Load detection system and method suitable for warmer

Abstract

The invention belongs to the technical field of heaters, and particularly relates to a load detection system and method suitable for a heater, and the method comprises the steps: 1, data collection: collecting operation data of the heater in continuous operation in real time through a sensor; 2, data preprocessing: cleaning and sorting the operation data collected according to the time sequence, and removing abnormal values and noise; step 3, feature extraction; step 4, model training; 5, real-time monitoring is carried out; step 6, early warning and control; the load detection system comprises a data acquisition module, a data processing module, a real-time monitoring module and an early warning and control module. According to the method, the actual working state of the load of the warmer can be continuously self-updated through the prediction model, the real-time working state of the warmer is judged according to the actual working state, the temperature of the second load is autonomously regulated and controlled according to the actual environment temperature so that the second load can meet the preset requirement of a user, early warning is conducted in the abnormal state, the load power supply is disconnected, and user safety is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of heaters, and in particular relates to a load detection system and method suitable for heaters. Background Art

[0002] Air-heating heaters are heating devices that heat the air through electric heating elements and then blow the hot air into the room through a fan. Air-heating heaters mainly use electric heating elements (such as PTC ceramic heating elements) to generate heat, and then quickly blow the hot air out through a built-in fan to achieve a rapid heating effect. This design enables air-heating heaters to increase the indoor temperature in a short period of time, and is especially suitable for environments that require rapid heating. This type of air-heating heater is generally installed in the bathroom, which results in its working environment being generally high temperature and high humidity, and there is a lot of water vapor.

[0003] In traditional air-heating heaters, load 1 (blowing motor) and load 2 (heating module) are operated in conjunction. The correct way to use them is to turn on load 1 (blowing motor) first, and then load 2 (heating module). This ensures that the temperature inside the heater box is controlled within a certain range, thereby ensuring the safety and reliability of the heater. Otherwise, the heat generated by the heating module cannot be taken out of the machine and accumulates inside the machine, causing the temperature inside the machine to rise sharply, thereby affecting the reliability and stability of the entire product, and even causing accidents in serious cases.

[0004] In the detection and control system of traditional air-heating heaters, the detection method is relatively simple, and only the real-time current, voltage and power parameters of load one (blowing motor) and load two (heating module) are detected to determine whether the operation is abnormal. Due to the working environment of the air-heating heater, its working state will be greatly affected by external water vapor, temperature and humidity. It is inaccurate to judge its working state only by current, voltage and power parameters, and it is impossible to make an effective judgment on its working state. Therefore, there are safety hazards. At the same time, the control accuracy is low, and users are prone to inaccurate temperature control when using it. Summary of the invention

[0005] The purpose of the present invention is to provide a load detection system and method suitable for a heater, which can continuously self-update the actual working status of the heater load through a prediction model, and judge its real-time working status based on this, and autonomously adjust the temperature of load 2 according to the actual ambient temperature to meet the user's preset requirements, and issue an early warning and disconnect the load power supply in an abnormal state to ensure user safety.

[0006] The technical solution adopted by the present invention is as follows: A load detection method applicable to a heater comprises the following steps: Step 1: Data collection: collect the operating data of the heater in real time through sensors; Step 2: Data preprocessing: Clean and organize the operation data collected in time series to remove outliers and noise; Step 3, feature extraction: extracting features related to the heater load from the preprocessed operating data; Step 4: Model training: Use machine learning algorithms to train prediction models to predict heater load conditions; Step 5: Real-time monitoring: Input the real-time collected operation data into the prediction model to obtain the current load of the heater; Step 6: Early warning and control: Make corresponding judgments based on the current load of the heater; if the load exceeds the preset threshold, the system will send out an alarm signal and take corresponding control measures; otherwise, it will continue to run.

[0007] As a preferred solution, the load includes load one and load two, load one is a blower motor, and load two is a heating module, and the two operate in association with each other.

