Heat pump water heater temperature control method based on artificial intelligence

Abstract

The invention discloses a heat-pump water heater temperature control method based on artificial intelligence, relates to the technical field of heat-pump water heater temperature control, and aims to solve the problem that the COP (coefficient of performance) is reduced due to a low-temperature environment, the exhaust temperature of a compressor and the opening degree of an expansion valve are optimized in real time through a GA (genetic algorithm), and operating parameters of a heat-pump water heater are dynamically adjusted by taking COP maximization as a target; the operation frequency of a compressor and the flow of a water pump are dynamically adjusted through fuzzy control logic in combination with a GA optimization result, the current water temperature and the temperature set by a user, the stability of the water supply temperature is guaranteed, an optimal control strategy is calculated through defuzzification, and the adaptability under different environment conditions is improved; an ANN is adopted to learn and predict a user behavior pattern, an LSTM network is utilized to analyze time sequences of historical water consumption data, target temperature and environment temperature, future water consumption peak and temperature requirements are predicted, a preheating function is started in advance, and insufficient hot water supply in the water consumption peak is avoided.

Description

Technical Field

[0001] The invention relates to the technical field of temperature control of heat pump water heaters, and in particular to a temperature control method of a heat pump water heater based on artificial intelligence. Background Art

[0002] Heat pump water heaters are widely used in residential and commercial places. They rely on extracting heat from the ambient air and converting low-grade thermal energy into high-temperature hot water through a heat pump cycle, thereby achieving efficient and energy-saving hot water supply.

[0003] Existing heat pump water heaters are mostly based on fixed logic for temperature control, such as switch control or PID adjustment, which achieves operating water temperature control by maintaining the target water temperature within the set range. However, this method is not capable of coping in low-temperature environments. In winter or extremely cold conditions, the sudden drop in outside temperature will cause the operating efficiency of the heat pump to drop significantly. The heat pump system needs to overcome a larger temperature difference to heat the cold water to the set temperature. The energy consumption of the compressor increases significantly, and the energy efficiency ratio (COP) decreases significantly. It is easy to have insufficient hot water supply, resulting in increased system operating costs.

[0004] For the above-mentioned low-temperature scenarios, auxiliary electric heating is currently used as a supplement to maintain the set water temperature by increasing the heating capacity of the water heater. However, this method will obviously increase energy consumption significantly, which runs counter to the original intention of the heat pump to save energy. Therefore, there is an urgent need for an artificial intelligence-based heat pump water heater temperature control method to solve this problem. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides a heat pump water heater temperature control method based on artificial intelligence to solve

[0007] When the outdoor temperature is low in winter, the heating efficiency of traditional heat pump water heaters drops significantly, resulting in increased operating energy consumption.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0009] The embodiment of the present invention provides a temperature control method for a heat pump water heater based on artificial intelligence, which comprises:

[0010] Step S1, real-time monitoring of environmental data, including outdoor temperature, humidity, and water inlet and outlet temperature difference of the water tank, is performed through the thermostat and the online centralized controller. The environmental data is transmitted to the online centralized controller via the ZigBee / ZWave wireless protocol and uploaded to the cloud server via WiFi / wireless communication;

[0011] Step S2, using a genetic algorithm GA, taking energy efficiency ratio COP maximization as the optimization goal, and adjusting the compressor exhaust temperature and expansion valve opening in real time;

[0012] Step S3, introducing fuzzy control logic, and adjusting the operating frequency of the compressor and the flow rate of the water pump in real time through the online centralized controller;

[0013] Step S4, using artificial neural network (ANN) technology to learn and predict the user's behavior pattern on the central control screen, and at the same time perform intelligent fault monitoring on the heat pump water heater.

[0014] As a preferred solution of the temperature control method of a heat pump water heater based on artificial intelligence described in the present invention, the calculation logic in step S2 is:

[0015] The cloud server receives environmental data, uses GA for iterative optimization, searches for the best operating parameters and generates optimization instructions.

[0016] The online centralized controller receives optimization instructions and performs dynamic adjustments.

