are shielded with a medium probability (for example, 50%). Among the remaining executable actions, the second strategy is determined as the target energy consumption optimization strategy. On the basis of ensuring the normal operation of the equipment, the second strategy further optimizes the energy consumption of each system. For example, the power of the water massage system is reduced by 15%, the frequency of bubble generation is reduced by 10%, and the brightness of the lighting atmosphere system is appropriately reduced by 10% to reduce energy consumption.

[0162] When the equipment operation characteristic data is the third operation characteristic data, the energy consumption contribution of the equipment is low and the corrosion situation is relatively mild. A more aggressive energy consumption optimization strategy can be adopted. At this time, according to the probabilistic action shielding rule driven by the corrosion degree, the actions that may cause the corrosion of equipment components to increase are shielded with a lower probability (for example, 20%). Among the remaining executable actions, the third strategy is determined as the target energy consumption optimization strategy. The third strategy reduces energy consumption as much as possible while ensuring that the equipment operation performance is not affected. For example, the power of the water massage system is reduced by 20%, the frequency of bubble generation is reduced by 20%, the set temperature of the temperature control system is lowered by 1°C, the brightness of the lighting atmosphere system is dimmed by 20%, and the volume of the audio playback system is reduced by 10%.

[0163] The specific steps for resource allocation for the target energy consumption optimization strategy are as follows:

[0164] The power resources of the equipment are rationally allocated based on the adjustment requirements of the various equipment components involved in the target energy consumption optimization strategy. For example, if the target strategy requires reducing the power of the water massage system by 15%, the power allocated to the water massage system in the power distribution module is adjusted from the original 4kW to 3.4kW, and the saved 0.6kW of power is allocated to other systems that require increased power or as a backup power resource.

[0165] At the same time, adjust the control signal resources of the equipment. For example, for a temperature control system, adjust the control signals of the heater and refrigerator according to the requirements of the target strategy so that they can operate stably under the new operating parameters. This may involve changing the start and stop frequency of the heater and refrigerator, adjusting their power output, and other control signal adjustments.

[0166] In addition, the maintenance resources of the equipment are rationally planned. Since the implementation of the energy consumption optimization strategy will affect the operating status and life of the equipment components, the corresponding maintenance plan is arranged according to the characteristics of the target strategy. For example, if the target strategy reduces the power of the water massage system, it may cause the operating pressure of the water pump to change. It is necessary to increase the number of inspections of the water pump to check whether it has abnormal vibration, water leakage, etc., and perform maintenance in a timely manner.

[0167] When the resource allocation results meet the preset energy consumption optimization conditions, the spa equipment is controlled according to the target energy consumption optimization strategy. The preset energy consumption optimization conditions include:

[0168] The operating parameters of each component of the equipment are within the allowable range. After the power of the water massage system is adjusted, the water flow rate cannot be lower than the minimum water flow rate required for normal operation of the equipment, otherwise it may affect the water massage effect; after the water temperature of the temperature control system is adjusted, it must be maintained in a temperature range suitable for human use, such as 36℃-40℃.

[0169] The overall energy consumption of the equipment is reduced by a certain percentage. For example, according to the target energy consumption optimization strategy, it is expected that energy consumption can be reduced by 10%. After actual resource allocation, the energy consumption data of the equipment is monitored in real time to verify whether its energy consumption reduction ratio reaches or is close to the expected value. If the actual energy consumption reduction ratio is lower than 80% of the expected value, it is necessary to re-evaluate the resource allocation results and possibly adjust the target energy consumption optimization strategy.

[0170] The equipment's operating performance indicators must meet the requirements. For example, the massage intensity of the water massage system, the bubble density and uniformity of the bubble generation system, the brightness and color effects of the lighting atmosphere system, and the sound quality of the audio playback system must meet user requirements and equipment design standards. Performance indicators are evaluated through user feedback and the equipment's own performance monitoring system. If a performance indicator is found to have dropped beyond the allowable range, such as a drop in massage intensity of more than 10%, the control strategy needs to be revised to balance the relationship between energy consumption optimization and equipment operating performance.