[0008] As a preferred solution, the operating data includes real-time current, voltage, and power parameters of the load 1 and the load 2; It also includes the load integrated surface temperature, the load second temperature, the current ambient temperature, and the heater body internal temperature and humidity parameters; The surface temperature of the blower motor and the temperature and humidity parameters of the heater body are collected by installing sensors on the surface of the blower motor and the heater body.

[0009] As a preferred solution, the data preprocessing in step 2 includes the following steps: Step 2.1, stability judgment: obtain the operating data, and judge the stability of the operating data by referring to the benchmark parameters in the database; if the data is stable, proceed to step 3, otherwise proceed to step 2.2; Step 2.2, differential stability: Use differential operation to stabilize the running data and further determine the stability of the running data; if the data is stable, proceed to step 3, otherwise continue to step 2.2 until the required stability is achieved; Step 2.3, white noise test: perform white noise test on the stable operation data, remove the white noise operation data, and retain the stable non-white noise operation data.

[0010] As a preferred solution, the machine learning algorithm in step 4 is an ARIMA model algorithm, and the specific steps of model training include: Step 4.1, Stability judgment: Get the historical data in the database, and judge the stability of the historical data by referring to the benchmark parameters in the database; if the data is stable, proceed to step 3, otherwise proceed to step 2.2; Step 4.2, differential stability: Use differential operation to stabilize the historical data and further determine the stability of the historical data; if the data is stable, proceed to step 3, otherwise continue to step 2.2 until the required stability is achieved; Step 4.3, white noise test: perform white noise test on stable historical data, remove white noise historical data, and retain stable non-white noise historical data; Step 4.4, model identification: fitting the prediction model to the stationary non-white noise historical data; Step 4.5, prediction model test: check whether the residual data fitted by the prediction model is white noise data, and at the same time test the significance of the prediction model fitting results with reference to the benchmark parameters in the database, so as to continuously correct and train the prediction model.

[0011] As a preferred solution, the expression of the difference operation is: First-order difference: ; Second order difference: ; Order difference: ; in, Represents a time series at a point in time The value of is the difference operator, represents the first-order difference, that is, the time point Compared with the previous time point The change in value between represents the second-order difference, represents the k-th order difference.

[0012] As a preferred solution, the expression of the ARIMA model algorithm is: ; in, Indicates at a point in time The value of the dependent variable, is a constant term representing the average level or intercept in the time series, is the autoregressive coefficient, is the error term, is the moving average coefficient, and are the number of autoregressive terms and the number of moving average terms, respectively.

[0013] As a preferred solution, it also includes step 7, data feedback: feeding back the early warning operation data to the database, serving as a historical data sample for training the prediction model; Among them, when the historical data samples reach a predetermined threshold, the early historical data samples are deleted to maintain the vitality of the historical data samples and reduce the computing pressure.

[0014] A load detection system for a heater, using the above-mentioned load detection method for a heater, comprises: A data acquisition module, used to collect the operation data of the heater in continuous operation in real time according to the time series through sensors; A data processing module is used to clean and organize the operation data collected in time series, remove abnormal values ​​and noise, extract features related to the heater load from the pre-processed operation data, and obtain processed operation data; A real-time monitoring module is used to input the real-time collected operation data into the prediction model to obtain the current load condition of the heater; The early warning and control module is used to make corresponding judgments based on the current load of the heater; if the load exceeds the preset threshold, the system will send out an alarm signal and take corresponding control measures; otherwise, it will continue to run.

[0015] As a preferred solution, the early warning method in the early warning and control module is a buzzer alarm, and the control measure is to autonomously adjust the corresponding temperature of load two through the control system so that the current ambient temperature reaches the expected value or disconnect the load power supply through the control system, and reconnect the load power supply when the load integrated surface temperature, load two temperature and the temperature and humidity parameters inside the heater body are reduced to a predetermined threshold.

