A method for monitoring and early warning of ultrasonic cleaning machine operation

Through real-time monitoring and compensation algorithms, the temperature and bubble data in the ultrasonic cleaning machine are processed, and the accuracy of turbidity monitoring of cleaning liquid pollutants is solved, and efficient cleaning liquid management is achieved, which extends the service life and improves the equipment performance.

CN118467916BActive Publication Date: 2024-12-03深圳市林科超声波洗净设备有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410931118.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-12-03
Estimated Expiration
2044-07-12

AI Technical Summary

Technical Problem

During ultrasonic cleaning, the temperature of the cleaning solution and bubbles interfere with the measurement accuracy of the turbidity sensor, resulting in inaccurate monitoring of the pollutant concentration, and the existing turbidity compensation algorithm has a large amount of calculation and insufficient real-time performance.

Method used

By obtaining real-time temperature and bubble data, the turbidity compensation algorithm is used to correct the turbidity measurement value, trigger the alarm mechanism and replacement prompts, record and store monitoring data, form a historical database, and optimize the service life estimate of the cleaning liquid.

Benefits of technology

It improves the accuracy and real-time performance of cleaning liquid quality monitoring, extends the service life of cleaning liquid, ensures the safety of equipment and operations, and improves the performance and economic benefits of ultrasonic cleaning machines.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118467916B_ABST
    Figure CN118467916B_ABST
Patent Text Reader

Abstract

The present application provides an ultrasonic cleaning machine operation monitoring and early warning method, comprising: obtaining real-time temperature data and bubble data of a cleaning fluid of the ultrasonic cleaning machine, and using them as input parameters of a turbidity compensation algorithm; measuring the turbidity value of the cleaning fluid in real time through a turbidity sensor to obtain original turbidity measurement data; comparing the compensated turbidity measurement value with a preset pollutant concentration threshold to determine whether the degree of contamination of the cleaning fluid exceeds the standard; if the pollutant concentration exceeds the threshold, triggering a cleaning fluid replacement prompt to notify the user to replace the cleaning fluid in time to ensure the cleaning effect; continuously monitoring the turbidity change trend over a period of time, determining the degree of aging of the cleaning fluid through a data analysis algorithm, and estimating its remaining usable time; recording and storing turbidity monitoring data, compensation parameters, and abnormal conditions to form a historical database for cleaning fluid quality monitoring.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an ultrasonic cleaning machine operation monitoring and early warning method. Background Art

[0002] During the cleaning process of an ultrasonic cleaning machine, the concentration of pollutants in the cleaning fluid is a key factor affecting the cleaning quality. In order to ensure the cleaning effect, the concentration of pollutants in the cleaning fluid needs to be monitored in real time. The commonly used method for monitoring pollutant concentration is to use a turbidity sensor to measure the turbidity of the cleaning fluid and calculate the concentration of pollutants based on the turbidity value. However, in the actual ultrasonic cleaning process, factors such as the temperature and bubbles of the cleaning fluid will affect the measurement accuracy of the turbidity sensor.

[0003] The temperature change of the cleaning liquid will cause the refractive index of the liquid to change, which will affect the measurement results of the turbidity sensor. In addition, a large number of bubbles will be generated during the ultrasonic cleaning process, and the presence of bubbles will also interfere with the measurement of the turbidity sensor. Bubbles will change the scattering and absorption characteristics of the liquid medium to light, making it impossible for the turbidity sensor to accurately reflect the actual pollutant concentration.

[0004] Therefore, in order to achieve accurate monitoring of pollutant concentrations, it is necessary to introduce a turbidity compensation algorithm based on turbidity measurement. The turbidity compensation algorithm needs to comprehensively consider the influence of multiple factors such as temperature and bubbles, and correct the turbidity measurement value by establishing a mathematical model to eliminate the interference of temperature and bubbles. However, it is not easy to establish an accurate turbidity compensation model. It requires a lot of experimental data support, and the parameters of the model also need to be adjusted and optimized according to different cleaning processes. At the same time, the turbidity compensation algorithm has a large amount of calculation, which puts higher requirements on the real-time performance of the monitoring system. Summary of the invention

[0005] The present invention provides an ultrasonic cleaning machine operation monitoring and early warning method, which mainly includes:

[0006] Obtain the real-time temperature data and bubble data of the ultrasonic cleaning machine cleaning fluid and use them as input parameters of the turbidity compensation algorithm;

[0007] The turbidity value of the cleaning liquid is measured in real time by the turbidity sensor to obtain the original turbidity measurement data;

[0008] According to the obtained temperature parameters and bubble parameters, the turbidity compensation algorithm is used to correct and calculate the original turbidity measurement value to obtain the compensated turbidity measurement value;

[0009] If the temperature parameter or bubble parameter exceeds the preset threshold range, it is judged as an abnormal situation, triggering the alarm mechanism to prompt the user to check and handle it;

[0010] Compare the compensated turbidity measurement value with the preset pollutant concentration threshold to determine whether the degree of contamination of the cleaning fluid exceeds the standard;

[0011] If the pollutant concentration exceeds the threshold, the cleaning fluid replacement reminder will be triggered to notify the user to replace the cleaning fluid in time to ensure the cleaning effect;

[0012] Continuously monitor the turbidity change trend over a period of time, determine the aging degree of the cleaning fluid through data analysis algorithms, and estimate its remaining usable time;

[0013] The turbidity monitoring data, compensation parameters and abnormal conditions are recorded and stored to form a historical database for cleaning fluid quality monitoring.

[0014] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0015] The present invention discloses a turbidity monitoring and cleaning fluid quality management system for an ultrasonic cleaning machine. The unique business scenario problem solved by the system is how to monitor and adjust the turbidity of the cleaning fluid in real time to ensure the cleaning effect and optimize the service life of the cleaning fluid. The present invention achieves this goal through the following steps:

[0016] 1. Obtain real-time temperature data and bubble data of ultrasonic cleaning fluid, which directly affects the accuracy of turbidity measurement.

[0017] 2. Use a turbidity sensor to measure the turbidity value of the cleaning fluid in real time and obtain the original turbidity measurement data.