[0017] As a preferred solution of the artificial intelligence-based heat pump water heater temperature control method of the present invention, the step of adjusting the compressor exhaust temperature and the expansion valve opening in real time with the energy efficiency ratio COP maximization as the optimization goal is as follows:

[0018] Define the energy efficiency ratio COP, the calculation formula of COP is:

[0019] COP=Q h / W c ,

[0020] Among them, COP represents the energy efficiency ratio of the heat pump, Q h Indicates the total heat output of the heat pump per unit time, W c represents the power consumption of the compressor,

[0021] Definition of output heat Q h , Q h The calculation formula is:

[0022] Q h =m·c·ΔT,

[0023] Where m represents the mass flow rate of water, c represents the specific heat capacity of water, ΔT represents the temperature difference between the inlet and outlet water of the water tank, and T out Indicates the water outlet temperature of the water tank, T in Indicates the water inlet temperature of the water tank, satisfying ΔT=T out -T in ;

[0024] Compressor power consumption W c The calculation formula is:

[0025] W c =P c +P e ,

[0026] Among them, P c Indicates the compressor power, satisfying P c =M·R·T / η c , M represents the mass flow rate of gas in the compressor, R represents the gas constant, T represents the compressor exhaust temperature, η c represents the adiabatic efficiency of the compressor, P e Indicates the additional power consumption of the expansion valve;

[0027] Construct the genetic algorithm objective function, the objective function is:

[0028]

[0029] Where x represents the set of parameters that need to be optimized, including the compressor exhaust temperature T and the expansion valve opening α;

[0030] Through the objective function f(x) of maximizing the energy efficiency ratio, the parameter set x = {T, α} is optimized to adjust the exhaust temperature of the compressor and the opening of the expansion valve in real time under environmental changes;

[0031] The genetic algorithm parameter optimization steps are:

[0032] Define the initial parameter set x = {T, α},

[0033] Substitute each set of parameters into the objective function f(x) to calculate the fitness value.

[0034] Select the parent generation according to the fitness value probability, and cross-generate the offspring.

[0035] Random variation parameter T and expansion valve opening α,

[0036] If the number of iterations reaches the threshold or the fitness value does not change significantly, output the optimal solution x * ={T * ,α *};

[0037] The online centralized controller receives the genetic algorithm optimization instructions and adjusts the actual operating parameters.

[0038] As a preferred solution of the artificial intelligence-based temperature control method for heat pump water heater described in the present invention, the input variables of the fuzzy control logic include the optimization instructions of the genetic algorithm, the current water temperature and the target temperature set by the user, and logical inference is performed on the input variables to dynamically output the adjustment strategy; allowing the user to control the equipment through the central control screen.

[0039] As a preferred solution of the temperature control method of a heat pump water heater based on artificial intelligence described in the present invention, the step of introducing fuzzy control logic and adjusting the operating frequency of the compressor and the water pump flow rate in real time through the online centralized controller is as follows:

[0040] Define the input variables of fuzzy control. The input variables of fuzzy control include:

[0041] The target compressor exhaust temperature T given in the genetic algorithm optimization instruction target and expansion valve opening α target , current water temperature T current , and the user-set target temperature T user ;

[0042] Define the error E and error change rate ΔE as:

[0043] E=T user -T current ,

[0044] ΔE=dE / dt,

[0045] Where E represents the current temperature error, T user Indicates the target temperature set by the user, T current represents the current water temperature, ΔE represents the rate of change of error, and t represents time;

[0046] Design fuzzy rules, the output variables of the fuzzy controller include the compressor operating frequency f c and pump flow q p , the input set of fuzzy rules includes error E and error change rate ΔE,

[0047] If E is large and ΔE is positive, increase f c And increase q p ,

[0048] If E is zero and ΔE is zero, then f holds c and q p constant,

[0049] The membership function μ is defined as:

[0050] If E∈ large positive interval, then μ(E)=1, if E Valid interval, then μ(E)=0,

[0051] Among them, μ(E) represents the membership of E to the fuzzy set, and the value of μ ranges from 0 to 1;

[0052] The control quantity f output by fuzzy reasoning c and q pis a fuzzy set. The actual value is calculated by defuzzification and the final value is calculated by weighted average method. The calculation formula is:

[0053]

[0054] Among them, μ i represents the membership degree of the i-th fuzzy rule, f c，i represents the compressor frequency corresponding to the i-th rule, q p，i represents the pump flow rate corresponding to the i-th rule, i represents the fuzzy rule number, and there are n rules in total;

[0055] The online centralized controller adjusts the actual operating parameters according to the defuzzification results. The adjustment formula is:

[0056] f c →f current ,q p →q current ,

[0057] Among them, f current Indicates the current operating frequency of the compressor q current Indicates the current flow rate of the pump.