[0016] The technical effects achieved by the present invention are: The present invention collects the real-time current, voltage, power parameters of load one and load two in the heater, as well as the temperature of the load integrated surface and the temperature parameters of load two, and utilizes the constantly revised and trained ARIMA model. During the use of the heater, it can accurately predict and determine its current use status, and promptly issue an early warning and take corresponding control measures to ensure user safety. Compared with the traditional control mode, the control mode can be adaptively adjusted according to the mechanical state inside the heater to reduce the judgment error and improve the accuracy of use status detection; secondly, combined with the current ambient temperature and the temperature and humidity parameters of the heater body, the temperature of load two and the wind force of load one are autonomously regulated to meet the user's preset requirements, so as to ensure the accuracy of temperature control and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1is a structural diagram of a load detection method in an embodiment of the present invention; Figure 2 is a structural schematic diagram of a specific method of step 2 of the load detection method in an embodiment of the present invention; Figure 3 is a structural schematic diagram of a specific method of step 4 of the load detection method in an embodiment of the present invention; Figure 4 It is a structural schematic diagram of a load detection system in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

[0019] like Figure 1-Figure 3 As shown, a load detection method applicable to a heater comprises the following steps: Step 1: Data collection: collect the operating data of the heater in real time through sensors; Step 2: Data preprocessing: Clean and organize the operation data collected in time series to remove outliers and noise; Step 3, feature extraction: extracting features related to the heater load from the preprocessed operating data; Step 4: Model training: Use machine learning algorithms to train prediction models to predict heater load conditions; Step 5: Real-time monitoring: Input the real-time collected operation data into the prediction model to obtain the current load of the heater; Step 6: Early warning and control: Make corresponding judgments based on the current load of the heater; if the load exceeds the preset threshold, the system will send out an alarm signal and take corresponding control measures; otherwise, it will continue to run.

[0020] Among them, the load includes load one and load two, load one is a blower motor, and load two is a heating module, and the two operate in a linked manner.

[0021] Among them, the machine learning algorithm in step 4 is the ARIMA model algorithm, that is, the difference autoregressive moving average model, which is a statistical model for time series prediction. It converts the non-stationary time series into a stationary time series through differential operation, and then establishes an autoregressive moving average (ARMA) model for prediction, and then can judge whether there is a problem with the load operation of the heater according to the temperature increase of the load 2 heating module over time and related parameters, so as to ensure the stable operation of the equipment and the safety of users.

[0022] In the actual operation process, the expression of the ARIMA model algorithm is: ; in, Indicates at a point in time The value of the dependent variable, which in time series analysis usually represents a quantity that changes over time, in this case the change in temperature; is a constant term, representing the average level or intercept in the time series; are the autoregressive coefficients, which represent the current time point The dependent variable Before Time points ( , ,…, ) , ,…, The linear relationship between them, these coefficients measure the degree of influence of historical data on current values; is the error term (also called residual term, noise term), representing the time point When the model fails to explain the random fluctuations or changes, it is usually assumed to be white noise that follows a normal distribution; are the moving average coefficients, which represent the current time point The error term Before Time points ( , ,…, ) , ,…, The linear relationship between them, these coefficients measure the impact of historical errors on current errors; and They are the number of autoregressive terms and the number of moving average terms, which determine the amount of historical data used in the model. In the ARIMA model, the selection of these two parameters is usually based on the characteristics of the data and the modeling objectives.

[0023] It should be noted that the operating data includes the real-time current, voltage and power parameters of load one and load two; it also includes the temperature of the load integrated surface, the temperature of load two, the current ambient temperature and the temperature and humidity parameters of the heater body.

[0024] Among them, the current, voltage and power parameters can be directly obtained through the circuit elements through the intelligent control system, and the surface temperature of the hair dryer motor and the temperature and humidity parameters inside the heater body are collected by installing sensors on the surface of the hair dryer motor and the body of the heater.