[0171] When the above preset energy consumption optimization conditions are met, the target energy consumption optimization strategy is converted into a specific device control signal and sent to each execution component of the device, for example:

[0172] For the water massage system, the control signal adjusts the speed of the water pump to reduce its power to the target value. This can be achieved through a variable frequency speed regulator. The corresponding water pump speed is calculated according to the target power value, and the corresponding frequency control signal is sent to the variable frequency speed regulator.

[0173] For temperature control systems, control signals adjust the power output and valve opening of the heater and cooler so that the water temperature is adjusted according to the new operating parameters. For example, the heating power of the heater is adjusted by a thyristor power regulator, and the refrigerant flow rate is changed by an electric control valve, thereby achieving precise control of the water temperature.

[0174] For the bubble generation system, the control signal adjusts the air supply pressure of the air pump and the opening frequency of the solenoid valve to reduce the frequency of bubble generation or change the bubble size. For example, through the pressure transmitter and solenoid valve driver, the corresponding control signal is sent according to the target bubble parameters to achieve energy-optimized operation of the bubble system.

[0175] For the lighting atmosphere system, the control signal adjusts the brightness and color of the lighting module. For example, a PWM (pulse width modulation) signal or a DMX512 signal is sent through the lighting controller to adjust the brightness and RGB color value of the light, so as to create a suitable atmosphere effect while reducing energy consumption.

[0176] For the audio playback system, the control signal adjusts the volume and audio signal processing parameters of the audio playback device. For example, a control signal is sent through the volume adjustment interface of the audio amplifier to reduce the volume output. At the same time, the audio signal spectrum characteristics are adjusted through the audio signal processor to optimize the sound quality effect to compensate for the sound quality degradation that may be caused by the volume reduction.

[0177] Taking a large-scale integrated spa equipment in a spa center as an example, during operation, the equipment collects its energy consumption data and equipment status data in real time through installed sensors. After wavelet decomposition processing, the transient energy consumption characteristics are extracted as [0.8, 1.2, 0.5] (unit: kW), and the current harmonic distortion rate is calculated to be 15%.

[0178] The equipment operation mode database is queried to determine that the initial energy consumption optimization strategy includes three candidate energy consumption optimization strategies: Strategy A (reduce the power of the water massage system by 20% and increase the power of the bubble generation system by 10%), Strategy B (reduce the power of the water massage system by 10% and reduce the power of the lighting atmosphere system by 15%), and Strategy C (increase the power of the temperature control system by 5% and reduce the power of the bubble generation system by 20%).

[0179] The parameters of the three candidate energy consumption optimization strategies are input into the energy consumption benefit estimation model respectively, and the estimated energy consumption benefit data are obtained: strategy A is 8%, strategy B is 6%, and strategy C is 5%.

[0180] The macro-micro motion hierarchical data of the spa equipment were obtained. The macro motions included the water circulation motion of the water massage system, with a frequency of 10 times per minute and an amplitude of 0.5m, and the light flashing motion of the light atmosphere system, with a frequency of 2 times per second and a brightness variation amplitude of 30%. The micro motions included the bubble generation motion of the bubble generation system, with a frequency of 50 times per second and a bubble size variation amplitude of 10%, and the heating motion of the heating tube of the temperature control system, with a frequency of 3 times per minute and a temperature variation amplitude of 2°C. At the same time, the corrosion data of the equipment components were detected: the corrosion degree of the water massage system components was 10%, the corrosion degree of the bubble generation system components was 8%, the corrosion degree of the light atmosphere system components was 5%, and the corrosion degree of the temperature control system components was 7%.

[0181] A benefit analysis is performed on the candidate energy consumption optimization strategies. The preset energy consumption benefit threshold is set to 6%. The estimated energy consumption benefits of strategies A and B are greater than the threshold, and they are determined to be intermediate energy consumption optimization strategies.

[0182] Calculate the energy contribution of the equipment's motion: the water massage system contributes 40%, the bubble generation system contributes 30%, the lighting and atmosphere system contributes 20%, and the temperature control system contributes 10%. Combined with the corrosion level, calculate the corrosion impact coefficient: the water massage system's corrosion impact coefficient is 0.1, the bubble generation system's is 0.08, the lighting and atmosphere system's is 0.05, and the temperature control system's is 0.07. The overall assessment value calculated using the formula is 0.18.

