Control method of electronic pump, liquid control equipment and computer readable storage medium
By acquiring the operating data of the electronic pump and updating the parameters of the self-learning model, the problem of flow characteristic drift caused by wear and aging of the electronic pump was solved, and continuous improvement in accuracy was achieved.
Patent Information
- Application Number
- CN202511445774.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-10
- Publication Date
- 2026-01-13
AI Technical Summary
During use, electric pumps experience flow characteristic drift due to factors such as wear and aging, making accuracy issues difficult to resolve.
By acquiring operating data of the electronic pump, including liquid flow rate, timestamps, and control parameters, the control parameters are calibrated, and a self-learning model is used to update the parameters, constructing a flow-control parameter characteristic curve, and adjusting the control parameters in real time to improve accuracy.
It enables continuous calibration of the electronic pump during use, improves control accuracy, ensures stable flow characteristics, and adapts to individual differences and environmental changes.
Smart Images
Figure CN121322360A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of electronic pump processing technology, and in particular to electronic pump control methods, liquid control devices, and computer-readable storage media. Background Technology
[0002] Electric pumps are typically used for flow control. This involves providing them with specific control parameters to cause them to output a corresponding flow rate. However, electric pumps often suffer from accuracy issues.
[0003] During the use of an electric pump, the flow characteristics may drift due to factors such as wear and aging. Summary of the Invention
[0004] The electronic pump control method, liquid control device, and computer-readable storage medium provided in this application can improve the control accuracy of the electronic pump.
[0005] In a first aspect, this application provides a control method for an electronic pump, the control method comprising: acquiring operating data of the electronic pump during operation; wherein the operating data includes at least two of liquid flow rate, timestamp, and control parameters; calibrating the control parameters based on the operating data; and updating the parameters of a self-learning model corresponding to the electronic pump based on the operating data; wherein the self-learning model is used to predict the target liquid flow rate of the electronic pump at a future time; and controlling the operation of the electronic pump based on the calibrated control parameters and / or the target liquid flow rate.
[0006] The control parameter calibration based on working data includes: during idle periods, traversing the control of the electronic pump according to the parameter range corresponding to the control parameters to obtain the liquid flow rate corresponding to several control parameters within the parameter range, and constructing the initial flow rate-control parameter characteristic curve; when the electronic pump is in use, collecting the current control parameters and the current liquid flow rate; and using the current control parameters and the current liquid flow rate to reconstruct the local part of the initial flow rate-control parameter characteristic curve.
[0007] Specifically, during idle periods, the electronic pump is traversed according to the parameter range corresponding to the control parameters to obtain the liquid flow rate corresponding to several control parameters within the parameter range, and an initial flow rate-control parameter characteristic curve is constructed. This includes: during idle periods, traversing the parameter range corresponding to the control parameters according to a preset step to obtain several target control parameters; controlling the electronic pump to work according to each target control parameter and recording the liquid flow rate corresponding to each target control parameter; and constructing an initial flow rate-control parameter characteristic curve based on each target control parameter and the corresponding liquid flow rate.
[0008] The local reconstruction of the initial flow-control parameter characteristic curve using the current control parameters and the current liquid flow rate includes: obtaining the desired liquid flow rate corresponding to the current control parameters; calculating the deviation rate based on the current liquid flow rate and the desired liquid flow rate; and performing local reconstruction of the initial flow-control parameter characteristic curve using the current control parameters and the current liquid flow rate when the deviation rate is greater than a first deviation threshold and less than a second deviation threshold.
[0009] The method further includes: when the deviation rate is greater than the second deviation threshold, performing a global reconstruction of the initial flow-control parameter characteristic curve.
[0010] The process involves updating the parameters of the self-learning model corresponding to the electronic pump based on the working data. This includes: obtaining new parameters using the forgetting factor, the current parameters of the self-learning model, and the working data; wherein the forgetting factor gradually increases with the amount of working data; and replacing the current parameters with the new parameters to update the parameters of the self-learning model.
[0011] The method further includes: inputting working data into a self-learning model, using the self-learning model to analyze the attenuation trend of liquid flow rate, and predicting the calibration timing; the calibration timing is used to indicate the time corresponding to the calibration timing, at which the step of calibrating control parameters based on working data is performed.
[0012] The working data also includes liquid temperature. The method further includes: calculating the target temperature point corresponding to the electronic pump during use based on the liquid temperature; wherein the temperature corresponding to the target temperature point is the expected temperature corresponding to the liquid temperature; and calibrating the control parameters corresponding to the target temperature point.
[0013] The method includes: receiving model parameters sent from the cloud; training a network model in the cloud based on the working data of multiple electronic pumps; and updating the self-learning model using the model parameters.
[0014] In a second aspect, this application provides a liquid control device, which includes: an electronic pump; a flow sensor for collecting the liquid flow rate when the electronic pump is operating; a processor connected to the electronic pump; and a memory connected to the processor for storing a computer program; the processor is used to execute the computer program to implement the method provided in the first aspect.
[0015] Thirdly, this application provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, is used to implement the method provided in the first aspect.