[0058] As a preferred solution of the artificial intelligence-based temperature control method for a heat pump water heater described in the present invention, the behavior pattern includes historical user settings and operation data recorded by the central control screen.

[0059] As a preferred solution of the temperature control method of a heat pump water heater based on artificial intelligence described in the present invention, the learning and prediction methods are as follows:

[0060] ANN uses time series analysis to predict future peak water usage periods and temperature demands, and activates the preheating function before the peak period.

[0061] As a preferred solution of the artificial intelligence-based heat pump water heater temperature control method of the present invention, the steps of using artificial neural network ANN technology to learn and predict the user's behavior pattern on the central control screen are:

[0062] The input layer X receives multidimensional time series data, represented as:

[0063] X=[T user (t), H(t), T env (t)],

[0064] Among them, T user (t) represents the target temperature time series set by the user, H(t) represents the historical water consumption time series, T env (t) represents the ambient temperature time series, t represents time,

[0065] Perform hidden layer modeling and use LSTM long short-term memory network to process time series data. The processing formula is:

[0066] h t =σ(W h ·[X t ,h t-1 ]+b h ),

[0067] Among them, h t represents the output of the hidden layer at time t, W h represents the weight matrix of the hidden layer, X t represents the data of the input matrix at time t, h t-1 represents the hidden layer state at the previous moment, b h represents the bias vector, σ represents the activation function;

[0068] The output layer predicts the future peak water consumption period t peak and target temperature T peak , denoted by Y:

[0069] Y=W o ·h t +b o ,

[0070] Among them, Y represents the output vector, including t peak and T peak , W o represents the output layer weight matrix, b o Represents the output layer bias vector;

[0071] The loss function is optimized and defined as the mean square error MSE loss function, expressed as:

[0072]

[0073] Among them, L represents the loss value, Y pred，i represents the predicted value of the i-th sample, Y true，i represents the true value of the i-th sample, N represents the total number of training samples, and i represents the sample number;

[0074] Adjust the weight W through the back-propagation algorithm h , W o and bias b h , b o .

[0075] As a preferred solution of the temperature control method of a heat pump water heater based on artificial intelligence described in the present invention, the method of performing intelligent fault monitoring is:

[0076] The equipment operating parameters are monitored, including the compressor operating frequency, water pump temperature and current parameters. The equipment operating parameters are uploaded to the cloud server. ANN identifies abnormal modes based on the equipment operating parameters, including high-frequency start and stop or water pump overheating, and issues an alarm on the central control screen, while automatically switching to protection mode.

[0077] As a preferred solution of the temperature control method of a heat pump water heater based on artificial intelligence described in the present invention, the step of performing intelligent fault monitoring on the heat pump water heater is:

[0078] Define monitoring parameters. The monitoring equipment operation parameters include:

[0079] Compressor operating frequency f c , water pump temperature T p , pump current I p And the start-stop frequency F on / off ;

[0080] ANN is used for abnormal pattern recognition, and the input vector X is defined as:

[0081] X=[f c , T p , I p , F on / off ],

[0082] The intermediate state h of the model is calculated through the hidden layer t :

[0083] h t =tanh(W h ·X+b h ),

[0084] Among them, h t represents the hidden layer output, W h represents the weight matrix, b h represents bias, tanh represents activation function;

[0085] The output layer Y is defined as the anomaly probability:

[0086] Y=σ(W o ·h t +b o ),

[0087] Among them, Y represents the output result, the range is [0, 1], and it represents the probability of abnormality. o represents the output layer weight, b o represents the output layer bias, σ represents the Sigmoid activation function,

[0088] If Y>θ, it is judged as abnormal, and θ is the threshold;

[0089] Based on the model’s output probabilities, identify abnormal patterns, including:

[0090] If F on / off >F threshold , then high frequency start and stop,

[0091] If T p >T threshold , the water pump is overheated.

[0092] like The current is abnormal;

[0093] When an anomaly is detected, protection strategies are executed, including:

[0094] The central control screen issues an alarm message, indicating the specific fault type and recommended treatment measures.

