Double-inlet and double-outlet valve temperature control method based on machine learning

Abstract

The invention discloses a double-inlet and double-outlet water valve temperature control method based on machine learning, which relates to the technical field of water temperature control and comprises the following steps: identifying a current working stage and entering a corresponding control step; when it is identified that the water source temperature is not synchronized, the cold and hot water initial flow proportion is adjusted through hot water inlet temperature prediction of a water source temperature prediction model, and the initial position of a double-water-inlet valve is adjusted according to the target flow; when the temperature regulation and control stage is identified, dynamically optimizing the cold and hot water mixing proportion based on the real-time inlet water temperature of the double water inlets, generating a valve regulation instruction according to the target flow, and performing position regulation on the double water inlet valves in an incremental position regulation mode; and when the temperature is identified as the stable temperature stage, the real-time outlet water temperature change rate of the double water outlets is monitored, and when the change rate exceeds a dynamic response threshold value, a feedforward-feedback composite control strategy is adopted to adjust the valve positions of the double water inlets under fluctuation suppression. According to the invention, the fluctuation range of the outlet water temperature is greatly reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water temperature control, and particularly relates to a double-inlet double-outlet water valve temperature control method based on machine learning. BACKGROUND

[0002] In the double-inlet double-outlet water valve temperature control system in the fields of bathroom, heating and ventilation, the traditional method generally adopts PID control or logic control based on fixed rules. Such methods have significant defects in practical application: first, in the cold start stage, due to the lack of prediction ability for water source temperature fluctuation, the controller needs to passively wait until the temperature stabilizes before starting to adjust, resulting in a long initial response delay, and users have to endure the water temperature fluctuation; second, when the flow suddenly changes or the water inlet temperature is disturbed, the large step adjustment of the valve position easily causes overshoot, causing the outlet water temperature to repeatedly oscillate, not only reducing comfort, but also accelerating the mechanical wear of the valve; finally, after entering the stable state, the traditional method is difficult to effectively cope with water pressure sudden changes, seasonal water source temperature changes and other disturbance factors, due to the lack of feedforward compensation mechanism, only relying on lagging feedback adjustment, which prolongs the temperature fluctuation recovery time and increases energy consumption. Although the prior art attempts to introduce adaptive PID parameters, it still cannot systematically solve the problem of cooperative control of water source prediction, incremental approximation and disturbance feedforward compensation, which restricts the improvement of user experience and energy efficiency. SUMMARY

[0003] To solve the problem of cooperative control of water source prediction, incremental approximation and disturbance feedforward compensation, the present application proposes a double-inlet double-outlet water valve temperature control method based on machine learning, which adjusts the water inlet flow of the corresponding water inlet through independent valve control, comprising: S1: receiving the target outlet water temperature and target flow set by the user, and collecting the hot water inlet temperature and flow, cold water inlet temperature and flow, double-outlet outlet water temperature and flow in real time; S2: based on the collected data, identifying the current working stage and entering the corresponding control step, the working stage including the water source temperature unsynchronized stage, the temperature regulation stage and the temperature stable stage; S3: when identifying the water source temperature unsynchronized stage, adjusting the initial flow ratio of hot and cold water through the hot water inlet temperature prediction of the water source temperature prediction model, and adjusting the initial position of the double-inlet valve according to the target flow; S4: when identifying the temperature regulation stage, dynamically optimizing the hot and cold water mixing ratio based on the real-time water inlet temperature of the double-inlet, and generating valve adjustment instructions according to the target flow to adjust the position of the double-inlet valve in an incremental position adjustment mode; S5: when identifying the temperature stable stage, monitoring the real-time outlet water temperature change rate of the double-outlet, and when the change rate exceeds the dynamic response threshold, using a feedforward-feedback composite control strategy to adjust the position of the double-inlet valve under fluctuation suppression.

[0004] The application realizes full-process adaptive management in a double-inlet double-outlet water valve system for the first time through deep coupling of a machine learning model and a dynamic control strategy. The core breakthrough lies in decoupling the temperature control process into three stages of unsynchronized water source temperature, temperature regulation, and temperature stabilization, and dynamically switching control mechanisms according to the characteristics of different stages: in the initial water source fluctuation stage, a time series prediction model is used to predict the stable temperature of hot water and accurately calculate the initial flow ratio, directly overcoming the repeated oscillation problem caused by the hysteresis of water source temperature in traditional methods; in the temperature regulation stage, an incremental position adjustment algorithm is used to realize accurate tracking of flow; in the stable stage, an anti-noise feedforward-feedback strategy is introduced to lock the steady-state operating point, so that the fluctuation range of outlet water temperature is greatly reduced.