[0016] The beneficial effects of the embodiments of this application are as follows: Unlike the prior art, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application acquire operating data of the electronic pump during operation. This operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters. Control parameters are calibrated based on at least two of the liquid flow rate, timestamp, and control parameters to improve the accuracy of the electronic pump control parameters. Furthermore, the self-learning model corresponding to the electronic pump is updated based on at least two of the liquid flow rate, timestamp, and control parameters, allowing the self-learning model to be updated in real time. This improves the accuracy of the self-learning model in predicting the target liquid flow rate of the electronic pump at future moments. Therefore, based on the calibrated control parameters and / or the target liquid flow rate, the electronic pump is controlled, thereby improving the control precision of the electronic pump. In other words, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application can continuously calibrate the control precision of the electronic pump during its use. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating the first embodiment of the electronic pump control method provided in this application; Figure 2 This is a flowchart illustrating the second embodiment of the electronic pump control method provided in this application; Figure 3 This is a flowchart illustrating the third embodiment of the electronic pump control method provided in this application; Figure 4 This is a flowchart illustrating the fourth embodiment of the electronic pump control method provided in this application; Figure 5 This is a flowchart illustrating the fifth embodiment of the electronic pump control method provided in this application; Figure 6 This is a flowchart illustrating the sixth embodiment of the electronic pump control method provided in this application; Figure 7 This is a schematic diagram of the structure of an embodiment of the liquid control device provided in this application; Figure 8 This is a schematic diagram of the structure of an embodiment of the water purifier provided in this application; Figure 9 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. Detailed Implementation
[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, it should be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all structures. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.
[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0020] Electric pumps are typically used for flow control. This involves providing them with specific control parameters to cause them to output a corresponding flow rate. However, electric pumps often suffer from accuracy issues.
[0021] During the use of an electric pump, the flow characteristics may drift due to factors such as wear and aging.
[0022] Based on this, this application proposes to obtain operating data of an electronic pump during operation; wherein the operating data includes at least two of liquid flow rate, timestamp, and control parameters; the control parameters are calibrated based on at least two of the liquid flow rate, timestamp, and control parameters to improve the accuracy of the electronic pump control parameters; and the self-learning model corresponding to the electronic pump is updated based on at least two of the liquid flow rate, timestamp, and control parameters, so that the self-learning model can be updated in real time, improving the accuracy of the self-learning model in predicting the target liquid flow rate of the electronic pump at future times, and thus controlling the operation of the electronic pump based on the calibrated control parameters and / or target liquid flow rate, thereby improving the control precision of the electronic pump. That is, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application can continuously calibrate the control precision of the electronic pump during its use. See any of the following embodiments for specific technical solutions.
[0023] See Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the electronic pump control method provided in this application. The control method includes: Step 11: Obtain the operating data of the electronic pump during operation; the operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters.
[0024] In some embodiments, working data includes liquid flow rate and timestamps.
[0025] In some embodiments, the working data includes timestamps and control parameters.
[0026] In some embodiments, the operating data includes liquid flow rate and control parameters.
[0027] In some embodiments, working data includes liquid flow rate, timestamps, and control parameters.
[0028] Liquid flow rate refers to the amount of liquid passing through an electronic pump per unit time.
[0029] The timestamp indicates the time of sampling liquid flow rate and control parameters.
[0030] The control parameter can be a PWM signal. The operation of the electronic pump is controlled by the PWM signal.
[0031] In some embodiments, the electronic pump may be a pump body such as a water pump used for drawing liquids.
[0032] Step 12: Calibrate the control parameters based on the working data.
[0033] In some embodiments, control parameters typically correspond to desired data. Taking the flow rate of an electronic pump as an example, its control parameters usually control the operation of the electronic pump to control its flow rate. That is, there is a substantial correspondence between control parameters and flow rates. For example, the desired flow rate corresponding to control parameter A is 'a', and the desired flow rate corresponding to control parameter B is 'b'. The liquid flow rate in the working data is the actual liquid flow rate collected by the flow sensor. The optimal control of an electronic pump is for the actual liquid flow rate to equal the desired flow rate. However, due to factors such as the usage time and aging of the electronic pump, the actual liquid flow rate may differ from the desired flow rate. Therefore, it is necessary to calibrate the control parameters based on the working data to make the actual liquid flow rate infinitely close to or even equal to the desired flow rate.
[0034] In some embodiments, the difference between the liquid flow rate in the working data and the desired flow rate corresponding to the control parameters can be calculated, and then the control parameters can be compensated based on the difference to obtain new control parameters. Controlling the electronic pump using these new control parameters allows the actual liquid flow rate to equal the desired flow rate. This can be expressed as follows: The desired flow rate corresponding to control parameter A is a. After calibrating the control parameter, the desired flow rate corresponding to control parameter A' is now a.
[0035] Step 13: Update the parameters of the self-learning model corresponding to the electronic pump based on the working data.
[0036] In some embodiments, the self-learning model is used to predict the target liquid flow rate of the electronic pump at future moments.
[0037] In some embodiments, the self-learning model may employ a recursive update algorithm to update the parameters of the self-learning model using newly acquired working data.
[0038] In some embodiments, recursive least squares or Kalman filtering can be used to update the parameters of the self-learning model in real time.
[0039] Step 14: Control the operation of the electronic pump according to the calibrated control parameters and / or target liquid flow rate.