[0025] See attached Figure 2 The data preprocessing in step 2 specifically includes the following steps: Step 2.1, stability judgment: obtain the operating data, and judge the stability of the operating data by referring to the benchmark parameters in the database; if the data is stable, proceed to step 3, otherwise proceed to step 2.2; Step 2.2, differential stability: Use differential operation to stabilize the running data and further determine the stability of the running data; if the data is stable, proceed to step 3, otherwise continue to step 2.2 until the required stability is achieved; Step 2.3, white noise test: perform white noise test on the stable operation data, remove the white noise operation data, and retain the stable non-white noise operation data.

[0026] In this embodiment, the white noise test method is the Ljung-Box test. It first assumes that the running data is white noise, and then calculates the statistic and P value based on the assumption. If the P value is less than the set significance level (such as 0.05), the null hypothesis is rejected and it is considered that the sequence is not white noise. On the contrary, if the P value is greater than the significance level, the null hypothesis cannot be rejected and it is considered that the sequence may be white noise.

[0027] Of course, in other embodiments, the normal distribution method may be used to check whether the operating data is white noise. The corresponding checking method may be selected according to specific application requirements.

[0028] In this embodiment, in step three, features related to the heater load are extracted from the preprocessed operating data, specifically, the real-time current, voltage, power parameters of load one and load two, as well as the load integrated surface temperature and load two temperature parameters are extracted; and the current ambient temperature and the heater body internal temperature and humidity parameters are extracted as auxiliary parameters. When the current, voltage, power parameters, the load integrated surface temperature and the load two temperature parameters cannot accurately determine the operating status of the heater, the current ambient temperature and the heater body internal temperature and humidity parameters are added for re-determination to ensure accuracy while reducing computing pressure.

[0029] See attached Figure 3 The specific steps of model training in step 4 include: Step 4.1, Stability judgment: Get the historical data in the database, and judge the stability of the historical data by referring to the benchmark parameters in the database; if the data is stable, proceed to step 3, otherwise proceed to step 2.2; Step 4.2, differential stability: Use differential operation to stabilize the historical data and further determine the stability of the historical data; if the data is stable, proceed to step 3, otherwise continue to step 2.2 until the required stability is achieved; Step 4.3, white noise test: perform white noise test on stable historical data, remove white noise historical data, and retain stable non-white noise historical data; Step 4.4, model identification: fitting the prediction model to the stationary non-white noise historical data; Step 4.5, prediction model test: check whether the residual data fitted by the prediction model is white noise data, and at the same time test the significance of the prediction model fitting results with reference to the benchmark parameters in the database, so as to continuously correct and train the prediction model.

[0030] Through continuous correction and training of the prediction model, the dynamic characteristics of the operating data can be better captured. During the correction process, the structure of the model (such as autoregressive terms, difference orders, and moving average terms) may change, which helps to better understand the dynamic mechanism behind the data, minimize prediction errors, and improve the accuracy of autonomous regulation.

[0031] In order to further optimize the prediction model, the seventh step, data feedback, is also included: feeding back the early warning operation data to the database to serve as historical data samples for training the prediction model; Among them, when the historical data samples reach a predetermined threshold, the early historical data samples are deleted to keep the historical data samples active and reduce the computing pressure. By increasing and updating the number of samples, the prediction model can be updated in real time according to the aging of load one and load two during continuous use. As more data is added and the model parameters are fine-tuned, the prediction error of the model will gradually decrease, thereby improving the accuracy of the prediction and ensuring precise control of the heater.

[0032] In step 2 and step 4, the expression of the difference operation is: First-order difference: ; Second order difference: ; Order difference: ; in, Represents a time series at a point in time The value of is the difference operator, represents the first-order difference, that is, the time point Compared with the previous time point The change in value between represents the second-order difference, represents the k-th order difference.

[0033] In this embodiment, the first-order difference is used to stabilize the operating data. The differential operation mathematical model in the ARIMA model is one of the key steps to convert a non-stationary time series into a stationary time series through differential operation. The number and form of differential operations depend on the stationarity test results and prediction targets of the time series. In the actual calculation process, the number of calculations needs to be determined according to the specific stationarity of the data until the predetermined stationarity is achieved.

[0034] Of course, in other embodiments, a higher-order differential calculation method may be selected as needed to adapt to different requirements.