[0005] Further, in the S2 step, the determination method for identifying the working stage is: comparing the duration that the outlet water temperature is within the preset range of the target outlet water temperature with the duration threshold ; at the same time, the temperature fluctuation standard deviation of the hot water inlet temperature and the cold water inlet temperature is calculated respectively , and the preset hot water fluctuation standard deviation and the preset cold water temperature fluctuation standard deviation are compared; When or , it is determined that the water source temperature is in the unsynchronized stage; When and , but , it is determined that the temperature is in the regulation stage; When and , it is determined that the temperature is in the stable stage.

[0006] Further, in the S3 step, the water source temperature prediction model is a time series prediction model based on historical temperature data to predict the stable temperature value, and the initial flow ratio of cold and hot water is calculated by the following formula: In the formula, is the initial flow ratio of cold and hot water, is the target outlet water temperature, is the cold water inlet temperature, is the stable temperature value predicted by the water source temperature prediction model based on the historical hot water inlet temperature.

[0007] Further, in the S4 step, the incremental position adjustment specifically includes: calculating the deviation of the current inlet flow from the required flow corresponding to the valve adjustment instruction , the valve opening adjustment amount is calculated by the proportional coefficient and the deviation limiting function ; According to the deviation , the current valve position is updated to the new valve position , and the adjustment direction is determined by the positive and negative of the deviation.

[0008] Further, the deviation limiting function is expressed as the following formula: In the formula, is the proportional coefficient, is the deviation limiting function, which limits the deviation in the interval , is the single maximum allowable adjustment amount, and when is less than the dead zone threshold , the valve adjustment is stopped.

[0009] Further, the proportional coefficient is set according to the valve nonlinear characteristic segmentation: When the valve is in the first opening interval, ; When the valve is outside the first opening interval, , is the nonlinear coefficient.

[0010] Further, in the S5 step, the feed-forward feedback compound control strategy includes a feed-forward disturbance compensation term that generates a pre-adjustment instruction for water pressure fluctuations and / or sudden changes in water temperature, and is expressed as the following formula: In the formula, is the pre-adjustment instruction, is the steady state working point, is the temperature deviation between the real-time changing outlet water temperature and the target outlet water temperature, are the proportional term coefficient, integral term coefficient, and differential term coefficient, respectively, is the feed-forward compensation coefficient, is the normalized disturbance amplitude based on water pressure fluctuations and / or water temperature.

[0011] Further, the steady state working point is obtained by the following steps: Record the valve position that satisfies in the temperature stabilization stage; Take the average of the valve position in the last ten control cycles as the steady state working point.

[0012] Further, the differential item calculation adopts an anti-noise algorithm: In the formula, is a control period.

[0013] Further, it also includes an abnormality processing step: When any outlet temperature exceeds a first warning temperature, increase the cold water inlet flow rate proportion; When any outlet temperature exceeds a second warning temperature, close the hot water inlet valve; When any inlet flow is interrupted, close the double inlet valve.

[0014] Compared with the prior art, the present application has at least the following beneficial effects: (1) The double-inlet double-outlet valve temperature control method based on machine learning proposed by the present application realizes full-process adaptive management in a double-inlet double-outlet valve system for the first time through deep coupling of a machine learning model and a dynamic control strategy. The core breakthrough lies in decoupling the temperature control process into three stages of unsynchronized water source temperature, temperature regulation, and temperature stabilization, and dynamically switching control mechanisms according to the characteristics of different stages: in the initial water source fluctuation stage, a time series prediction model is used to predict the stable temperature of hot water and accurately calculate the initial flow rate proportion, directly overcoming the repeated oscillation problem caused by the hysteresis of the water source temperature in traditional methods; in the temperature regulation stage, an incremental position adjustment algorithm is used to realize accurate tracking of the flow rate; in the stable stage, an anti-noise feedforward-feedback strategy is introduced to lock the steady-state operating point, so that the outlet temperature fluctuation amplitude is greatly reduced;

[0015] (2) Through a working stage determination criterion based on the double thresholds of temperature fluctuation standard deviation and compliance duration, the double-valve system is provided with a state self-sensing capability, changing the limitation of the traditional method in which a single control strategy deals with complex working conditions throughout the process; (3) The water source temperature prediction model eliminates source interference through historical data modeling, the nonlinear valve compensation in the incremental adjustment mechanism improves device compatibility, and the anti-noise differential algorithm in the feedforward-feedback strategy and the disturbance amplitude normalization processing cooperatively suppress the influence of environmental mutations. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a method step diagram of a double-inlet double-outlet valve temperature control method based on machine learning. DETAILED DESCRIPTION

[0017] The following is a specific embodiment of the present application and further describes the technical solutions of the present application in conjunction with the drawings, but the present application is not limited to these embodiments.