[0040] In some embodiments, after the self-learning model predicts the target liquid flow rate of the electronic pump at a future time, the control parameters corresponding to the future time can be compensated based on the target liquid flow rate to obtain new control parameters, and then the electronic pump can be controlled to work.
[0041] In this embodiment, operating data of the electronic pump during operation is acquired. This operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters. Control parameters are calibrated based on at least two of the liquid flow rate, timestamp, and control parameters to improve the accuracy of the electronic pump's control parameters. Furthermore, the self-learning model corresponding to the electronic pump is updated based on at least two of the liquid flow rate, timestamp, and control parameters, allowing the self-learning model to update in real time and improving the accuracy of its prediction of the target liquid flow rate of the electronic pump at future moments. Based on the calibrated control parameters and / or the target liquid flow rate, the electronic pump is controlled, thereby improving the control precision of the electronic pump. In other words, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application can continuously calibrate the control precision of the electronic pump during its use.
[0042] See Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the electronic pump control method provided in this application. The control method includes: Step 21: Obtain the operating data of the electronic pump during operation; the operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters.
[0043] Step 22: During idle periods, the electronic pump is traversed according to the parameter range corresponding to the control parameters to obtain the liquid flow rate corresponding to several control parameters within the parameter range, and the initial flow rate-control parameter characteristic curve is constructed.
[0044] In some embodiments, during idle periods, the system iterates through the parameter range corresponding to the control parameters according to a preset step to obtain several target control parameters; the electronic pump is controlled to work according to each target control parameter, and the liquid flow rate corresponding to each target control parameter is recorded; an initial flow rate-control parameter characteristic curve is constructed based on each target control parameter and the corresponding liquid flow rate.
[0045] In some embodiments, idle periods can be obtained based on big data. For example, 2:00-3:00 AM every night can be defined as an idle period.
[0046] In some embodiments, taking the control parameter as a PWM signal as an example, its parameter range essentially refers to the duty cycle range. For example, the duty cycle range can be (0, 100%), (10%, 100%), or (20%, 100%).
[0047] In some embodiments, the liquid flow rate corresponding to each target control parameter can be recorded by controlling the electronic pump to operate for a preset duration using each target control parameter, and then obtaining the average liquid flow rate within the preset duration. In some embodiments, the preset duration can be 10 seconds, 15 seconds, 20 seconds, 25 seconds, or 30 seconds, depending on the actual situation of the electronic pump.
[0048] In some embodiments, step 22 can be performed at the beginning of the use of the electronic pump as an initial calibration. For example, step 22 is performed within 7 days of the first use of the electronic pump.
[0049] In other embodiments, step 22 can be performed at any stage of the use of the electric pump.
[0050] Step 23: When the electronic pump is in use, collect the current control parameters and the current liquid flow rate.
[0051] In some embodiments, such as when a user is using an electronic pump, the electronic pump operates to collect current control parameters and current liquid flow rate.
[0052] Step 24: Reconstruct the local part of the initial flow-control parameter characteristic curve using the current control parameters and the current liquid flow rate.
[0053] In some embodiments, the desired liquid flow rate corresponding to the current control parameter can be obtained; the deviation rate can be calculated based on the current liquid flow rate and the desired liquid flow rate; when the deviation rate is greater than a first deviation threshold and less than a second deviation threshold, the initial flow rate-control parameter characteristic curve can be partially reconstructed using the current control parameter and the current liquid flow rate.
[0054] That is, when the deviation rate is greater than the first deviation threshold but less than the second deviation threshold, it indicates that the electric pump cannot provide a suitable liquid flow rate under the current control parameters. Therefore, a local reconstruction of the initial flow-control parameter characteristic curve based on the current control parameters and the current liquid flow rate is required to correct the initial flow-control parameter characteristic curve. Local reconstruction refers to calibrating the local feature points corresponding to the current control parameters in the initial flow-control parameter characteristic curve.
[0055] In some embodiments, when the deviation rate is greater than a second deviation threshold, the initial flow-control parameter characteristic curve is globally reconstructed.
[0056] When the deviation rate exceeds the second deviation threshold, large-scale anomalies may occur, requiring a global reconstruction of the initial flow-control parameter characteristic curve. For example, under the current state of the electronic pump, the pump is traversed according to the parameter range corresponding to the control parameters to obtain the liquid flow rate corresponding to several control parameters within the parameter range, thus constructing a new initial flow-control parameter characteristic curve.
[0057] Step 25: Update the parameters of the self-learning model corresponding to the electronic pump based on the working data.
[0058] The self-learning model is used to predict the target liquid flow rate of the electronic pump at future moments.
[0059] Step 26: Control the operation of the electronic pump according to the calibrated control parameters and / or target liquid flow rate.
[0060] In some embodiments, steps 25 and 26 can be referred to in other embodiments of this application, and will not be described in detail here.