[0035] like Figure 1-Figure 4 As shown, a load detection system applicable to a heater, using the above-mentioned load detection method applicable to a heater, comprises: The data acquisition module is used to collect the operating data of the heater in real time according to the time series through the sensor. Specifically, it is necessary to collect the real-time current, voltage, and power parameters of load one and load two, the load integrated surface temperature, the load two temperature, the current ambient temperature, and the body temperature and humidity parameters of the heater, and transmit these data signals to the data processing module.

[0036] The data processing module is used to clean and organize the operation data collected in time series, remove abnormal values ​​through differential operation, and test the noise through Ljung-Box test method and remove it from the operation data, and then extract the characteristics related to the heater load (i.e., the real-time current, voltage, power parameters of load one and load two, and the temperature of the load integrated surface and the temperature parameter of load two) from the pre-processed operation data to obtain the processed operation data; If the above data cannot accurately predict the operating status of the heater, it is necessary to extract the current ambient temperature and the temperature and humidity parameters of the heater body.

[0037] The real-time monitoring module is used to input the real-time collected operation data into the prediction model to obtain the current load conditions of the heater, including low-load operation, medium-load operation and high-load operation.

[0038] The early warning and control module is used to make corresponding judgments according to the current load of the heater; under the above three operating conditions, the surface temperature of the load 1 blower motor and the temperature rise state of the load 2 heating module need to be synchronously matched. If they do not match the preset standard parameters in the data, it means that its operating state is abnormal; If the abnormal value of the load operation status exceeds the preset threshold, the system will issue an alarm signal and take corresponding control measures; Secondly, if the temperature and humidity parameters inside the heater exceed the preset threshold, it means that the environment is no longer suitable for the operation of load 1 and load 2, and there is a safety hazard. The system will also send out an alarm signal and take corresponding control measures. Otherwise, it will continue to operate. Secondly, during the actual operation, when the ambient temperature parameters are different from the parameters set by the user, it is necessary to autonomously adjust the working status of the load 1 blowing motor and the load 2 heating module through the control system to make the environment stable and meet the user's preset requirements, thereby ensuring the accuracy of the heater temperature control. At the same time, the control method is continuously optimized during the continuous use and aging of the load 1 blowing motor and the load 2 heating module to achieve the control purpose.

[0039] Wherein, the warning mode in the warning and control module is a buzzer alarm, and the control measure is to autonomously adjust the corresponding temperature of load 2 or the wind force of load 1 through the control system so that the ambient temperature reaches the expected value; Of course, when the load operating status is abnormal or the temperature and humidity parameters inside the heater body exceed the preset thresholds, the load power supply is disconnected through the control system, and the load power supply is reconnected when the load integrated surface temperature, load two temperature and the temperature and humidity parameters inside the heater body drop to the preset thresholds, so as to ensure the safe operation of the heater and reduce safety hazards.

[0040] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A load detection method suitable for a heater, characterized in that: The following steps are involved: Step 1: Data collection: collect the operating data of the heater in real time through sensors; Step 2: Data preprocessing: Clean and organize the operation data collected in time series to remove outliers and noise; Step 3, feature extraction: extracting features related to the heater load from the preprocessed operating data; Step 4: Model training: Use machine learning algorithms to train prediction models to predict heater load conditions; Step 5: Real-time monitoring: Input the real-time collected operation data into the prediction model to obtain the current load of the heater; Step 6: Early warning and control: Make corresponding judgments based on the current load of the heater; if the load exceeds the preset threshold, the system will send out an alarm signal and take corresponding control measures; otherwise, it will continue to run.

2. A load detection method suitable for a heater according to claim 1, characterized in that: The loads include load one and load two, load one is a blower motor, and load two is a heating module, and the two operate in association with each other.