[0095] Adjust operating parameters and reduce compressor frequency f c and pump flow q p ,

[0096] Upload the operation data before and after the abnormality occurs to the cloud;

[0097] After the fault is resolved, the ANN model is retrained using the newly added data.

[0098] The beneficial effects of the present invention are as follows: in view of the problem that the energy efficiency ratio (COP) decreases due to low temperature environment, the present invention optimizes the compressor exhaust temperature and the expansion valve opening in real time through GA, and dynamically adjusts the operating parameters of the heat pump water heater with the goal of maximizing COP; at the same time, a COP calculation formula is defined, and the output heat and compressor power consumption are comprehensively affected to make the energy efficiency optimization more accurate; in addition, through fuzzy control logic, combined with GA optimization results, current water temperature and user-set temperature, the compressor operating frequency and water pump flow are dynamically adjusted to ensure the stability of the water supply temperature, and the optimal control strategy is calculated through defuzzification to improve the adaptability under different environmental conditions, further improving user experience and equipment safety; ANN is used to learn and predict user behavior patterns, and LSTM network is used to analyze the time series of historical water use data, target temperature and ambient temperature, predict future water use peak and temperature demand, start the preheating function in advance, and avoid insufficient hot water supply during water use peak period; at the same time, ANN is also used for intelligent fault monitoring, identifying abnormal modes through equipment operating parameters, and issuing alarms on the central control screen, adjusting operating parameters in real time or switching to protection mode to reduce fault risks.

[0099] In summary, the present invention effectively overcomes the shortcomings of traditional control methods in low-temperature environments, optimizes the operating efficiency of heat pumps, and significantly reduces energy consumption and maintenance costs through intelligent prediction and fault monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0101] Figure 1 It is a schematic flow chart of the temperature control method of a heat pump water heater based on artificial intelligence of the present invention. DETAILED DESCRIPTION

[0102] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0103] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0104] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0105] Example 1, reference Figure 1 , this embodiment provides a temperature control method for a heat pump water heater based on artificial intelligence, comprising the following steps:

[0106] Step S1, real-time monitoring of environmental data, including outdoor temperature, humidity, and water inlet and outlet temperature difference of the water tank, is performed through the thermostat and the online centralized controller. The environmental data is transmitted to the online centralized controller via the ZigBee / ZWave wireless protocol and uploaded to the cloud server via WiFi / wireless communication;

[0107] Step S2, using a genetic algorithm GA, taking energy efficiency ratio COP maximization as the optimization goal, and adjusting the compressor exhaust temperature and expansion valve opening in real time;

[0108] The calculation logic in step S2 is:

[0109] The cloud server receives environmental data, uses GA for iterative optimization, searches for the best operating parameters and generates optimization instructions.

[0110] The online centralized controller receives the optimization instruction and performs dynamic adjustment;

[0111] Taking the energy efficiency ratio (COP) maximization as the optimization goal, the steps of real-time adjustment of the compressor exhaust temperature and expansion valve opening are as follows:

[0112] Define the energy efficiency ratio COP, the calculation formula of COP is:

[0113] COP=Q h / W c ,

[0114] Among them, COP represents the energy efficiency ratio of the heat pump, Q h Indicates the total heat output of the heat pump per unit time, W c Indicates the power consumption of the compressor.

[0115] Definition of output heat Q h , Q h The calculation formula is:

[0116] Q h =m·c·ΔT,

[0117] Where m represents the mass flow rate of water, c represents the specific heat capacity of water, ΔT represents the temperature difference between the inlet and outlet water of the water tank, and T out Indicates the water outlet temperature of the water tank, T in Indicates the water inlet temperature of the water tank, satisfying ΔT=T out -T in ;

[0118] Compressor power consumption W c The calculation formula is:

[0119] W c =P c +P e ,

[0120] Among them, P c Indicates the compressor power, satisfying P c =M·R·T / η c , M represents the mass flow rate of gas in the compressor, R represents the gas constant, T represents the compressor exhaust temperature, η c represents the adiabatic efficiency of the compressor, P e Indicates the additional power consumption of the expansion valve;

[0121] Construct the genetic algorithm objective function, the objective function is:

[0122]

[0123] Where x represents the set of parameters that need to be optimized, including the compressor exhaust temperature T and the expansion valve opening α;

[0124] By maximizing the objective function f(x) of the energy efficiency ratio, the parameter set x = {T, α} is optimized to adjust the exhaust temperature of the compressor and the opening of the expansion valve in real time under environmental changes;

[0125] The genetic algorithm parameter optimization steps are:

[0126] Define the initial parameter set x = {T, α},

[0127] Substitute each set of parameters into the objective function f(x) to calculate the fitness value.