[0061] In this embodiment, operating data of the electronic pump during operation is acquired. This operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters. During idle periods, a global initial flow rate-control parameter characteristic curve is constructed for the electronic pump. Then, during pump operation, local reconstructions of the initial flow rate-control parameter characteristic curve are performed in real time to improve the accuracy of the pump's control parameters. Furthermore, the self-learning model corresponding to the electronic pump is updated based on at least two of the liquid flow rate, timestamp, and control parameters. This allows the self-learning model to be updated in real time, improving the accuracy of its prediction of the target liquid flow rate of the electronic pump in the future. Based on the calibrated control parameters and / or the target liquid flow rate, the electronic pump is controlled, thereby improving the control precision of the electronic pump. In other words, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application can continuously calibrate the control precision of the electronic pump during its operation.
[0062] See Figure 3 , Figure 3This is a flowchart illustrating the first embodiment of the electronic pump control method provided in this application. The control method includes: Step 31: Obtain the operating data of the electronic pump during operation; the operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters.
[0063] Step 32: Calibrate the control parameters based on the working data.
[0064] In some embodiments, steps 31 to 32 can be referred to any embodiment of this application, and will not be described in detail here.
[0065] Step 33: Use the forgetting factor, the current parameters of the self-learning model, and the working data to obtain new parameters; whereby the forgetting factor gradually increases with the amount of working data.
[0066] In some embodiments, the new model parameter = α × old model parameter + (1-α) × new measurement value.
[0067] Here, α is the forgetting factor, initially α=0.7, gradually increasing to 0.95 as the amount of data increases. The new measurement values represent the working data, and the old model parameters represent the current parameters of the self-learning model.
[0068] Step 34: Replace the current parameters with the new parameters to update the parameters of the self-learning model.
[0069] The self-learning model is used to predict the target liquid flow rate of the electronic pump at future moments.
[0070] Step 35: Control the operation of the electronic pump according to the calibrated control parameters and / or target liquid flow rate.
[0071] In some embodiments, steps 34 and 35 can be referred to in other embodiments of this application, and will not be described in detail here.
[0072] In this embodiment, operating data of the electronic pump during operation is acquired. This operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters. Control parameters are calibrated based on at least two of the liquid flow rate, timestamp, and control parameters to improve the accuracy of the electronic pump's control parameters. Furthermore, the self-learning model corresponding to the electronic pump is updated based on at least two of the liquid flow rate, timestamp, and control parameters, allowing the self-learning model to update in real time and improving the accuracy of its prediction of the target liquid flow rate of the electronic pump at future moments. Based on the calibrated control parameters and / or the target liquid flow rate, the electronic pump is controlled, thereby improving the control precision of the electronic pump. In other words, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application can continuously calibrate the control precision of the electronic pump during its use.
[0073] In some embodiments, see Figure 4 The method also includes: Step 41: Input the working data into the self-learning model, use the self-learning model to analyze the attenuation trend of liquid flow rate, and predict the timing of calibration.
[0074] The calibration timing is used to indicate the moment corresponding to the calibration timing.
[0075] In some embodiments, the self-learning model can also perform liquid flow rate attenuation trend analysis, and then predict the calibration timing based on the attenuation trend. For example, the calibration timing is preset to next Monday.
[0076] Step 42: When the calibration time is reached, calibrate the control parameters based on the working data.
[0077] When the calibration timing is reached, the current control parameters are compensated based on the working data to obtain new control parameters. This completes the calibration of the control parameters.
[0078] In this embodiment, a self-learning model is used to analyze the attenuation trend of liquid flow and predict the calibration timing. When the calibration timing is reached, the control parameters are calibrated based on the working data, thereby improving the pertinence of the calibration of the electronic pump control parameters and optimizing the calibration method.
[0079] In some embodiments, the working data also includes liquid temperature, see [reference]. Figure 5 The method also includes: Step 51: Based on the liquid temperature, calculate the target temperature point corresponding to the electric pump during use; where the temperature corresponding to the target temperature point is the desired temperature corresponding to the liquid temperature.
[0080] For example, an electric pump can be a heating pump (such as a hot water pump), which can heat liquids. Therefore, the target temperature points for the electric pump during use can be statistically determined, such as 45 degrees Celsius, 50 degrees Celsius, and 85 degrees Celsius.
[0081] Step 52: Calibrate the control parameters corresponding to the target temperature point.
[0082] Since these target temperatures are commonly encountered during the operation of the electric pump, it is necessary to improve the flow control accuracy at these temperatures. Therefore, the control parameters corresponding to these target temperatures are calibrated.
[0083] In this embodiment, the target temperature points corresponding to the use of the electronic pump are statistically analyzed, and the control parameters corresponding to the target temperature points are calibrated. This achieves targeted calibration of the electronic pump control parameters. Moreover, these temperature points are commonly used temperature points of the electronic pump, so calibration can be prioritized for commonly used temperature points, thereby improving the user experience of the electronic pump.
[0084] In some embodiments, see Figure 6 The method also includes: Step 61: Receive the model parameters sent from the cloud; the network model in the cloud is trained based on the working data of multiple electronic pumps.
[0085] In some embodiments, cloud-based network models can be trained using big data in the cloud, thereby obtaining better model parameters from the cloud.
[0086] In some embodiments, the self-learning model corresponding to this electronic pump also sends its corresponding model parameters to the cloud. The parameters are then compared in the cloud. If the cloud's model parameters are superior to the electronic pump's model parameters, the cloud sends its model parameters to the device corresponding to the electronic pump. If the cloud's model parameters are weaker than the electronic pump's model parameters, no action is taken. The cloud continues to train the model using big data to obtain even better model parameters.