[0018] To clearly illustrate the implementation logic of the technical scheme of the present application, a double-inlet double-outlet valve temperature control system is taken as an embodiment for description. The system constructs a fluid circuit through a hot water inlet channel, a cold water inlet channel and two independently regulated outlet channels, and runs the adaptive temperature control method proposed by the present application by a microcontroller. The following implementation process will focus on the core implementation process of the three-stage control framework and the collaborative control algorithm cluster. In the implementation, the system acquires the sampling data of the hot water inlet temperature sensor, the cold water inlet temperature sensor and the double-outlet temperature sensor in real time, and simultaneously collects real-time flow signals through a flow detection device, and the valve actuator adopts a stepper motor to realize accurate opening degree control. It should be noted that the skilled person in the art can apply the technical principle disclosed in the present embodiment to other scenes requiring double-valve collaborative temperature control, such as industrial heat exchange systems and medical constant temperature equipment, and the innovative architecture and algorithm logic have universality. As shown in Figure 1 , the present application proposes a double-inlet double-outlet valve temperature control method based on machine learning, mainly including the following steps:

[0019] S1: receiving the target outlet water temperature and target flow rate set by the user, and collecting the hot water inlet temperature and flow rate, cold water inlet temperature and flow rate, and double-outlet outlet water temperature and flow rate in real time; S2: based on the collected data, identifying the current working stage and entering the corresponding control step, the working stage including a water source temperature unsynchronized stage, a temperature regulation stage and a temperature stable stage; S3: when identifying the water source temperature unsynchronized stage, adjusting the initial flow rate ratio of cold and hot water through hot water inlet temperature prediction of the water source temperature prediction model, and adjusting the initial position of the double-inlet valve according to the target flow rate; S4: when identifying the temperature regulation stage, dynamically optimizing the cold and hot water mixing ratio based on the real-time inlet water temperature of the double-inlet, and generating valve adjustment instructions according to the target flow rate to adjust the position of the double-inlet valve in an incremental position adjustment mode; S5: when identifying the temperature stable stage, monitoring the real-time outlet water temperature change rate of the double-outlet, and when the change rate exceeds the dynamic response threshold, adjusting the valve position of the double-inlet in a fluctuation suppression mode using a feedforward-feedback composite control strategy.

[0020] Specifically, the user sets the target outlet water temperature and the total target flow rate through the intelligent terminal, and specifies the flow distribution ratio parameters of the double-outlet (for example, outlet A accounts for 60% and outlet B accounts for 40%). After system initialization, the real-time data acquisition process is started: the temperature sensor group synchronously collects the hot water inlet temperature and instantaneous flow rate , the cold water inlet temperature and instantaneous flow rate , and independently monitor the real-time temperature of the dual water outlets A and B And the corresponding flow At the same time, the valve position sensor continuously feeds back the current opening position of the four electric control valves , corresponding to the flow control status of the hot water inlet, cold water inlet, outlet A, and outlet B. All collected data are transmitted to the central processing unit via a high-speed bus for dynamic analysis.

[0021] Based on the above real-time data, the system executes the phase identification logic: First, the standard deviation of the fluctuation of the hot water and cold water inlet temperatures is calculated. and , and compare them with the preset hot water fluctuation standard deviation and preset cold water temperature fluctuation standard deviation Compare and calculate the temperature of the two water outlets to maintain within the preset range of the target water outlet temperature. (Only taking 2℃ as an example) , and the duration threshold Compare: When detected or When , the system determines that it has entered the water source temperature unsynchronized stage; When detected and ,but When , the system determines that it has entered the temperature control stage; When detected and ,and When , the system determines that it has entered the temperature stabilization stage; Thus, the subsequent corresponding control strategy execution process is triggered based on the stage identification results.