[0128] Select the parent generation according to the fitness value probability, and cross-generate the offspring.

[0129] Random variation parameter T and expansion valve opening α,

[0130] If the number of iterations reaches the threshold or the fitness value does not change significantly, output the optimal solution x * ={T * , α *};

[0131] The online centralized controller receives the genetic algorithm optimization instructions and adjusts the actual operating parameters;

[0132] Specifically, the genetic algorithm continuously optimizes the parameter set x in the optimization step to dynamically maximize the energy efficiency ratio COP. The optimization result x * Upload to the online centralized controller to adjust the compressor exhaust temperature and expansion valve opening in real time; through the online centralized controller, the compressor exhaust temperature and expansion valve opening are kept in the optimal state by dynamically responding to environmental changes, which can maximize the energy efficiency ratio (COP) of the heat pump and improve the overall operating efficiency.

[0133] Step S3, introducing fuzzy control logic, and adjusting the operating frequency of the compressor and the flow rate of the water pump in real time through the online centralized controller;

[0134] The input variables of the fuzzy control logic include the optimization instructions of the genetic algorithm, the current water temperature, and the target temperature set by the user. The input variables are logically inferred and the adjustment strategy is dynamically output. The user is allowed to control the equipment through the central control screen.

[0135] The steps of introducing fuzzy control logic and adjusting the compressor operating frequency and water pump flow rate in real time through the online centralized controller are as follows:

[0136] Define the input variables of fuzzy control. The input variables of fuzzy control include:

[0137] The target compressor exhaust temperature T given in the genetic algorithm optimization instruction target and expansion valve opening α target , current water temperature T current , and the user-set target temperature T user;

[0138] Define the error E and error change rate ΔE as:

[0139] E=T user -T current ,

[0140] ΔE=dE / dt,

[0141] Where E represents the current temperature error, T user Indicates the target temperature set by the user, T current represents the current water temperature, ΔE represents the rate of change of error, and t represents time;

[0142] Design fuzzy rules, the output variables of the fuzzy controller include the compressor operating frequency f c and pump flow q p , the input set of fuzzy rules includes error E and error change rate ΔE,

[0143] If E is large and ΔE is positive, increase f c And increase q p ,

[0144] If E is zero and ΔE is zero, then f holds c and q p constant,

[0145] The membership function μ is defined as:

[0146] If E∈ large positive interval, then μ(E)=1, if E Valid interval, then μ(E)=0,

[0147] Among them, μ(E) represents the membership of E to the fuzzy set, and the value of μ ranges from 0 to 1;

[0148] The control quantity f output by fuzzy reasoning c and q p is a fuzzy set. The actual value is calculated by defuzzification and the final value is calculated by weighted average method. The calculation formula is:

[0149]

[0150] Among them, μ i represents the membership degree of the i-th fuzzy rule, f c，i represents the compressor frequency corresponding to the i-th rule, q p，i represents the pump flow rate corresponding to the i-th rule, i represents the fuzzy rule number, and there are n rules in total;

[0151] The online centralized controller adjusts the actual operating parameters according to the defuzzification results. The adjustment formula is:

[0152] f c →f current ,q p →q current ,

[0153] Among them, f current Indicates the current operating frequency of the compressor q current Indicates the current flow of the water pump;

[0154] Specifically, the fuzzy control logic is based on the input variables E and ΔE, combined with the fuzzy rule base for reasoning, and the defuzzification method is used to calculate the actual output value. The online centralized controller dynamically adjusts the compressor operating frequency and water pump flow rate according to the calculation results to maintain efficient operation under different environmental conditions.