[0018] 3. According to the acquired temperature parameters and bubble parameters, the turbidity compensation algorithm is used to correct the original turbidity measurement value to obtain the compensated turbidity measurement value. This step is the core, which ensures the accuracy of the measurement data and reduces the interference of external factors.

[0019] 4. If the temperature parameter or bubble parameter exceeds the preset threshold range, it is judged as an abnormal situation, triggering the alarm mechanism to prompt the user to check and handle it. This function ensures the safe operation of the system.

[0020] 5. By comparing the compensated turbidity measurement value with the preset pollutant concentration threshold, it is determined whether the degree of contamination of the cleaning fluid exceeds the standard.

[0021] 6. If the pollutant concentration exceeds the threshold, the cleaning fluid replacement reminder will be triggered, notifying the user to replace the cleaning fluid in time to ensure the cleaning effect.

[0022] 7. Continuously monitor the turbidity trend over a period of time, determine the aging degree of the cleaning fluid through data analysis algorithms, and estimate its remaining usable time.

[0023] 8. Record and store turbidity monitoring data, compensation parameters, and abnormal conditions to form a historical database for cleaning fluid quality monitoring.

[0024] The overall technical effects or functions are summarized as follows: The present invention provides an efficient cleaning fluid quality monitoring method, which not only improves the cleaning efficiency, but also extends the service life of the cleaning fluid through real-time data monitoring and intelligent compensation algorithms, while ensuring the safety of equipment and operation. This system can significantly improve the performance and economic benefits of ultrasonic cleaning machines. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 The present invention is a flow chart of an ultrasonic cleaning machine operation monitoring and early warning method.

[0026] Figure 2 It is a schematic diagram of an ultrasonic cleaning machine operation monitoring and early warning method of the present invention.

[0027] Figure 3 It is another schematic diagram of an ultrasonic cleaning machine operation monitoring and early warning method of the present invention. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be described clearly and completely below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0029] like Figure 1-3 In this embodiment, an ultrasonic cleaning machine operation monitoring and early warning method may specifically include:

[0030] S101, obtaining real-time temperature data and bubble data of the cleaning fluid of the ultrasonic cleaning machine, and using them as input parameters of the turbidity compensation algorithm.

[0031] The real-time temperature data and bubble data of the ultrasonic cleaning machine cleaning fluid are obtained and used as input parameters of the turbidity compensation algorithm. The specific implementation process is as follows: by installing a temperature sensor in the ultrasonic cleaning machine cleaning fluid, the temperature data of the cleaning fluid is obtained in real time. According to the obtained temperature data, the least squares method is used to fit the functional relationship between temperature and time to obtain the curve equation of temperature changing with time. By installing a bubble sensor in the ultrasonic cleaning machine cleaning fluid, the bubble data in the cleaning fluid is obtained in real time. According to the obtained bubble data, the Kalman filter algorithm is used to filter the bubble data to obtain a smooth bubble data curve. The temperature data and bubble data are used as input parameters of the turbidity compensation algorithm. The multivariate linear regression algorithm is used to establish a mathematical model between temperature, bubbles and turbidity to obtain the turbidity compensation equation. According to the turbidity compensation equation, the turbidity of the ultrasonic cleaning machine cleaning fluid is compensated in real time to obtain the compensated turbidity value. The compensated turbidity value is compared with the preset turbidity threshold. If the compensated turbidity value exceeds the threshold, it is judged that the cleaning fluid needs to be replaced; if the compensated turbidity value does not exceed the threshold, it is judged that the cleaning fluid can continue to be used. According to the judgment results, the cleaning liquid of the ultrasonic cleaning machine is automatically controlled to change, ensuring that the turbidity of the cleaning liquid is always within a reasonable range to ensure the cleaning effect.

[0032] Specifically, for example, by installing a temperature sensor in the cleaning fluid of an ultrasonic cleaning machine, the temperature data of the cleaning fluid is obtained every 5 seconds, and 120 data points are obtained continuously. According to the obtained temperature data, the least squares method is used to fit the functional relationship between temperature and time, and the curve equation of temperature change over time is obtained as follows: T = 0.5t + 30, where T is temperature in degrees Celsius and t is time in seconds. For example, by installing a bubble sensor in the cleaning fluid of an ultrasonic cleaning machine, the bubble data in the cleaning fluid is obtained every 2 seconds, and 300 data points are obtained continuously. According to the obtained bubble data, the Kalman filter algorithm is used to filter the bubble data. The mean value of the bubble data before filtering is 1000 / ml, and the mean value of the bubble data after filtering is 980 / ml. The temperature data and bubble data are used as input parameters of the turbidity compensation algorithm. The mathematical model between temperature, bubbles and turbidity is established by using the multivariate linear regression algorithm, and the turbidity compensation equation is obtained as follows: Turbidity = 2 × Temperature + 0.5 × Bubbles + 10, where Turbidity is turbidity in NTU, Temperature is temperature in degrees Celsius, and Bubbles is the number of bubbles in pieces / ml. According to the turbidity compensation equation, the turbidity of the cleaning fluid of the ultrasonic cleaning machine is compensated in real time. For example, the current cleaning fluid temperature is 45 degrees Celsius, the number of bubbles is 1200 pieces / ml, and the turbidity value after compensation is 79NTU after substitution. The compensated turbidity value is compared with the preset turbidity threshold of 80 NTU. If the compensated turbidity value exceeds the threshold, it is judged that the cleaning fluid needs to be replaced; if the compensated turbidity value does not exceed the threshold, it is judged that the cleaning fluid can continue to be used. For example, if the current compensated turbidity value is 79 NTU and does not exceed the preset threshold of 80 NTU, it is judged that the cleaning fluid can continue to be used. According to the judgment result, the replacement of the cleaning fluid of the ultrasonic cleaning machine is automatically controlled. For example, when the turbidity value after three consecutive compensations exceeds the preset threshold, the cleaning fluid is automatically replaced to ensure that the turbidity of the cleaning fluid is always within a reasonable range and to ensure the cleaning effect.

[0033] S102, measuring the turbidity value of the cleaning liquid in real time through a turbidity sensor to obtain original turbidity measurement data.