[0183] The first threshold value of the preset device operation characteristic threshold is set to 0.2, and the second threshold value is set to 0.1. Since the comprehensive evaluation value 0.18 is less than the first threshold value 0.2 and greater than the second threshold value 0.1, the device operation characteristic data is determined to be the second operation characteristic data.

[0184] According to the probabilistic action shielding rule driven by the corrosion degree, strategy B in the intermediate energy consumption optimization strategy is determined to be the target energy consumption optimization strategy that meets the preset matching conditions.

[0185] Perform resource allocation for Strategy B and adjust the power distribution module to reduce the power of the water massage system by 10% and the power of the lighting atmosphere system by 15%. When the resource allocation results meet the conditions for normal equipment operation and energy consumption reduction of more than 6%, send a control signal to the device to achieve energy-optimized operation according to Strategy B.

[0186] The same or similar reference numerals correspond to the same or similar components;

[0187] The terms used in the drawings to describe positional relationships are for illustrative purposes only and should not be construed as limiting this patent;

[0188] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in the field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to enumerate all the implementation methods here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A method for optimizing energy consumption of spa equipment based on reinforcement learning, characterized in that: The method comprises: Acquire real-time energy consumption data and device status data of the hydrotherapy equipment during operation, process the real-time energy consumption data based on a wavelet decomposition algorithm, extract transient energy consumption characteristics, and calculate the current harmonic distortion rate of the equipment as an indicator of the equipment operation status; Determining an initial energy consumption optimization strategy based on the transient energy consumption characteristics, the current harmonic distortion rate, and a preset device operation mode library, wherein the initial energy consumption optimization strategy includes multiple candidate energy consumption optimization strategies; Obtain the strategy parameters corresponding to each candidate energy consumption optimization strategy and the energy consumption benefit estimation model built based on reinforcement learning; Inputting the policy parameters corresponding to each candidate energy consumption optimization strategy into the energy consumption benefit estimation model for benefit estimation processing to obtain estimated energy consumption benefit data for each candidate energy consumption optimization strategy; Obtaining macro-micro motion layered data and corrosion data of the spa equipment components, performing decision analysis based on the estimated energy consumption benefit data, macro-micro motion layered data, and corrosion data of each candidate energy consumption optimization strategy, and selecting a target energy consumption optimization strategy from the multiple candidate energy consumption optimization strategies; Performing resource allocation on the target energy consumption optimization strategy to obtain a resource allocation result; When the resource configuration result meets the preset energy consumption optimization condition, the hydrotherapy equipment is controlled according to the target energy consumption optimization strategy.

2. The optimization method according to claim 1, characterized in that Determining the initial energy consumption optimization strategy includes: Obtaining a device operation mode database, wherein the device operation mode database includes a preset transient energy consumption characteristic range, a current harmonic distortion rate threshold, and a corresponding energy consumption optimization strategy; Based on the transient energy consumption characteristics and current harmonic distortion rate, the equipment operation mode database is queried to obtain a preset energy consumption optimization strategy that matches the current equipment operation status, and the matching preset energy consumption optimization strategy is determined to be the initial energy consumption optimization strategy; the equipment operation mode database is constructed based on historical equipment operation status data and corresponding effective energy consumption optimization strategies.

3. The optimization method according to claim 2, characterized in that The method for constructing the equipment operation mode database includes: Acquire a sample set of equipment operating status information, wherein the sample set includes transient energy consumption characteristic samples and current harmonic distortion rate samples under different working conditions; Setting at least one corresponding energy consumption optimization strategy based on each device operation status information sample; The device operation mode database is constructed based on each device operation status information sample and the energy consumption optimization strategy corresponding to each sample.

4. The optimization method according to claim 1, characterized in that The decision analysis based on the estimated energy consumption benefit data, macro-micro action hierarchical data and corrosion degree data of each candidate energy consumption optimization strategy, and screening out a target energy consumption optimization strategy from the multiple candidate energy consumption optimization strategies includes: Performing benefit analysis on the estimated energy consumption benefit data of each candidate energy consumption optimization strategy to obtain the estimated energy consumption benefit of each candidate energy consumption optimization strategy; In the case where the estimated energy consumption benefit is greater than the preset energy consumption benefit threshold, determining the candidate energy consumption optimization strategy corresponding to the estimated energy consumption benefit greater than the preset energy consumption benefit threshold as the intermediate energy consumption optimization strategy; Performing feature evaluation on the macro-micro motion hierarchical data and combining it with corrosion data of equipment components to obtain equipment operation feature data; A matching analysis is performed on the intermediate energy consumption optimization strategy and the equipment operation characteristic data, and based on a probabilistic action shielding rule driven by the corrosion degree, the intermediate energy consumption optimization strategy whose matching analysis result meets the preset matching conditions is determined as the target energy consumption optimization strategy.