[0087] Step 62: Update the self-learning model using the model parameters.
[0088] In this embodiment, the parameters of the local self-learning model are updated collaboratively in the cloud so that the local self-learning model can more closely match the flow characteristics of the electronic pump, thereby improving the subsequent prediction accuracy and calibration accuracy.
[0089] In one application scenario, the device corresponding to the electronic pump may include the following structure: flow detection unit, pump control unit, and intelligent control center.
[0090] Flow detection unit: A high-precision flow sensor can be installed at the rear end of the RO filter element, with a measurement range of 0-1000ml / min and an accuracy of ±2%.
[0091] Electronic pump control unit: The electronic pump is controlled by a PWM driver with a PWM frequency of 1kHz and a duty cycle resolution of 0.1%.
[0092] Intelligent Control Center: The main controller integrates a self-learning module and a calibration control module, and runs an automatic calibration algorithm.
[0093] The automatic calibration workflow can be as follows: During the initial calibration phase (within 7 days of first use): Step 1: System initialization, establish the default PWM-flow mapping table.
[0094] Step 2: Perform a full calibration every night between 2:00 and 3:00.
[0095] Step 3: Iterate through the PWM duty cycle (10%-100%) in 10% increments.
[0096] Step 4: Collect data for 30 seconds at each point and record the average flow rate.
[0097] Step 5: Build and store the initial characteristic curves.
[0098] During routine implicit calibration: Triggering condition: When the user uses the electronic pump to output liquid, such as hot water or water.
[0099] Step 1: Record the current PWM value and target flow rate.
[0100] Step 2: Obtain the actual flow rate using a flow sensor.
[0101] Step 3: Calculate the deviation rate: ε = (measured value - target value) / target value.
[0102] Step 4: If |ε|>5%, update the mapping relationship of the working point.
[0103] Step 5: Trigger the local curve reconstruction algorithm.
[0104] The self-learning algorithm is implemented as follows: A recursive update algorithm can be used, as follows: New model parameters = α × old model parameters + (1-α) × new measurement value. Where α is the forgetting factor, initially α=0.7, gradually increasing to 0.95 as the amount of data increases.
[0105] And fitting the characteristic curve: for example, using cubic spline interpolation to ensure curve smoothness. Specifically as follows: .parameter ~ Solve using the least squares method.
[0106] The intelligent optimization strategy is as follows: Predictive calibration: Analyze flow decay trends and predict the timing of the next calibration.
[0107] Working point optimization: Collect statistics on commonly used temperature points and focus on optimizing the accuracy of these areas.
[0108] Anomaly detection: Full calibration is triggered when mutations exceed 20%.
[0109] Furthermore, the self-learning model can be as follows: The self-learning model is a multi-factor compensation model: Flow = Base_Flow × K_temp × K_pressure × K_time × K_viscosity.
[0110] - K_temp: Temperature compensation coefficient.
[0111] - K_pressure: Pressure compensation coefficient.
[0112] - K_time: Time decay coefficient.
[0113] - K_viscosity: Viscosity compensation coefficient.
[0114] Furthermore, self-learning models can be constructed from deep learning networks. The structure of a deep learning network is as follows: Input layer: Input PWM value, water temperature, usage time, and historical data.
[0115] Hidden layers: 2 layers, each with 16 neurons.
[0116] Output layer: Predicted flow values.
[0117] Deep learning network training method: online incremental learning.
[0118] Deep learning networks support cloud-based collaborative learning. Local models are periodically uploaded to the cloud, where data from multiple devices is aggregated to train a general model, and optimized parameters are pushed back to the local devices.
[0119] The corresponding device for the electric pump also features a fault diagnosis enhancement system. This system can assess the health of the electric pump, such as by establishing an electric pump performance degradation model. This model can then be used to predict the remaining service life and provide early warnings of replacement timing.
[0120] The corresponding devices for electronic pumps also have intelligent fault diagnosis functions, such as identifying abnormal flow patterns, automatically locating the cause of faults, and providing maintenance suggestions.
[0121] In some embodiments, the calibration algorithm may also employ reinforcement learning to optimize the calibration strategy through a reward mechanism, and introduce federated learning to improve model accuracy through multi-device collaboration. Bayesian optimization may also be used to select the optimal calibration timing.
[0122] In some embodiments, the following hardware may be included in the device corresponding to the electronic pump: Add a miniature pressure sensor to achieve joint pressure-flow calibration.
[0123] Using an ultrasonic flow meter improves measurement accuracy.
[0124] An integrated accelerometer is used to detect the vibration characteristics of the water pump.
[0125] Furthermore, the technical solution of this application can be applied to the collaborative calibration of multi-pump systems, to the self-learning of the characteristic curves of variable frequency pumps, and extended to other fluid control fields.
[0126] Furthermore, the data from the electronic pump in this application can be used in scenarios such as predicting equipment failure based on big data analysis, learning user water usage habits and providing personalized services, as well as analyzing water quality change trends and predicting filter life.