[0034] The turbidity value of the cleaning fluid is measured in real time according to the turbidity sensor to obtain the original turbidity measurement data. The median filter algorithm is used to filter the original turbidity measurement data, remove outliers, and obtain the filtered turbidity measurement data. The least squares method is used to fit the filtered turbidity measurement data to obtain a fitting curve of turbidity change over time. The turbidity change rate is calculated according to the fitting curve to determine whether the turbidity change rate exceeds the preset threshold. If it exceeds the threshold, it is determined that the cleaning fluid is contaminated and needs to be replaced; if it does not exceed the threshold, it is determined that the cleaning fluid is not contaminated and can continue to be used. The smoothed value of turbidity is calculated by the exponentially weighted moving average algorithm to reduce the fluctuation of turbidity measurement data and obtain the smoothed turbidity value. According to the smoothed turbidity value, the control limit algorithm is used to determine whether the turbidity exceeds the control limit. If it exceeds the control limit, a cleaning fluid contamination alarm is issued; if it does not exceed the control limit, it is considered that the cleaning fluid is not contaminated and the turbidity change continues to be monitored. The change trend of turbidity in the future period is predicted by the autoregressive moving average model to obtain the predicted value of turbidity. According to the predicted turbidity value and the current turbidity value, a linear discriminant analysis algorithm is used to determine whether the cleaning fluid needs to be replaced. If it needs to be replaced, an instruction to replace the cleaning fluid is issued; if it does not need to be replaced, the turbidity changes continue to be monitored.

[0035] Specifically, for example, the turbidity sensor collects data every 1 second for 100 consecutive times to obtain the original turbidity measurement data. For example, for a set of original turbidity measurement data: 2, 5, 1, 3, 7, 4, 6, the median filter algorithm is used, and the window size is selected as 5, then the filtered result is: 2, 4, 5, 4, 6. Among them, 7 is removed as an outlier. For example, the least squares fitting is performed on the 50 sets of filtered turbidity measurement data, and the fitting curve of turbidity change over time is: y = 0.2x + 15, where x is time and y is turbidity value. For example, according to the fitting curve, the current turbidity change rate is calculated to be 0.2 NTU / s, the preset turbidity change rate threshold is 0.5 NTU / s, and the current turbidity change rate does not exceed the threshold. It is judged that the cleaning fluid is not contaminated and can continue to be used. For example, for a set of turbidity measurement data: 2, 5, 3, 4, 6, the exponential weighted moving average algorithm is used, and the weighting coefficient is selected as 2, then the smoothed turbidity values ​​are: 2, 26, 25, 28, 34. For example, based on the smoothed turbidity values, the control limit algorithm is used, the upper control limit is set to 8 NTU, the lower control limit is set to 8 NTU, and the current turbidity value is 34 NTU, which does not exceed the control limit. It is considered that the cleaning fluid is not contaminated, and the turbidity changes continue to be monitored. For example, through the autoregressive moving average model, the turbidity change trend in the next 5 minutes is predicted, and the turbidity prediction values ​​are: 36, 38, 41, 44, 47. For example, based on the turbidity prediction value and the current turbidity value of 34 NTU, the linear discriminant analysis algorithm is used to set the turbidity replacement threshold to 5 NTU. The predicted value does not exceed the threshold, and it is judged that the cleaning fluid does not need to be replaced for the time being, and the turbidity changes continue to be monitored.

[0036] S103, according to the obtained temperature parameters and bubble parameters, a turbidity compensation algorithm is used to correct and calculate the original turbidity measurement value to obtain a compensated turbidity measurement value.

[0037] According to the obtained temperature parameters and bubble parameters, the turbidity compensation algorithm is used to correct the original turbidity measurement value to obtain the compensated turbidity measurement value. The specific implementation process is as follows: the water temperature data measured by the temperature sensor and the bubble data measured by the bubble sensor are obtained. According to the obtained water temperature data, the water temperature compensation coefficient is obtained by the table lookup method. If the water temperature is higher than the set threshold, the water temperature compensation coefficient is set to a positive value; if the water temperature is lower than the set threshold, the water temperature compensation coefficient is set to a negative value. According to the obtained bubble data, the bubble compensation coefficient is obtained by the table lookup method. If the bubble data is greater than the set threshold, the bubble compensation coefficient is set to a positive value; if the bubble data is less than the set threshold, the bubble compensation coefficient is set to a negative value. The original measurement value of the turbidity sensor is obtained. The obtained water temperature compensation coefficient and bubble compensation coefficient are substituted into the formula of the turbidity compensation algorithm, and the original turbidity measurement value is corrected and calculated to obtain the compensated turbidity measurement value. The turbidity compensation algorithm adopts a weighted average algorithm, and the water temperature compensation coefficient and the bubble compensation coefficient are used as weights, and the weighted average is performed with the original turbidity measurement value to obtain the final compensated turbidity measurement value. Determine whether the compensated turbidity measurement value is within the normal range. If the compensated turbidity measurement value exceeds the normal range, it is limited to the upper or lower limit of the normal range to obtain the final turbidity measurement result. If the compensated turbidity measurement value is within the normal range, it is directly output as the final turbidity measurement result. The final turbidity measurement result is transmitted to the subsequent data processing module for further analysis and control.

[0038] Specifically, for example, the water temperature data is 25℃ and the bubble data is 10 / minute. For example, through the table lookup method, when the water temperature is 25℃, the water temperature compensation coefficient is 0.2; when the bubble data is 10 / minute, the bubble compensation coefficient is 0.1. For example, the original measurement value of the turbidity sensor is 11.5 NTU. The formula of the turbidity compensation algorithm is: compensated turbidity = original turbidity × (1 + water temperature compensation coefficient + bubble compensation coefficient). Substituting the values, compensated turbidity = 3.8 × (1 + 0.2 + 0.1) = 5 NTU. For example, the normal turbidity range is set to 0 to 10 NTU. If the compensated turbidity measurement value of 5 NTU is within the normal range, it is directly output as the final turbidity measurement result. For example, the turbidity measurement result 5 NTU is transmitted to the data processing module for generating water quality reports and controlling water treatment equipment.

[0039] S104: If the temperature parameter or the bubble parameter exceeds the preset threshold range, it is judged as an abnormal situation, triggering an alarm mechanism to prompt the user to check and handle it.