5. The optimization method according to claim 4, characterized in that: Obtaining the equipment operation characteristic data includes: Calculate the energy consumption contribution of equipment motion based on macro-micro motion hierarchical data, and calculate the corrosion influence coefficient based on the corrosion degree of equipment components; Calculate the comprehensive evaluation value based on the equipment's energy consumption contribution value and corrosion influence coefficient; The comprehensive evaluation value is compared with a preset device operation characteristic threshold to obtain a comparison result, and the device operation characteristic data is determined based on the comparison result.

6. The optimization method according to claim 5, characterized in that: The preset device operation characteristic threshold includes a first threshold and a second threshold, the first threshold is greater than the second threshold, the device operation characteristic data includes first operation characteristic data, second operation characteristic data, and third operation characteristic data, and comparing the comprehensive evaluation value with the preset device operation characteristic threshold to obtain a comparison result, and determining the device operation characteristic data based on the comparison result includes: When the comprehensive evaluation value is greater than or equal to the first threshold, determining the device operation characteristic data as the first operation characteristic data; When the comprehensive evaluation value is less than the first threshold and greater than or equal to the second threshold, determining that the device operation characteristic data is the second operation characteristic data; When the comprehensive evaluation value is less than the second threshold, the device operation characteristic data is determined to be the third operation characteristic data.

7. The optimization method according to claim 4, characterized in that: The intermediate energy consumption optimization strategy includes a first strategy, a second strategy, and a third strategy; the equipment operation characteristic data includes first operation characteristic data, second operation characteristic data, and third operation characteristic data; performing matching analysis on the intermediate energy consumption optimization strategy and the equipment operation characteristic data, and determining, based on a probabilistic action shielding rule driven by the corrosion degree, the intermediate energy consumption optimization strategy whose matching analysis result satisfies a preset matching condition as the target energy consumption optimization strategy, includes: When the device operation characteristic data is the first operation characteristic data, determining the first strategy as the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree; When the device operation characteristic data is the second operation characteristic data, determining the second strategy as the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree; When the equipment operation characteristic data is the third operation characteristic data, the third strategy is determined to be the target energy consumption optimization strategy according to a probabilistic action shielding rule driven by the corrosion degree.

8. The optimization method according to claim 1, characterized in that: The training method of the energy consumption benefit estimation model includes: Obtain an initial energy consumption benefit estimation model based on an Actor network constrained by thermodynamic equations and a lightweight spiking neural network; Inputting the strategy parameters of the sample energy consumption optimization strategy into the initial energy consumption benefit estimation model to perform benefit estimation processing to obtain estimated energy consumption benefit data corresponding to the sample energy consumption optimization strategy; Matching the energy consumption benefit label data of the sample energy consumption optimization strategy with the estimated energy consumption benefit data of the corresponding sample energy consumption optimization strategy to obtain a matching result; If the matching result does not meet the preset condition, determining the difference information between the energy consumption benefit label data of the sample energy consumption optimization strategy and the estimated energy consumption benefit data of the corresponding sample energy consumption optimization strategy; Based on the difference information, the network parameters of the initial energy consumption benefit estimation model are updated, and the network parameters are constrained and adjusted in combination with domain knowledge until the matching result meets the preset conditions, and the initial energy consumption benefit estimation model corresponding to the matching result that meets the preset conditions is determined to be the energy consumption benefit estimation model.

9. An electronic device, characterized in that: It includes a processor and a memory, wherein the memory stores at least one instruction or at least one program, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the energy consumption optimization method of spa equipment based on reinforcement learning as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions. When the computer instructions are executed, the method for optimizing energy consumption of hydrotherapy equipment based on reinforcement learning according to any one of claims 1 to 8 is executed.

Images