[0127] See Figure 7 , Figure 7 This is a schematic diagram of an embodiment of the liquid control device provided in this application. The liquid control device 100 includes: an electronic pump 10; a flow sensor 20 for collecting the liquid flow rate when the electronic pump 10 is operating; a processor 30 connected to the electronic pump 10; and a memory 40 connected to the processor 30 for storing computer programs. The processor 30 executes the computer programs to implement the following methods: Acquire operating data of the electronic pump during operation; wherein the operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters; calibrate the control parameters based on the operating data; and update the parameters of the self-learning model corresponding to the electronic pump based on the operating data; wherein the self-learning model is used to predict the target liquid flow rate of the electronic pump at a future time; and control the operation of the electronic pump based on the calibrated control parameters and / or the target liquid flow rate.
[0128] In some embodiments, the processor 30 is further configured to execute a computer program to implement the following methods: During idle periods, the electronic pump is traversed and controlled according to the parameter range corresponding to the control parameters to obtain the liquid flow rate corresponding to several control parameters within the parameter range, and an initial flow rate-control parameter characteristic curve is constructed. When the electronic pump is in use, the current control parameters and the current liquid flow rate are collected. The local part of the initial flow rate-control parameter characteristic curve is reconstructed using the current control parameters and the current liquid flow rate.
[0129] In some embodiments, the processor 30 is further configured to execute a computer program to implement the following methods: During idle periods, the system iterates through the parameter range corresponding to the control parameters according to a preset step to obtain several target control parameters; the electronic pump is controlled to work according to each target control parameter, and the liquid flow rate corresponding to each target control parameter is recorded; based on each target control parameter and the corresponding liquid flow rate, an initial flow rate-control parameter characteristic curve is constructed.
[0130] In some embodiments, the processor 30 is further configured to execute a computer program to implement the following methods: Obtain the desired liquid flow rate corresponding to the current control parameters; calculate the deviation rate based on the current liquid flow rate and the desired liquid flow rate; when the deviation rate is greater than the first deviation threshold and less than the second deviation threshold, perform a local reconstruction of the initial flow rate-control parameter characteristic curve using the current control parameters and the current liquid flow rate.
[0131] In some embodiments, the processor 30 is further configured to execute a computer program to implement the following methods: When the deviation rate is greater than the second deviation threshold, the initial flow-control parameter characteristic curve is globally reconstructed.
[0132] In some embodiments, the processor 30 is further configured to execute a computer program to implement the following methods: New parameters are obtained using the forgetting factor, the current parameters of the self-learning model, and the working data; the forgetting factor gradually increases with the amount of working data; the new parameters are used to replace the current parameters to update the parameters of the self-learning model.
[0133] In some embodiments, the processor 30 is further configured to execute a computer program to implement the following methods: The working data is input into the self-learning model, which is then used to analyze the decline trend of the liquid flow rate and predict the calibration timing. The calibration timing indicates the moment at which the control parameters are calibrated based on the working data.
[0134] In some embodiments, the working data also includes liquid temperature, and the processor 30 is further configured to execute a computer program to implement the following methods: Based on the liquid temperature, the target temperature points corresponding to the electric pump during operation are statistically analyzed; the temperature corresponding to the target temperature point is the desired temperature corresponding to the liquid temperature; and the control parameters corresponding to the target temperature points are calibrated.
[0135] In some embodiments, the processor 30 is further configured to execute a computer program to implement the following methods: Receive model parameters sent from the cloud; the network model in the cloud is trained based on the working data of multiple electronic pumps; update the self-learning model using the model parameters.
[0136] In some embodiments, the processor 30 is also configured to execute a computer program to implement the method of any embodiment of this application.
[0137] In some embodiments, the liquid control device may be a water purifier, water heater, or other device capable of controlling the flow rate of liquid. For example, the water purifier may be an under-sink hot water tank purifier.
[0138] See Figure 8 , Figure 8 This is a schematic diagram of an embodiment of the water purifier provided in this application. The water purifier 200 includes: a diaphragm pump 1, a filter element 2, a flow sensor 3, a heating tank 4, a hot water pump 5, a mixing chamber 6, a water outlet 7, and a controller 8. The controller 8 can control the diaphragm pump 1 and the hot water pump 5 according to the scheme provided in this application.
[0139] See Figure 9 , Figure 9 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. The computer-readable storage medium 90 is used to store a computer program 91, which, when executed by a processor, implements the following method: Acquire operating data of the electronic pump during operation; wherein the operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters; calibrate the control parameters based on the operating data; and update the parameters of the self-learning model corresponding to the electronic pump based on the operating data; wherein the self-learning model is used to predict the target liquid flow rate of the electronic pump at a future time; and control the operation of the electronic pump based on the calibrated control parameters and / or the target liquid flow rate.
[0140] In some embodiments, when the computer program 91 is executed by the processor, it is also used to implement the following method: during idle periods, the electronic pump is traversed and controlled according to the parameter range corresponding to the control parameters to obtain the liquid flow rate corresponding to several control parameters within the parameter range, and an initial flow rate-control parameter characteristic curve is constructed; when the electronic pump is in use, the current control parameters and the current liquid flow rate are collected; and the local part of the initial flow rate-control parameter characteristic curve is reconstructed using the current control parameters and the current liquid flow rate.