[0040] According to the preset temperature parameter and bubble parameter threshold range, the naive Bayes classification algorithm is used to monitor and analyze the temperature and bubble parameters in real time. The historical data of temperature and bubble parameters are modeled by Gaussian mixture model to obtain the normal distribution model of temperature and bubble parameters. The real-time collected temperature and bubble parameters are compared with the normal distribution model to calculate the likelihood probability of temperature and bubble parameters. If the likelihood probability of temperature parameter or bubble parameter is lower than the preset threshold, it is judged as an abnormal situation. When the temperature parameter or bubble parameter is detected to be abnormal, the hidden Markov model is triggered to predict the state transition of the abnormal situation. According to the duration and severity of the abnormal situation predicted by the hidden Markov model, the decision tree algorithm is used to generate the corresponding alarm strategy. If the predicted abnormal situation lasts for a long time or has a high severity, a high-level alarm is triggered to prompt the user to check and handle it immediately; if the predicted abnormal situation lasts for a short time or has a low severity, a low-level alarm is triggered to prompt the user to pay attention. At the same time as the alarm, the support vector machine algorithm is used to classify the type of abnormal situation to determine whether the abnormal situation is caused by abnormal temperature parameters or abnormal bubble parameters. According to the type of abnormal situation, the decision tree algorithm is used to generate corresponding processing suggestions to guide users to conduct targeted inspections and processing. During the user's inspection and processing, the Kalman filter algorithm is used to track and predict the temperature and bubble parameters in real time to evaluate the effectiveness of the user's processing measures. If the user's processing measures are effective, the abnormal situation is resolved and the alarm is cancelled; if the user's processing measures are ineffective, the alarm will continue to be triggered until the abnormal situation is resolved.

[0041] Specifically, for example, the normal temperature parameter threshold range is preset to be 20°C to 30°C, and the bubble parameter threshold range is 0 to 5. The system collects a set of temperature parameters of 25°C and a bubble parameter of 3. The likelihood probability of this set of data calculated by the Gaussian mixture model is 95, which is higher than the preset threshold of 9 and is judged to be normal. For another example, the system collects a set of temperature parameters of 35°C and a bubble parameter of 8. The likelihood probability of this set of data calculated by the Gaussian mixture model is 2, which is lower than the preset threshold of 9 and is judged to be abnormal. The hidden Markov model is triggered for prediction, and it is obtained that the abnormal situation may last for 5 minutes and the severity is level 3. According to the alarm strategy generated by the decision tree algorithm, a high-level alarm needs to be triggered for severity levels above level 3, prompting the user to check and handle it immediately. At the same time, the support vector machine algorithm determines that the abnormality is caused by the bubble parameter and generates a processing suggestion "Check whether the pipeline is leaking". In the process of the user handling the abnormality, the Kalman filter algorithm predicts that the temperature parameter will fall back to 33°C and the bubble parameter will drop to 6 in 1 minute. After 5 minutes, the user completes the inspection and finds that the pipeline joint is loose and causes leakage, so the joint is re-fixed. After 10 minutes, the system collected data showing that the temperature parameter was 29°C and the bubble parameter was 4. The abnormal situation was resolved and the alarm was cancelled.

[0042] S105: Compare the compensated turbidity measurement value with the preset pollutant concentration threshold to determine whether the contamination degree of the cleaning fluid exceeds the standard.

[0043] The turbidity value measured by the turbidity sensor is obtained and used as the turbidity measurement value. According to the turbidity measurement value and the preset turbidity compensation value, the compensated turbidity measurement value is calculated by a weighted average algorithm. The compensated turbidity measurement value is compared with the preset pollutant concentration threshold to determine whether the degree of contamination of the cleaning liquid exceeds the standard. If the compensated turbidity measurement value is greater than or equal to the preset pollutant concentration threshold, it is determined that the degree of contamination of the cleaning liquid exceeds the standard, and the cleaning liquid needs to be replaced or purified; if the compensated turbidity measurement value is less than the preset pollutant concentration threshold, it is determined that the degree of contamination of the cleaning liquid does not exceed the standard, and the current cleaning liquid can continue to be used for cleaning operations. After determining whether the degree of contamination of the cleaning liquid exceeds the standard, according to the judgment result, the control system issues corresponding control instructions to control the cleaning device to perform corresponding operations, such as continuing to use the current cleaning liquid for cleaning, replacing new cleaning liquid, or purifying the current cleaning liquid, so as to ensure the effect and quality of the cleaning operation.

[0044] Specifically, the turbidity value measured by the turbidity sensor is obtained and used as the turbidity measurement value. For example, the turbidity value measured by the turbidity sensor is 80 NTU. According to the turbidity measurement value and the preset turbidity compensation value, the compensated turbidity measurement value is calculated by a weighted average algorithm. For example, if the turbidity measurement value is 80 NTU, the preset turbidity compensation value is 5 NTU, the weight assigned to the turbidity measurement value is 8, and the weight assigned to the turbidity compensation value is 2, then the compensated turbidity measurement value = 80 × 8 +5 × 2 = 65 NTU. Compare the compensated turbidity measurement value with the preset pollutant concentration threshold to determine whether the degree of contamination of the cleaning solution exceeds the standard. For example, if the preset pollutant concentration threshold is 60 NTU, and the compensated turbidity measurement value is 65 NTU, then 65 NTU > 60 NTU, and it is determined that the degree of contamination of the cleaning solution exceeds the standard, and the cleaning solution needs to be replaced or purified. If the compensated turbidity measurement value is less than the preset pollutant concentration threshold, it is determined that the degree of contamination of the cleaning solution does not exceed the standard, and the current cleaning solution can continue to be used for cleaning operations. For example, if the turbidity measurement value after compensation is 58 NTU, which is less than the threshold value of 60 NTU, the pollution degree does not exceed the standard, and the current cleaning fluid can continue to be used. After determining whether the pollution degree of the cleaning fluid exceeds the standard, the control system issues corresponding control instructions based on the judgment result to control the cleaning device to perform corresponding operations. For example, if the judgment result is that the cleaning fluid pollution exceeds the standard, the control system issues an instruction to control the cleaning device to replace the new cleaning fluid; if it is determined that the pollution does not exceed the standard, the control system issues an instruction to control the cleaning device to continue using the current cleaning fluid for cleaning operations.