[0155] Step S4, using artificial neural network ANN technology to learn and predict the user's behavior pattern on the central control screen, and at the same time perform intelligent fault monitoring on the heat pump water heater;

[0156] The behavior mode includes historical user settings and operation data recorded by the central control screen;

[0157] The way to learn and predict is:

[0158] ANN uses time series analysis to predict future peak water usage periods and temperature requirements, and activates the preheating function before the peak period;

[0159] The steps of using artificial neural network ANN technology to learn and predict the user's behavior pattern on the central control screen are as follows:

[0160] The input layer X receives multidimensional time series data, represented as:

[0161] X=[T user (t), H(t), T env (t)],

[0162] Among them, T user (t) represents the target temperature time series set by the user, H(t) represents the historical water consumption time series, T env (t) represents the ambient temperature time series, t represents time,

[0163] Perform hidden layer modeling and use LSTM long short-term memory network to process time series data. The processing formula is:

[0164] h t =σ(W h ·[X t ,h t-1 ]+b h ),

[0165] Among them, ht represents the output of the hidden layer at time t, W h represents the weight matrix of the hidden layer, X t represents the data of the input matrix at time t, h t-1 represents the hidden layer state at the previous moment, b h represents the bias vector, σ represents the activation function;

[0166] The output layer predicts the future peak water consumption period t peak and target temperature T peak , denoted by Y:

[0167] Y=W o ·h t +b o ,

[0168] Among them, Y represents the output vector, including t peak and T peak , W o represents the output layer weight matrix, b o Represents the output layer bias vector;

[0169] The loss function is optimized and defined as the mean square error MSE loss function, expressed as:

[0170]

[0171] Among them, L represents the loss value, Y pred，i represents the predicted value of the i-th sample, Y true，i represents the true value of the i-th sample, N represents the total number of training samples, and i represents the sample number;

[0172] Adjust the weight W through the back-propagation algorithm h , W o and bias b h , b o ;

[0173] Specifically, ANN combines time series data of user behavior patterns and uses LSTM network to capture the temporal correlation between historical behavior and future demand, and finally predicts the future peak water usage time and temperature demand. The loss function ensures the accuracy of the prediction results, helps preheat in advance, and improves user experience.

[0174] The way to carry out intelligent fault monitoring is:

[0175] Monitor equipment operating parameters, including compressor operating frequency, water pump temperature and current parameters. The equipment operating parameters are uploaded to the cloud server. ANN identifies abnormal modes based on the equipment operating parameters, including high-frequency start and stop or water pump overheating, and issues an alarm on the central control screen, while automatically switching to protection mode.

[0176] The steps for intelligent fault monitoring of heat pump water heaters are:

[0177] Define monitoring parameters. The monitoring equipment operation parameters include:

[0178] Compressor operating frequency f c , water pump temperature T p , pump current I p And the start-stop frequency F on / off ;

[0179] ANN is used for abnormal pattern recognition, and the input vector X is defined as:

[0180] X=[f c , T p , I p , F on / off ],

[0181] The intermediate state h of the model is calculated through the hidden layer t :

[0182] h t =tanh(W h ·X+b h ),

[0183] Among them, h t represents the hidden layer output, W h represents the weight matrix, b h represents bias, tanh represents activation function;

[0184] The output layer Y is defined as the anomaly probability:

[0185] Y=σ(W o ·h t +b o ),

[0186] Among them, Y represents the output result, the range is [0, 1], and it represents the probability of abnormality. o represents the output layer weight, b o represents the output layer bias, σ represents the Sigmoid activation function,

[0187] If Y>θ, it is judged as abnormal, and θ is the threshold;

[0188] Based on the model’s output probabilities, identify abnormal patterns, including:

[0189] If F on / off >F threshold , then high frequency start and stop,

[0190] If T p >T thresho1d , the water pump is overheated.

[0191] like The current is abnormal;

[0192] When an anomaly is detected, protection strategies are executed, including:

[0193] The central control screen issues an alarm message, indicating the specific fault type and recommended treatment measures.

[0194] Adjust operating parameters and reduce compressor frequency f c and pump flow q p ,

[0195] Upload the operation data before and after the abnormality occurs to the cloud;

[0196] After the fault is resolved, the ANN model is retrained using the newly added data;

[0197] Specifically, in this step, ANN is used to identify abnormal patterns based on the equipment operating parameters to quickly detect high-frequency start and stop, water pump overheating, and abnormal current faults, reducing the risk of equipment damage; the self-learning process further improves the level of intelligence and makes fault monitoring more robust.