[0141] In some embodiments, when the computer program 91 is executed by the processor, it is also used to implement the following method: during idle periods, traversing within the parameter range corresponding to the control parameters according to a preset step to obtain several target control parameters; controlling the electronic pump to work according to each target control parameter and recording the liquid flow rate corresponding to each target control parameter; and constructing an initial flow rate-control parameter characteristic curve based on each target control parameter and the corresponding liquid flow rate.
[0142] In some embodiments, when the computer program 91 is executed by the processor, it is further configured to implement the following method: obtain the desired liquid flow rate corresponding to the current control parameters; calculate the deviation rate based on the current liquid flow rate and the desired liquid flow rate; and when the deviation rate is greater than a first deviation threshold and less than a second deviation threshold, perform a local reconstruction of the initial flow-control parameter characteristic curve using the current control parameters and the current liquid flow rate.
[0143] In some embodiments, when the computer program 91 is executed by the processor, it is also used to implement the following method: when the deviation rate is greater than a second deviation threshold, to globally reconstruct the initial flow-control parameter characteristic curve.
[0144] In some embodiments, when the computer program 91 is executed by the processor, it is also used to implement the following method: obtaining new parameters using a forgetting factor, the current parameters of the self-learning model, and working data; wherein the forgetting factor gradually increases with the amount of working data; and updating the parameters of the self-learning model by replacing the current parameters with the new parameters.
[0145] In some embodiments, when the computer program 91 is executed by the processor, it is also used to implement the following method: inputting working data into a self-learning model, using the self-learning model to analyze the attenuation trend of liquid flow rate, and predicting the calibration timing; the calibration timing is used to indicate the step of calibrating control parameters based on working data at the time corresponding to the calibration timing.
[0146] In some embodiments, the working data also includes liquid temperature, and when the computer program 91 is executed by the processor, it is also used to implement the following methods: based on the liquid temperature, statistically analyze the target temperature point corresponding to the electronic pump during use; wherein the temperature corresponding to the target temperature point is the desired temperature corresponding to the liquid temperature; and calibrate the control parameters corresponding to the target temperature point.
[0147] In some embodiments, when the computer program 91 is executed by the processor, it is also used to implement the following methods: receiving model parameters sent from the cloud; training a network model in the cloud based on the working data of multiple electronic pumps; and updating the self-learning model using the model parameters.
[0148] In some embodiments, when executed by a processor, computer program 91 is also used to implement the method of any embodiment of this application.
[0149] In summary, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application acquire operating data of the electronic pump during operation. This operating data includes at least two of the following: liquid flow rate, timestamp, and control parameters. The method calibrates the control parameters based on at least two of the liquid flow rate, timestamp, and control parameters to improve the accuracy of the electronic pump control parameters. Furthermore, it updates the parameters of the self-learning model corresponding to the electronic pump based on at least two of the liquid flow rate, timestamp, and control parameters, enabling the self-learning model to update in real time and improving the accuracy of the self-learning model in predicting the target liquid flow rate of the electronic pump at future moments. Based on the calibrated control parameters and / or the target liquid flow rate, the method controls the operation of the electronic pump, thereby improving the control precision of the electronic pump. This addresses the problem of decreased flow control precision caused by individual differences in electronic pumps, aging and wear, and environmental changes.
[0150] Furthermore, the electronic pump control method, liquid control device, and computer-readable storage medium provided in this application construct a complete "sensing-calibration-learning-optimization" closed-loop system.
[0151] For example, it involves intelligent calibration triggering mechanisms, including implicit calibration, explicit calibration, and event-triggered calibration.
[0152] Implicit calibration: Calibration is performed during normal user operation, taking advantage of the opportunity when cold or hot water is dispensed separately.
[0153] Explicit calibration: Automatically perform the full calibration procedure during system idle periods (such as at night).
[0154] Event-triggered calibration: Calibration is initiated immediately upon detection of abnormal deviation.
[0155] For example, this involves multi-dimensional traffic characteristic modeling, including: Establish a basic mapping relationship between PWM duty cycle and flow rate; Introduce a temperature compensation factor (flow characteristics at different water temperatures). Consider the time decay factor (the impact of usage time on flow rate). Construct a three-dimensional flow characteristic surface: Flow = f(PWM, Temperature, Time).
[0156] For example, it involves the core of self-learning algorithms, including data acquisition, feature extraction, model updating, and prediction optimization.
[0157] Data acquisition: Record the PWM value, measured flow rate, water temperature, and timestamp each time water is discharged.
[0158] Feature extraction: Identifying the changing trends and patterns of traffic characteristics.
[0159] Model update: The model parameters are updated in real time using recursive least squares or Kalman filtering.
[0160] Predictive optimization: Based on historical data, predict future changes in traffic characteristics and compensate in advance.
[0161] For example, it involves intelligent data management, including hierarchical storage, data compression, and anomaly filtering.
[0162] Tiered storage: Frequently used work point data is stored in fast memory.
[0163] Data compression: Curve fitting is used to compress the amount of stored data.
[0164] Anomaly filtering: Identify and remove outlier data points to ensure model accuracy.