[0045] S106. If the pollutant concentration exceeds the threshold, a reminder to replace the cleaning fluid is triggered, notifying the user to replace the cleaning fluid in time to ensure the cleaning effect.

[0046] According to the pollutant concentration data collected by the sensor in real time, the Kalman filter algorithm is used to filter the data to obtain the pollutant concentration value after filtering. The pollutant concentration value after filtering is compared with the preset pollutant concentration threshold, and the logistic regression model is used to determine whether the pollutant concentration exceeds the threshold. If the pollutant concentration exceeds the threshold, the cleaning fluid replacement prompt is triggered, and the user is notified to replace the cleaning fluid in time through SMS or application push. At the same time, the pollutant concentration exceeding the standard event is recorded in the database, and the association rule mining algorithm is used to analyze the frequency and time law of the pollutant concentration exceeding the standard to obtain the pattern of the pollutant concentration exceeding the standard. According to the pattern of the pollutant concentration exceeding the standard, the decision tree model is used to generate the best time point recommendation for the replacement of the cleaning fluid, prompting the user to replace the cleaning fluid at the best time point to ensure the cleaning effect. After the user replaces the cleaning fluid, the Bayesian network model is used to evaluate the impact of the cleaning fluid replacement on the pollutant concentration, and determine whether the cleaning fluid replacement effectively reduces the pollutant concentration. If the cleaning fluid replacement fails to effectively reduce the pollutant concentration, the support vector machine model is used to analyze other factors affecting the pollutant concentration, generate suggestions for optimizing the cleaning process, and guide the user to further optimize the cleaning operation to ensure the best cleaning effect.

[0047] Specifically, for example, the sensor collects a set of pollutant concentration data of {12, 15, 8, 13, 11} mg / L. After being processed by the Kalman filter algorithm, the filtered pollutant concentration value is 13 mg / L. Assuming that the preset pollutant concentration threshold is 10 mg / L, the logistic regression model determines that the pollutant concentration exceeds the threshold, triggering the cleaning fluid replacement prompt. The pollutant concentration exceeding the standard event is recorded in the database. The database has accumulated the most recent 100 exceeding events, of which the number of exceeding standards from Monday to Friday is {15, 18, 13, 16, 20}. The association rule mining algorithm analysis found that the pollutant concentration exceeded the standard most frequently on Thursday and Friday, with the probability of exceeding the standard being 8 and 85, respectively. According to the pattern of pollutant concentration exceeding the standard, the decision tree model generates the best time to replace the cleaning fluid as Wednesday. After the user replaced the cleaning fluid, the Bayesian network model was used to evaluate and found that the mean pollutant concentration dropped from 13 mg / L to 1 mg / L, indicating that the replacement of the cleaning fluid effectively reduced the pollutant concentration. Assuming that after optimizing the cleaning process, the mean of the pollutant concentration is 5 mg / L and the variance is 8, the support vector machine model is used to analyze other factors affecting the pollutant concentration. It is found that for every 1 degree Celsius increase in temperature, the pollutant concentration will increase by 2 mg / L. The optimization suggestion is to control the ambient temperature of the cleaning process below 25 degrees Celsius.

[0048] S107. Continuously monitor the turbidity change trend over a period of time, determine the aging degree of the cleaning fluid through data analysis algorithms, and estimate its remaining usable time.

[0049] The turbidity monitoring data of the cleaning fluid over a period of time is obtained, and the least square method is used to fit the function curve of turbidity change over time. According to the slope change rate of the fitted function curve, the changing trend of the cleaning fluid aging speed is judged. When the slope change rate exceeds the preset threshold, it is considered that the cleaning fluid begins to age rapidly. The turbidity data of the historical cleaning fluid is predicted by the exponential smoothing method to obtain the predicted value of turbidity in the future period. If the predicted turbidity value exceeds the set replacement threshold, the remaining usable time of the cleaning fluid is estimated. According to the turbidity change rate and the used time of the cleaning fluid, the association rule mining algorithm is used to mine the association rules between the aging degree of the cleaning fluid and the used time and the turbidity change rate from the historical data. The aging degree of the current cleaning fluid is judged by the association rules. Combined with the changing trend of the cleaning fluid aging speed, the predicted value of turbidity and the judgment result of the aging degree, the fuzzy comprehensive evaluation method is used to evaluate the comprehensive aging state of the cleaning fluid and obtain the current aging degree level of the cleaning fluid. According to the aging degree level of the cleaning fluid and the predicted value of turbidity, the decision tree algorithm is used to generate a decision tree for the replacement time of the cleaning fluid. The remaining usable time of the cleaning fluid is estimated through the decision tree. The estimated remaining usable time is analyzed for confidence. The Monte Carlo simulation method is used to estimate the probability distribution of the remaining usable time through multiple random sampling simulations to obtain the confidence level. Based on the confidence level, the credibility of the estimated remaining usable time is determined.

[0050] Specifically, the least squares method is used to fit the function curve of turbidity changing with time: for example, the turbidity monitoring data of a cleaning fluid during use is fitted, and the functional relationship between turbidity y and use time x (unit: day) is obtained: y = 5x^2 + 2x + 10, and the correlation coefficient R^2 = 95. The changing trend of the cleaning fluid aging speed is judged according to the slope change rate of the fitting function curve: for example, the slope of the fitting curve is calculated to be 32, 42, and 54 when used for 30 days, 40 days, and 50 days, respectively. The slope change rate gradually increases and exceeds the preset threshold of 2. It is believed that the cleaning fluid begins to age rapidly after 40 days of use. The exponential smoothing method is used to predict the time series of turbidity data: for example, the smoothing coefficient α is set to 6, and the turbidity value of the next 10 days is predicted based on the turbidity data of the previous 50 days. The turbidity prediction value on the 60th day is 85, which exceeds the set replacement threshold of 80. It is estimated that the cleaning fluid can be used for 5 days. The association rule mining algorithm is used to determine the aging degree of the cleaning fluid: For example, by analyzing historical data, the rule is mined: the cleaning fluid that has been used for more than 40 days and the turbidity change rate is greater than 5 has a "serious" aging degree, with a support degree of 85% and a confidence level of 90%. The fuzzy comprehensive evaluation method is used to evaluate the aging state of the cleaning fluid: For example, by comprehensively considering the aging speed, turbidity prediction value and aging degree of the cleaning fluid, by establishing a fuzzy evaluation matrix, the current aging degree membership vector of the cleaning fluid is calculated to be (2,5,3), which belongs to the "moderate aging" level. The decision tree algorithm is used to estimate the remaining usable time of the cleaning fluid: For example, according to the aging degree level of the cleaning fluid and the turbidity prediction value, a decision tree is generated. When the aging degree of the cleaning fluid is "moderate" and the turbidity prediction value is 85, the decision tree outputs the result of "remaining 4 days of use". The Monte Carlo simulation method is used for confidence analysis: for example, through 1000 random sampling simulations, it is estimated that the probability that the remaining usable time of the cleaning fluid is 3 to 5 days is 90%, that is, the confidence level of the estimated result is.