[0022] When the system determines that the water source temperature has not been synchronized, the system starts the water source temperature prediction model to dynamically calibrate the inlet water temperature. The model uses the exponential smoothing algorithm to predict the hot water inlet temperature under steady state based on the cached historical temperature time series data (including at least the sampling sequence of the last 10 seconds). The specific implementation process is as follows: first, the hot water inlet temperature sequence is subjected to noise filtering to eliminate the abnormal values ​​of instantaneous measurement, and then the predicted value is calculated according to the following formula :

[0023] , , in, The trend slope of the hot water inlet temperature change in the last 3 seconds Flow fluctuation variance of hot water inlet , the historical weight coefficient dynamically adjusted by the autoregressive model, are the bias constant, temperature trend coefficient, and flow fluctuation coefficient obtained by pre-training, respectively, is the initial temperature reference value (taking the initial value of the current continuous 10-second temperature sequence, reflecting the thermal inertia of the system), is the conversion coefficient representing the cumulative effect of flow on temperature, determined by the volume and heat capacity characteristics of the pipeline, is the hot water inlet flow at time , and the integral term represents the cumulative effect of flow on temperature change, is the sampling time.

[0024] Based on the predicted temperature and the real-time cold water inlet temperature , the initial flow ratio instruction is dynamically generated using the following formula: , The mixing ratio determines the basic mixing ratio of cold and hot water. For example, when the user sets , the measured , and the predicted , calculate , that is, the cold and hot water need to be mixed in a ratio of 1:1.

[0025] The system then combines the total target flow and the dual-outlet distribution ratio to calculate the required flow of each inlet: Hot water inlet required flow , Cold water inlet required flow , Then adjust the flow of each branch outlet according to the dual-outlet distribution ratio (such as the previously set A:B=6:4).

[0026] Finally, perform valve position initialization control: based on the flow-valve opening calibration curve, convert the calculated target flow into initial opening instructions for the four valves. In this process, the system uses a flow feedforward compensation mechanism to eliminate the effects of pipeline pressure fluctuations, for example, when a 5% decrease in inlet pressure is detected, increase the opening instruction by 3%-5% according to the preset gain coefficient. After completion of the initialization, continuously monitor the stability of the water source temperature to create a prerequisite for switching to temperature regulation.

[0027] When the system determines that it has entered the temperature regulation stage, the system calculates the theoretical mixed water temperature in real time based on the principle of heat conservation: , where, As the theoretical benchmark value of the water mixing effect at the outlet, it is consistent with the actual outlet temperature. Perform dynamic comparison. When it is detected that the outlet water temperature of any outlet deviates from the theoretical reference value and exceeds the tolerance threshold, the ratio optimization program is triggered to update the flow mixing ratio. : , This optimization ratio This will replace the initialization phase ratio to ensure real-time response to water temperature changes. For example, if the cold water temperature suddenly drops by 2°C, the system will automatically increase the cold water flow ratio by 5%-8% to offset the temperature fluctuation.

[0028] Based on the updated mix ratio and target traffic , redistribute the required flow of each valve of the dual water inlet, and adjust the flow of each branch outlet again according to the dual outlet distribution ratio set by the user.

[0029] Then perform incremental valve position adjustment: the deviation between the actual flow and the required flow of each inlet valve is calculated. , calculate the valve opening adjustment amount through the deviation limit function : , in, is the proportionality coefficient, is the deviation limiting function, which means that the deviation Limited to the range middle, The maximum allowable adjustment for a single time. Less than the dead zone threshold Stop valve adjustment.

[0030] At the same time, the proportionality coefficient Dynamic switching according to the nonlinear characteristics of the valve (when the valve is in the first opening range, such as 30% to 70%, , when the valve is outside the first opening range, , is a nonlinear coefficient). The adjustment process is superimposed with a hysteresis compensation algorithm, which increases the pressure in advance when pipeline pressure fluctuations are detected. The value is 15% to avoid overshoot caused by water inertia. , the current valve position Updated to new valve position , where the adjustment direction is determined by the sign of the deviation.

[0031] During this phase, the system continuously monitors the control effect and and ,and If the condition is met, the temperature stabilizing phase is entered, otherwise the iteration optimization continues until convergence.