[0198] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A temperature control method for a heat pump water heater based on artificial intelligence, characterized in that: include, Step S1, real-time monitoring of environmental data, including outdoor temperature, humidity, and water inlet and outlet temperature difference of the water tank, is performed through the thermostat and the online centralized controller. The environmental data is transmitted to the online centralized controller via the ZigBee / ZWave wireless protocol and uploaded to the cloud server via WiFi / wireless communication; Step S2, using a genetic algorithm GA, taking energy efficiency ratio COP maximization as the optimization goal, and adjusting the compressor exhaust temperature and expansion valve opening in real time; Step S3, introducing fuzzy control logic, and adjusting the operating frequency of the compressor and the flow rate of the water pump in real time through the online centralized controller; Step S4, using artificial neural network (ANN) technology to learn and predict the user's behavior pattern on the central control screen, and at the same time perform intelligent fault monitoring on the heat pump water heater.

2. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 1, characterized in that: The calculation logic in step S2 is: The cloud server receives environmental data, uses GA for iterative optimization, searches for the best operating parameters and generates optimization instructions. The online centralized controller receives optimization instructions and performs dynamic adjustments.

3. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 2, characterized in that: The steps of adjusting the compressor exhaust temperature and the expansion valve opening in real time with the energy efficiency ratio COP maximization as the optimization goal are as follows: Define the energy efficiency ratio COP, the calculation formula of COP is: COP=Q h / W c , Among them, COP represents the energy efficiency ratio of the heat pump, Q h Indicates the total heat output of the heat pump per unit time, W c represents the power consumption of the compressor, Definition of heat output Q h , Q h The calculation formula is: Q h =m·c·ΔT, Where m represents the mass flow rate of water, c represents the specific heat capacity of water, ΔT represents the temperature difference between the inlet and outlet water of the water tank, and T out Indicates the water outlet temperature of the water tank, T in Indicates the water inlet temperature of the water tank, satisfying ΔT=T out -T in ; Compressor power consumption W c The calculation formula is: W c =P c +P e , Among them, P c Represents the compressor power, satisfying P c =M·R·T / η c , M represents the mass flow rate of gas in the compressor, R represents the gas constant, T represents the compressor exhaust temperature, η c represents the adiabatic efficiency of the compressor, P e Indicates the additional power consumption of the expansion valve; Construct the genetic algorithm objective function, the objective function is: Where x represents the set of parameters that need to be optimized, including the compressor exhaust temperature T and the expansion valve opening α; The genetic algorithm parameter optimization steps are: Define the initial parameter set x = {T, α}, Substitute each set of parameters into the objective function f(x) to calculate the fitness value. Select the parent generation according to the fitness value probability, and cross-generate the offspring. Random variation parameter T and expansion valve opening α, If the number of iterations reaches the threshold or the fitness value does not change significantly, output the optimal solution x * ={T * , α * }; The online centralized controller receives the genetic algorithm optimization instructions and adjusts the actual operating parameters.

4. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 3, characterized in that: The input variables of the fuzzy control logic include the optimization instructions of the genetic algorithm, the current water temperature and the target temperature set by the user, and the input variables are logically inferred to dynamically output the adjustment strategy; Allow users to control the device through the central control screen.

5. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 4, characterized in that: The steps of introducing fuzzy control logic and adjusting the operating frequency of the compressor and the flow rate of the water pump in real time through the online centralized controller are: Define the input variables of fuzzy control. The input variables of fuzzy control include: The target compressor exhaust temperature T given in the genetic algorithm optimization instruction target and expansion valve opening α target , current water temperature T current , and the user-set target temperature T user ; Define the error E and error change rate ΔE as: E=T user -T current , ΔE=dE / dt, Where E represents the current temperature error, T user Indicates the target temperature set by the user, T current represents the current water temperature, ΔE represents the rate of change of error, and t represents time; Design fuzzy rules, the output variables of the fuzzy controller include the compressor operating frequency f c and pump flow q p , the input set of fuzzy rules includes error E and error change rate ΔE, If E is large and ΔE is positive, increase f c And increase q p , If E is zero and ΔE is zero, then f holds c and q p constant, The membership function μ is defined as: If E∈ large positive interval, then μ(E)=1, if E Valid interval, then μ(E)=0, Among them, μ(E) represents the membership of E to the fuzzy set, and the value of μ ranges from 0 to 1; The control quantity f output by fuzzy reasoning c and q p is a fuzzy set. The actual value is calculated by defuzzification and the final value is calculated by weighted average method. The calculation formula is: Among them, μ i represents the membership degree of the i-th fuzzy rule, f c，i represents the compressor frequency corresponding to the i-th rule, q p，i represents the pump flow rate corresponding to the i-th rule, i represents the fuzzy rule number, and there are n rules in total; The online centralized controller adjusts the actual operating parameters according to the defuzzification results. The adjustment formula is: f c →f current ,q p →q current , Among them, f current Indicates the current operating frequency of the compressor q current Indicates the current flow rate of the pump.

6. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 5, characterized in that: The behavior pattern includes historical user settings and operation data recorded by the central control screen.

7. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 6, characterized in that: The way to learn and predict is: ANN uses time series analysis to predict future peak water usage periods and temperature demands, and activates the preheating function before the peak period.

8. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 7, characterized in that: The steps of using artificial neural network ANN technology to learn and predict the user's behavior pattern on the central control screen are: The input layer X receives multidimensional time series data, represented as: X=[T user (t),H(t),T env (t)], Among them, T user (t) represents the target temperature time series set by the user, H(t) represents the historical water consumption time series, T env (t) represents the ambient temperature time series, t represents time, Perform hidden layer modeling and use LSTM long short-term memory network to process time series data. The processing formula is: h t =σ(W h ·[X t ,h t-1 ]+b h ), Among them, h t represents the output of the hidden layer at time t, W h represents the weight matrix of the hidden layer, X t represents the data of the input matrix at time t, h t-1 represents the hidden layer state at the previous moment, b h represents the bias vector, σ represents the activation function; The output layer predicts the future peak water consumption period t peak and target temperature T peak , denoted by Y: Y=W o ·h t +b o , Among them, Y represents the output vector, including t peak and T peak , W o represents the output layer weight matrix, b o Represents the output layer bias vector; The loss function is optimized and defined as the mean square error MSE loss function, expressed as: Among them, L represents the loss value, Y pred,i represents the predicted value of the i-th sample, Y true，i represents the true value of the i-th sample, N represents the total number of training samples, and i represents the sample number; Adjust the weight W through the back-propagation algorithm h , W o and bias b h , b o .

9. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 8, characterized in that: The way to carry out intelligent fault monitoring is: The equipment operating parameters are monitored, including the compressor operating frequency, water pump temperature and current parameters. The equipment operating parameters are uploaded to the cloud server. ANN identifies abnormal modes based on the equipment operating parameters, including high-frequency start and stop or water pump overheating, and issues an alarm on the central control screen, while automatically switching to protection mode.

10. The temperature control method of a heat pump water heater based on artificial intelligence as claimed in claim 9, characterized in that: The steps of performing intelligent fault monitoring on the heat pump water heater are: Define monitoring parameters. The monitoring equipment operation parameters include: Compressor operating frequency f c , water pump temperature T p , pump current I p And the start-stop frequency F on / off ; ANN is used for abnormal pattern recognition, and the input vector X is defined as: X=[f c ,T p ,I p ,F on / off ], The intermediate state h of the model is calculated through the hidden layer t : h t =tanh(W h ·X+b h ), Among them, h t represents the hidden layer output, W h represents the weight matrix, b h represents bias, tanh represents activation function; The output layer Y is defined as the anomaly probability: Y=σ(W o ·h t +b o ), Among them, Y represents the output result, the range is [0, 1], and it represents the probability of abnormality. o represents the output layer weight, b o represents the output layer bias, σ represents the Sigmoid activation function, If Y>θ, it is judged as abnormal, and θ is the threshold; Based on the model’s output probabilities, identify abnormal patterns, including: If F on / off >F threshold , then high frequency start and stop, If T p >T threshold , the water pump is overheated. like The current is abnormal; When an anomaly is detected, protection strategies are executed, including: The central control screen issues an alarm message, indicating the specific fault type and recommended treatment measures. Adjust operating parameters and reduce compressor frequency f c and pump flow q p , Upload the operation data before and after the abnormality occurs to the cloud; After the fault is resolved, the ANN model is retrained using the newly added data.

Images