[0051] S108, recording and storing turbidity monitoring data, compensation parameters, and abnormal situation information to form a historical database for cleaning fluid quality monitoring.

[0052] The Kalman filter algorithm is used to filter the turbidity monitoring data to obtain the filtered turbidity data. According to the filtered turbidity data, combined with the compensation parameters, the compensated turbidity data is obtained by the least squares fitting algorithm. The compensated turbidity data is compared with the preset turbidity threshold to determine whether there is an abnormality. If the turbidity data exceeds the threshold range, the data is marked as abnormal data. Other relevant parameters for cleaning fluid quality monitoring, such as cleaning fluid temperature, pH value, and conductivity, are obtained, and these parameters are recorded together with the compensated turbidity data. The recorded turbidity data, compensation parameters, abnormal situation marks, and other relevant parameters are stored in a relational database to form a historical database for cleaning fluid quality monitoring. The data in the historical database are statistically analyzed, the principal component analysis method is used to extract key features, and a cleaning fluid quality prediction model is established through the support vector machine algorithm. The newly acquired cleaning fluid quality monitoring data is input into the prediction model to obtain the prediction results of the cleaning fluid quality. If the prediction results show that there may be problems with the cleaning fluid quality, an early warning message is generated. According to the early warning information and the fault diagnosis rules, the fault diagnosis is performed through the rule-based reasoning algorithm to obtain possible fault causes and solutions. The fault diagnosis results and warning information are recorded in the historical database, and the relevant information is fed back to the operator so that the corresponding maintenance or adjustment measures can be taken in time. The historical database is regularly analyzed for data mining, and the association rule mining algorithm is used to find the correlation between the cleaning fluid quality and various parameters, and the cleaning fluid quality monitoring and control strategy is optimized accordingly.

[0053] Specifically, for example, the Kalman filter algorithm is used to filter the turbidity monitoring data. The original turbidity data is 2, 5, 1, 6, 3, and the filtered turbidity data is 3, 4, 2, 5, 4. According to the filtered turbidity data, combined with the compensation parameter 0.2, the compensated turbidity data is obtained by the least squares fitting algorithm, such as 32, 42, 22, 52, 42. The compensated turbidity data is compared with the preset turbidity thresholds 0 and 6 to determine whether there is an abnormality. If the turbidity data exceeds the threshold range, such as 7, the data is marked as abnormal data. Other relevant parameters for cleaning fluid quality monitoring are obtained, such as cleaning fluid temperature 45°C, pH value 2, and conductivity 8 ms / cm, and these parameters are recorded together with the compensated turbidity data. The data in the historical database are statistically analyzed, and the principal component analysis method is used to extract key features. For example, the characteristic vectors of turbidity, temperature, and pH value are 6, 3, and 1, respectively, and a cleaning fluid quality prediction model is established through the support vector machine algorithm. The newly acquired cleaning fluid quality monitoring data, such as turbidity 8, temperature 48°C, and pH value 5, are input into the prediction model to obtain the prediction results of the cleaning fluid quality, such as the predicted quality score of 85. If the prediction results show that there may be problems with the cleaning fluid quality, such as the predicted quality score is lower than 8, an early warning message is generated. According to the early warning information, combined with the fault diagnosis rules, the fault diagnosis is performed through the rule-based reasoning algorithm. For example, it is inferred that the possible cause of the fault is that the cleaning fluid temperature is too high, and the solution is to reduce the cleaning fluid temperature to 40-45°C. Data mining analysis is performed on the historical database regularly, and the association rule mining algorithm is used to find the correlation between the cleaning fluid quality and various parameters. For example, when the temperature is higher than 50°C and the pH value is lower than 5, the cleaning fluid quality score is usually lower than 7. Based on this, the cleaning fluid quality monitoring and control strategy is optimized to control the temperature at 40-45°C and the pH value at.

[0054] Although the present invention has been described in detail above by general description and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements can be made to the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all belong to the scope of protection claimed by the present invention.