[0032] In the temperature stabilizing phase, the system monitors the real-time outlet temperature change rate of the double-outlet, and when the outlet temperature change rate of any outlet exceeds the dynamic response threshold, the feedforward-feedback composite control strategy is enabled. First, the steady-state operating point is obtained by recording the valve position that satisfies in the temperature stabilizing phase, and taking the average of the valve position in the last ten control cycles as the steady-state operating point , the pre-adjustment instruction is generated in real time according to the following formula :

[0033] , wherein, is the temperature deviation between the real-time outlet temperature and the target outlet temperature, are the proportional term coefficient, integral term coefficient, and differential term coefficient, respectively, is the feedforward compensation coefficient, is the normalized disturbance amplitude based on water pressure fluctuation and / or inlet water temperature. At the same time, the differential term adopts an anti-noise algorithm for calculation:

[0034] .

[0035] This algorithm decomposes temperature fluctuations into three compensation terms: the PID feedback term ( ) corrects cumulative deviation, and the feedforward term ( ) pre-compensates for sudden disturbances. For example, when a 10% drop in water pressure is detected, the valve opening can be automatically increased through the feedforward term, thereby improving response speed.

[0036] Further, when the temperature of any outlet exceeds the first warning temperature, the cold water valve opening is increased by a preset proportion in an emergency, and if the temperature breaks through the second warning temperature, the hot water valve power is immediately cut off and an alarm is sounded. After the disturbance is eliminated, the system automatically returns to steady-state control to maintain the temperature fluctuation range.

[0037] To sum up, the temperature control method of the double-inlet double-outlet water valve based on machine learning is proposed, which realizes the full-process adaptive management in the double-inlet double-outlet water valve system through the deep coupling of the machine learning model and the dynamic control strategy. The core breakthrough is to decouple the temperature control process into three stages of unsynchronized water source temperature, temperature regulation and temperature stability, and dynamically switch the control mechanism according to the characteristics of different stages: in the initial water source fluctuation stage, the time series prediction model is used to predict the stable temperature of hot water and accurately calculate the initial flow ratio, which directly overcomes the repeated oscillation problem caused by the hysteresis of the water source temperature in the traditional method; in the temperature regulation stage, the incremental position adjustment algorithm is used to realize the accurate tracking of the flow; in the stable stage, the anti-noise feedforward-feedback strategy is introduced to lock the steady-state working point, so that the fluctuation range of the outlet water temperature is greatly reduced.

[0038] Through the working stage determination criterion based on the double thresholds of temperature fluctuation standard deviation and compliance duration, the state self-sensing capability is provided for the double-valve system, and the limitation of the single control strategy in the traditional method for dealing with complex working conditions throughout the process is changed. The water source temperature prediction model eliminates the source interference through historical data modeling, the nonlinear valve compensation in the incremental adjustment mechanism improves the device compatibility, and the anti-noise differential algorithm in the feedforward-feedback strategy and the disturbance amplitude normalization processing cooperate to suppress the influence of environmental mutations.

[0039] It should be noted that all the directionality indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative position relationship, motion condition, etc. between the components in a certain specific posture (as shown in the drawings), and if the specific posture changes, the directionality indications will also change accordingly.

[0040] In addition, the descriptions such as "first", "second", "one" and the like in the present application are only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" can be explicitly or implicitly included at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, such as two, three, etc., unless otherwise specifically limited.

[0041] In the present application, unless otherwise specifically defined and limited, the terms "connection", "fixation" and the like should be understood in a broad sense, for example, "fixation" can be fixed connection, or detachable connection, or integral; can be mechanical connection, or electrical connection; can be direct connection, or indirect connection through intermediate medium; can be the internal communication of two elements or the interaction relationship between two elements, unless otherwise specifically limited. For ordinary skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0042] In addition, the technical solutions among various embodiments of the present application can be combined with each other, but it must be based on that a person skilled in the art can realize, when the combination of technical solutions appears contradictory or cannot be realized, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope required by the present application.

Claims

1. A temperature control method for dual-inlet and dual-outlet water valves based on machine learning, which controls the water flow rate of corresponding water inlets through independent valves, characterized in that: include: S1: Receives the target water outlet temperature and target flow set by the user, and collects the hot water inlet temperature and flow, cold water inlet temperature and flow, and dual water outlet water temperature and flow in real time; S2: Based on the collected data, the current working stage is identified and the corresponding control steps are entered. The working stages include the water source temperature non-synchronization stage, the temperature control stage, and the temperature stabilization stage; S3: When it is identified that the water source temperature is not synchronized, the initial flow ratio of hot and cold water is adjusted according to the hot water inlet temperature prediction of the water source temperature prediction model, and the initial position of the dual water inlet valve is adjusted according to the target flow; S4: When the temperature control stage is identified, the hot and cold water mixing ratio is dynamically optimized based on the real-time water inlet temperature of the dual water inlets, and valve adjustment instructions are generated according to the target flow rate. The position of the dual water inlet valve is adjusted in an incremental position adjustment manner; S5: When the temperature is identified as a stable stage, the real-time outlet water temperature change rate of the dual water outlets is monitored. When the change rate exceeds the dynamic response threshold, a feedforward-feedback composite control strategy is used to adjust the valve position of the dual water inlets under fluctuation suppression.