[0165] The electronic pump control method, liquid control device, and computer-readable storage medium provided in this application have the following effects: Automatic calibration effect: The system completes calibration automatically without human intervention; The calibration frequency is adaptive; frequent calibration is required in the initial stage of use, and the frequency is reduced after stabilization. Experimental data: Flow control accuracy improved by 85% after automatic calibration.
[0166] The more you use it, the more accurate it becomes: The longer it is used, the more calibration data is accumulated, and the more accurate the model becomes. Experimental results show that accuracy improved by 40% after one month of use, by 65% after three months, and reached its optimal level and remained stable after six months. Long-term accuracy maintenance: It can still maintain ±3% flow control accuracy after 2 years of use.
[0167] Significant cost-effectiveness: A high-precision water pump control effect is achieved by using a standard water pump with an automatic calibration scheme. No additional hardware is required; performance improvements are achieved solely through software algorithms.
[0168] Highly adaptable: Automatically adapts to pump aging: Real-time compensation for performance degradation; Environmental adaptability: Automatically compensates for the effects of changes in water temperature and pressure; Individualized adaptation: An independent flow characteristic model is established for each device.
[0169] Quantitative indicators: Temperature control accuracy: improved from ±5°C to ±1.5°C; Response time: The time for the first cup of water to reach the target temperature is less than 2 seconds; Calibration convergence speed: The new equipment completes the initial calibration within 7 days, with an accuracy of over 90%.
[0170] In the several embodiments provided in this application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.
[0171] If the integrated units in the other embodiments described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processing circuit component (processor) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0172] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.
Claims
1. A control method of an electronic pump, characterized by, The control method comprises: acquiring working data of the electronic pump in operation; wherein the working data comprises at least two of liquid flow, time stamp and control parameter; calibrating the control parameter according to the working data; and updating the self-learning model corresponding to the electronic pump with parameters according to the working data; wherein the self-learning model is used to predict the target liquid flow of the electronic pump at a future time; controlling the electronic pump to operate according to the calibrated control parameter and / or the target liquid flow.
2. The control method according to claim 1, characterized by, The calibration of the control parameter according to the working data comprises: in an idle period, performing traversal control on the electronic pump according to the parameter range corresponding to the control parameter, to obtain the liquid flow corresponding to a plurality of control parameters in the parameter range, and to construct an initial flow-control parameter characteristic curve; in use of the electronic pump, collecting the current control parameter and the current liquid flow; reconstructing a part of the initial flow-control parameter characteristic curve by using the current control parameter and the current liquid flow.
3. The control method according to claim 2, characterized by, The traversal control on the electronic pump according to the parameter range corresponding to the control parameter in the idle period to obtain the liquid flow corresponding to a plurality of control parameters in the parameter range and to construct an initial flow-control parameter characteristic curve comprises: in the idle period, performing traversal according to a preset step in the parameter range corresponding to the control parameter to obtain a plurality of target control parameters; controlling the electronic pump according to each target control parameter and recording the liquid flow corresponding to each target control parameter; constructing an initial flow-control parameter characteristic curve according to each target control parameter and the corresponding liquid flow.
4. The control method according to claim 2, characterized by, The reconstruction of a part of the initial flow-control parameter characteristic curve by using the current control parameter and the current liquid flow comprises: obtaining the expected liquid flow corresponding to the current control parameter; calculating a deviation rate according to the current liquid flow and the expected liquid flow; when the deviation rate is greater than a first deviation threshold and less than a second deviation threshold, reconstructing a part of the initial flow-control parameter characteristic curve by using the current control parameter and the current liquid flow.
5. The control method according to claim 4, characterized by The method further comprises: when the deviation rate is greater than the second deviation threshold, globally reconstructing the initial flow-control parameter characteristic curve.
6. The control method according to claim 1, characterized by, The updating of the self-learning model corresponding to the electronic pump with parameters according to the working data comprises: obtaining new parameters by using a forgetting factor, the current parameters of the self-learning model and the working data; wherein the forgetting factor gradually increases with the data volume of the working data; updating the self-learning model with parameters by replacing the current parameters with the new parameters.
7. The control method according to claim 1, characterized by, The method further comprises: inputting the working data into the self-learning model, analyzing the attenuation trend of the liquid flow by using the self-learning model, and predicting a calibration time; the calibration time is used to indicate that the calibration of the control parameter according to the working data is performed at the time corresponding to the calibration time.
8. The control method according to claim 1, characterized by, The working data further comprises liquid temperature, and the method further comprises: Based on the liquid temperature, the target temperature points corresponding to the electric pump during use are calculated; wherein, the temperature corresponding to the target temperature point is the desired temperature corresponding to the liquid temperature; The control parameters corresponding to the target temperature point are calibrated.
9. The control method according to any one of claims 1 to 8, characterized by, The method includes: Receive model parameters sent from the cloud; the network model in the cloud is trained based on the working data of multiple electronic pumps; The self-learning model is updated using the model parameters.
10. A liquid control device, characterized by, The liquid control device includes: Electric pump; A flow sensor is used to collect the liquid flow rate when the electronic pump is working; The processor is connected to the electronic pump; A memory, connected to the processor, for storing computer programs; The processor is used to execute the computer program to implement the method as described in any one of claims 1-9.
11. A computer readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program, which, when executed by a processor, is used to implement the method as described in any one of claims 1-9.