Claims

1. A method for monitoring and early warning of ultrasonic cleaning machine operation, It is characterized in that The method comprises: Obtain the real-time temperature data and bubble data of the cleaning fluid of the ultrasonic cleaning machine, and use them as input parameters of the turbidity compensation algorithm; obtain the temperature data of the cleaning fluid in real time by installing a temperature sensor in the cleaning fluid of the ultrasonic cleaning machine; According to the acquired temperature data, the functional relationship between temperature and time is fitted by the least square method to obtain the curve equation of temperature changing with time; By installing a bubble sensor in the cleaning fluid of the ultrasonic cleaning machine, the bubble data in the cleaning fluid can be obtained in real time; According to the acquired bubble data, the Kalman filter algorithm is used to filter the bubble data to obtain a smooth bubble data curve; The temperature data and the bubble data are used as input parameters of the turbidity compensation algorithm; The mathematical model between temperature, bubbles and turbidity was established by using the multivariate linear regression algorithm to obtain the turbidity compensation equation. According to the turbidity compensation equation, the turbidity of the ultrasonic cleaning machine cleaning fluid is compensated in real time to obtain the compensated turbidity value; The compensated turbidity value is compared with the preset turbidity threshold value. If the compensated turbidity value exceeds the threshold value, it is determined that the cleaning fluid needs to be replaced; If the turbidity value after compensation does not exceed the threshold, it is judged that the cleaning fluid can continue to be used; According to the judgment results, the cleaning liquid of the ultrasonic cleaning machine is automatically controlled to change, ensuring that the turbidity of the cleaning liquid is always within a reasonable range to ensure the cleaning effect; The turbidity value of the cleaning liquid is measured in real time by the turbidity sensor to obtain the original turbidity measurement data; According to the acquired temperature parameters and bubble parameters, the original turbidity measurement value is corrected and calculated by using the turbidity compensation algorithm to obtain the compensated turbidity measurement value; the water temperature data measured by the temperature sensor and the bubble data measured by the bubble sensor are obtained; According to the acquired water temperature data, the water temperature compensation coefficient is obtained by table lookup method; If the water temperature is higher than the set threshold, the water temperature compensation coefficient is set to a positive value; If the water temperature is lower than the set threshold, the water temperature compensation coefficient is set to a negative value; According to the obtained bubble data, the bubble compensation coefficient is obtained by table lookup method; If the bubble data is greater than the set threshold, the bubble compensation coefficient is set to a positive value; If the bubble data is less than the set threshold, the bubble compensation coefficient is set to a negative value; Get the raw measurement value of the turbidity sensor; Substituting the obtained water temperature compensation coefficient and bubble compensation coefficient into the formula of the turbidity compensation algorithm, the original turbidity measurement value is corrected and calculated to obtain the compensated turbidity measurement value; The turbidity compensation algorithm adopts a weighted average algorithm, which takes the water temperature compensation coefficient and the bubble compensation coefficient as weights and performs weighted average with the original turbidity measurement value to obtain the final compensated turbidity measurement value; Determine whether the turbidity measurement value after compensation is within the normal range; If the compensated turbidity measurement value exceeds the normal range, it is limited to the upper or lower limit of the normal range to obtain the final turbidity measurement result; If the compensated turbidity measurement value is within the normal range, it is directly output as the final turbidity measurement result; The final turbidity measurement results are transmitted to the subsequent data processing module for further analysis and control; If the temperature parameter or bubble parameter exceeds the preset threshold range, it is judged as an abnormal situation, triggering the alarm mechanism to prompt the user to check and handle it; Compare the compensated turbidity measurement value with the preset pollutant concentration threshold to determine whether the degree of contamination of the cleaning fluid exceeds the standard; If the pollutant concentration exceeds the threshold, the cleaning fluid replacement reminder will be triggered to notify the user to replace the cleaning fluid in time to ensure the cleaning effect; Continuously monitor the turbidity change trend over a period of time, determine the aging degree of the cleaning fluid through data analysis algorithms, and estimate its remaining usable time; The turbidity monitoring data, compensation parameters and abnormal conditions are recorded and stored to form a historical database for cleaning fluid quality monitoring.

2. The method according to claim 1, It is characterized in that The turbidity value of the cleaning liquid is measured in real time by the turbidity sensor to obtain the original turbidity measurement data, including: The median filter algorithm is used to filter the original turbidity measurement data, remove abnormal values, and obtain the filtered turbidity measurement data; The turbidity measurement data after filtering were fitted by the least square method to obtain the fitting curve of turbidity changing with time; Calculate the turbidity change rate according to the fitting curve, and determine whether the turbidity change rate exceeds a preset threshold. If it exceeds the threshold, it is determined that the cleaning fluid is contaminated and needs to be replaced. If the threshold is not exceeded, the cleaning fluid is judged to be uncontaminated and can continue to be used; The smoothed value of turbidity is calculated by exponentially weighted moving average algorithm to reduce the fluctuation of turbidity measurement data and obtain the smoothed turbidity value; Based on the smoothed turbidity value, the control limit algorithm is used to determine whether the turbidity exceeds the control limit. If it exceeds the control limit, a cleaning fluid contamination alarm is issued; If the control limit is not exceeded, the cleaning solution is considered to be uncontaminated and the turbidity changes continue to be monitored; The changing trend of turbidity in the future is predicted by the autoregressive moving average model to obtain the predicted value of turbidity; According to the predicted turbidity value and the current turbidity value, a linear discriminant analysis algorithm is used to determine whether the cleaning fluid needs to be replaced. If it needs to be replaced, an instruction to replace the cleaning fluid is issued; If replacement is not necessary, continue to monitor turbidity changes.

3. The method according to claim 1, It is characterized in that If the temperature parameter or the bubble parameter exceeds the preset threshold range, it is judged as an abnormal situation, triggering an alarm mechanism to prompt the user to check and handle it, including: According to the preset temperature parameters and bubble parameter threshold ranges, the naive Bayes classification algorithm is used to monitor and analyze the temperature and bubble parameters in real time; The historical data of temperature and bubble parameters are modeled by Gaussian mixture model to obtain the normal distribution model of temperature and bubble parameters; Compare the real-time collected temperature and bubble parameters with the normal distribution model, and calculate the likelihood probability of the temperature and bubble parameters; If the likelihood probability of the temperature parameter or the bubble parameter is lower than the preset threshold, it is judged as an abnormal situation; When abnormal temperature parameters or bubble parameters are detected, the hidden Markov model is triggered to predict the state transition of abnormal conditions; According to the duration and severity of abnormal situations predicted by the hidden Markov model, the decision tree algorithm is used to generate corresponding alarm strategies; If the predicted abnormal situation lasts for a long time or is of high severity, a high-level alarm will be triggered, prompting the user to check and handle it immediately; If the predicted abnormal situation lasts for a short time or is of low severity, a low-level alarm will be triggered to prompt the user to pay attention; While giving an alarm, the support vector machine algorithm is used to classify the type of abnormal situation and determine whether the abnormal situation is caused by abnormal temperature parameters or abnormal bubble parameters; According to the type of abnormal situation, the decision tree algorithm is used to generate corresponding processing suggestions to guide users to conduct targeted inspections and processing; During the user's inspection and treatment process, the Kalman filter algorithm is used to track and predict the temperature and bubble parameters in real time to evaluate the effectiveness of the user's treatment measures; If the user's handling measures are effective, the abnormal situation is resolved and the alarm is cancelled; If the user's handling measures are ineffective, the alarm will continue to be triggered until the abnormal situation is resolved.