3. A load detection method suitable for a heater according to claim 2, characterized in that: The operation data includes real-time current, voltage, and power parameters of the load 1 and the load 2; It also includes the load integrated surface temperature, the load second temperature, the current ambient temperature, and the heater body internal temperature and humidity parameters; The surface temperature of the blower motor and the temperature and humidity parameters of the heater body are collected by installing sensors on the surface of the blower motor and the heater body.

4. A load detection method suitable for a heater according to claim 1, characterized in that: The data preprocessing in step 2 includes the following steps: Step 2.1, stability judgment: obtain the operating data, and judge the stability of the operating data by referring to the benchmark parameters in the database; if the data is stable, proceed to step 3, otherwise proceed to step 2.2; Step 2.2, differential stability: Use differential operation to stabilize the running data and further determine the stability of the running data; if the data is stable, proceed to step 3, otherwise continue to step 2.2 until the required stability is achieved; Step 2.3, white noise test: perform white noise test on the stable operation data, remove the white noise operation data, and retain the stable non-white noise operation data.

5. A load detection method suitable for a heater according to claim 1, characterized in that: The machine learning algorithm in step 4 is an ARIMA model algorithm, and the specific steps of its model training include: Step 4.1, Stability judgment: Get the historical data in the database, and judge the stability of the historical data by referring to the benchmark parameters in the database; if the data is stable, proceed to step 3, otherwise proceed to step 2.2; Step 4.2, differential stability: Use differential operation to stabilize the historical data and further determine the stability of the historical data; if the data is stable, proceed to step 3, otherwise continue to step 2.2 until the required stability is achieved; Step 4.3, white noise test: perform white noise test on stable historical data, remove white noise historical data, and retain stable non-white noise historical data; Step 4.4, model identification: fitting the prediction model to the stationary non-white noise historical data; Step 4.5, prediction model test: check whether the residual data fitted by the prediction model is white noise data, and at the same time test the significance of the prediction model fitting results with reference to the benchmark parameters in the database, so as to continuously correct and train the prediction model.

6. A load detection method suitable for a heater according to claim 4 or 5, characterized in that: The expression of the difference operation is: First-order difference: ; Second order difference: ; Order difference: ; in, Represents a time series at a point in time The value of is the difference operator, represents the first-order difference, that is, the time point Compared with the previous time point The change in value between represents the second-order difference, represents the k-th order difference.

7. A load detection method suitable for a heater according to claim 5, characterized in that: The expression of the ARIMA model algorithm is: ； in, Indicates at a point in time The value of the dependent variable, is a constant term representing the average level or intercept in the time series, is the autoregressive coefficient, is the error term, is the moving average coefficient, and are the number of autoregressive terms and the number of moving average terms, respectively.

8. A load detection method suitable for a heater according to claim 1, characterized in that: It also includes step seven, data feedback: feeding back the early warning operation data to the database, serving as historical data samples for training the prediction model; Among them, when the historical data samples reach a predetermined threshold, the early historical data samples are deleted to maintain the vitality of the historical data samples and reduce the computing pressure.

9. A load detection system for a heater, using the load detection method for a heater as claimed in any one of claims 1 to 8, characterized in that: include: A data acquisition module, used to collect the operation data of the heater in continuous operation in real time according to the time series through sensors; A data processing module is used to clean and organize the operation data collected in time series, remove abnormal values ​​and noise, extract features related to the heater load from the pre-processed operation data, and obtain processed operation data; A real-time monitoring module is used to input the real-time collected operation data into the prediction model to obtain the current load condition of the heater; The early warning and control module is used to make corresponding judgments based on the current load of the heater; if the load exceeds the preset threshold, the system will send out an alarm signal and take corresponding control measures; otherwise, it will continue to run.

10. A load detection system suitable for a heater according to claim 9, characterized in that: The early warning method in the early warning and control module is a buzzer alarm, and the control measures are to autonomously adjust the corresponding temperature of load two through the control system so that the current ambient temperature reaches the expected value or disconnect the load power supply through the control system, and reconnect the load power supply when the load integrated surface temperature, load two temperature and the temperature and humidity parameters inside the heater body are reduced to a predetermined threshold.

Images