2. A temperature control method for a dual-inlet and dual-outlet water valve based on machine learning as claimed in claim 1, characterized in that: In step S2, the method for determining the working stage is as follows: Compare the duration that the outlet water temperature is within the preset range of the target outlet water temperature With duration threshold At the same time, calculate the temperature fluctuation standard deviation of hot water inlet temperature and cold water outlet temperature respectively , and the preset hot water fluctuation standard deviation and preset cold water temperature fluctuation standard deviation Make comparisons; when or When , it is determined to be the water source temperature unsynchronized stage; when and ,but When , it is determined to be the temperature control stage; when and ,and When , it is determined to be the temperature stable stage.

3. The temperature control method of a dual-inlet and dual-outlet water valve based on machine learning according to claim 1, characterized in that: In step S3, the water source temperature prediction model is a time series prediction model, which predicts a stable temperature value based on historical temperature data. The initial flow ratio of hot and cold water is calculated using the following formula: Where, is the initial flow ratio of hot and cold water, is the target outlet water temperature, is the cold water inlet temperature, It is the stable temperature value predicted by the water source temperature prediction model based on the historical hot water inlet temperature.

4. The temperature control method of a dual-inlet and dual-outlet water valve based on machine learning according to claim 1, characterized in that: In step S4, the incremental position adjustment specifically includes: According to the deviation between the current water inlet flow and the valve adjustment instruction corresponding to the required flow , calculate the valve opening adjustment amount through the proportional coefficient and deviation limit function ; according to , the current valve position Updated to new valve position , where the adjustment direction is determined by the sign of the deviation.

5. The temperature control method of a dual-inlet and dual-outlet water valve based on machine learning as claimed in claim 4, characterized in that: The deviation limiting function is expressed as the following formula: Where, is the proportionality coefficient, is the deviation limiting function, which means that the deviation Limited to the range middle, The maximum allowable adjustment for a single time. Less than the dead zone threshold Stop valve adjustment.

6. A temperature control method for a dual-inlet and dual-outlet water valve based on machine learning as claimed in claim 5, characterized in that: The proportionality coefficient Segmented setting based on valve nonlinear characteristics: When the valve is in the first opening range, ; When the valve is outside the first opening range, , is the nonlinear coefficient.

7. The method for controlling temperature of a dual-inlet and dual-outlet water valve based on machine learning according to claim 1, characterized in that: In step S5, the feedforward-feedback composite control strategy includes a feedforward disturbance compensation term, which generates a pre-adjustment instruction for water pressure fluctuations and / or sudden changes in inlet water temperature. The formula is: Where, For pre-adjustment instructions, is the steady-state operating point, is the temperature deviation between the outlet water temperature that changes in real time with time t and the target outlet water temperature, are the proportional term coefficient, the integral term coefficient, and the differential term coefficient respectively. is the feedforward compensation coefficient, is the normalized disturbance amplitude based on water pressure fluctuation and / or inlet water temperature.

8. The temperature control method of a dual-inlet and dual-outlet water valve based on machine learning according to claim 7, characterized in that: The steady-state operating point is obtained by the following steps: Record the temperature during the stabilization phase. Valve position ; Take the valve position for ten consecutive control cycles The mean value of is taken as the steady-state operating point.

9. The method for controlling temperature of a dual-inlet and dual-outlet water valve based on machine learning according to claim 7, characterized in that: The differential term calculation adopts the noise-resistant algorithm: Where, For the control cycle.

10. The method for controlling temperature of a dual-inlet and dual-outlet water valve based on machine learning according to claim 1, characterized in that: It also includes exception handling steps: When the temperature of any water outlet exceeds the first warning temperature, the flow rate ratio of the cold water inlet is increased; When the temperature of any water outlet exceeds the second warning temperature, close the hot water inlet valve; When the water flow at any water inlet is interrupted, close the dual water inlet valve.

Images