4. The method according to claim 1, It is characterized in that The method of comparing the compensated turbidity measurement value with a preset pollutant concentration threshold value to determine whether the degree of contamination of the cleaning fluid exceeds the standard includes: Obtaining a turbidity value measured by a turbidity sensor and using it as a turbidity measurement value; According to the turbidity measurement value and the preset turbidity compensation value, a weighted average algorithm is used to calculate the compensated turbidity measurement value; Compare the compensated turbidity measurement value with the preset pollutant concentration threshold to determine whether the degree of contamination of the cleaning fluid exceeds the standard; If the compensated turbidity measurement value is greater than or equal to the preset pollutant concentration threshold, it is determined that the degree of contamination of the cleaning fluid exceeds the standard and the cleaning fluid needs to be replaced or purified; If the compensated turbidity measurement value is less than the preset pollutant concentration threshold, it is judged that the degree of contamination of the cleaning fluid does not exceed the standard, and the current cleaning fluid can continue to be used for cleaning operations; After determining whether the degree of contamination of the cleaning fluid exceeds the standard, the control system issues corresponding control instructions based on the judgment result to control the cleaning device to perform corresponding operations, such as continuing to use the current cleaning fluid for cleaning and replacing it with a new cleaning fluid or purifying the current cleaning fluid to ensure the effect and quality of the cleaning operation.

5. The method according to claim 1, It is characterized in that If the pollutant concentration exceeds the threshold, a reminder to replace the cleaning fluid is triggered, notifying the user to replace the cleaning fluid in time to ensure the cleaning effect, including: According to the pollutant concentration data collected by the sensor in real time, the Kalman filter algorithm is used to filter the data to obtain the filtered pollutant concentration value; Compare the filtered pollutant concentration value with the preset pollutant concentration threshold, and determine whether the pollutant concentration exceeds the threshold through a logistic regression model; If the pollutant concentration exceeds the threshold, a reminder to replace the cleaning fluid is triggered, and the user is notified via SMS or app push to replace the cleaning fluid in time; At the same time, the pollutant concentration exceeding the standard event is recorded in the database, and the association rule mining algorithm is used to analyze the frequency and time pattern of the pollutant concentration exceeding the standard to obtain the pattern of the pollutant concentration exceeding the standard; According to the pattern of excessive pollutant concentration, the decision tree model generates the best time to replace the cleaning fluid, prompting the user to replace the cleaning fluid at the best time to ensure the cleaning effect; After the user changes the cleaning fluid, the Bayesian network model is used to evaluate the impact of the cleaning fluid change on the pollutant concentration and determine whether the cleaning fluid change effectively reduces the pollutant concentration; If the replacement of the cleaning fluid fails to effectively reduce the pollutant concentration, the support vector machine model is used to analyze other factors that affect the pollutant concentration, generate suggestions for optimizing the cleaning process, and guide the user to further optimize the cleaning operation to ensure the best cleaning effect.

6. The method according to claim 1, It is characterized in that The continuous monitoring of the turbidity change trend over a period of time, judging the aging degree of the cleaning fluid through a data analysis algorithm, and estimating the remaining usable time thereof, includes: Obtain the turbidity monitoring data of the cleaning liquid over a period of time, and use the least square method to fit the function curve of turbidity changing with time; According to the slope change rate of the fitting function curve, the changing trend of the aging speed of the cleaning fluid is determined; When the slope change rate exceeds a preset threshold, it is considered that the cleaning fluid begins to age rapidly; The exponential smoothing method is used to perform time series forecasting on the historical cleaning fluid turbidity data to obtain the predicted value of turbidity in the future period. If the predicted turbidity value exceeds the set replacement threshold, the remaining usable time of the cleaning fluid is estimated; According to the turbidity change rate and usage time of the cleaning fluid, the association rule mining algorithm is used to mine the association rules between the aging degree of the cleaning fluid and the usage time and turbidity change rate from the historical data; By using association rules, the aging degree of the current cleaning fluid is determined; Combining the changing trend of the cleaning fluid aging speed, the predicted value of turbidity and the judgment result of the aging degree, the fuzzy comprehensive evaluation method is used to evaluate the comprehensive aging state of the cleaning fluid and obtain the current aging degree of the cleaning fluid. According to the cleaning fluid aging level and turbidity prediction value, a decision tree for the cleaning fluid replacement time is generated using a decision tree algorithm; Estimate the remaining usable time of the cleaning fluid through the decision tree; Conduct confidence analysis on the estimated remaining usable time. Use the Monte Carlo simulation method to estimate the probability distribution of the remaining usable time through multiple random sampling simulations to obtain the confidence level. Based on the confidence level, determine the degree of confidence in the estimated remaining usable time.

7. The method according to claim 1, It is characterized in that The turbidity monitoring data, compensation parameters, and abnormal conditions are recorded and stored to form a historical database for cleaning fluid quality monitoring, including: The Kalman filter algorithm is used to filter the turbidity monitoring data to obtain filtered turbidity data; According to the filtered turbidity data and the compensation parameters, the compensated turbidity data is obtained by the least square fitting algorithm; Compare the compensated turbidity data with the preset turbidity threshold to determine whether there is an abnormality; If the turbidity data exceeds the threshold range, the data is marked as abnormal data; Obtain other relevant parameters for cleaning fluid quality monitoring, including cleaning fluid temperature, pH value and conductivity, and record these parameters together with the compensated turbidity data; The recorded turbidity data, compensation parameters, abnormal situation marks and other related parameters are stored in a relational database to form a historical database for cleaning fluid quality monitoring; Statistical analysis was performed on the data in the historical database, key features were extracted using principal component analysis, and a cleaning fluid quality prediction model was established using the support vector machine algorithm; Input the newly acquired cleaning fluid quality monitoring data into the prediction model to obtain the prediction result of the cleaning fluid quality; If the prediction results indicate that there may be problems with the quality of the cleaning fluid, an early warning message is generated; According to the early warning information, combined with the fault diagnosis rules, the fault diagnosis is carried out through the rule-based reasoning algorithm to obtain the possible fault causes and solutions; Record the fault diagnosis results and warning information in the historical database, and feed back the relevant information to the operator so that the corresponding maintenance or adjustment measures can be taken in time; Data mining and analysis are performed on the historical database regularly, and the association rule mining algorithm is used to discover the correlation between the quality of the cleaning fluid and various parameters, and the cleaning fluid quality monitoring and control strategy is optimized accordingly.