Visualization method and system applied to quality monitoring and statistics

Abstract

The invention relates to the technical field of quality monitoring, and discloses a visualization method and system applied to quality monitoring and statistics, and the method comprises the steps: continuously collecting the operation data of a first object; predicting the operation state of the first object based on the operation data of the first object, and generating a visual prediction result; and judging the running state of the first object by using the visual prediction result, and regulating and controlling the first object. By integrating four modules of real-time data acquisition, dynamic trend prediction and analysis, multi-dimensional interactive visualization and alarm and suggestion, comprehensive and real-time monitoring of production equipment is realized. The system can timely and accurately identify the equipment performance change trend, predict the equipment state change in advance, and effectively prevent faults. A multi-dimensional interactive visual interface improves the flexibility and efficiency of data analysis and exception troubleshooting of a user, so that the decision making process is more visual and accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of quality monitoring, and in particular to a visualization method and system applied to quality monitoring and statistics. Background Art

[0002] With the in-depth advancement of Industry 4.0 and intelligent manufacturing, enterprises have put forward higher requirements for quality monitoring and statistical analysis of the production process. Traditional quality monitoring methods mainly rely on manual regular inspections and manual data recording, which have problems such as low efficiency, poor accuracy, and lack of real-time performance. This method is difficult to detect quality problems in the production process in a timely manner, and problems are often not detected until they accumulate to a certain extent, resulting in waste of resources, fluctuations in product quality, and even serious production accidents.

[0003] In recent years, although some companies have introduced computer-aided quality monitoring systems, these systems are mostly limited to basic data collection and simple statistical analysis, and lack the ability to deeply mine data and predict trends. When faced with large amounts of rapidly changing equipment operation data, traditional systems are difficult to respond in a timely manner and cannot quickly extract valuable information from massive amounts of data. This makes it difficult for managers to gain timely insight into the changing trends of equipment performance, miss the best time for maintenance and regulation, and increase the risk of production line downtime and maintenance costs.

[0004] In addition, existing systems also have limitations in visualization, usually only providing static, single-dimensional chart displays, and lacking a multi-dimensional, interactive visualization interface. Users cannot flexibly select display dimensions and adjust timelines according to actual needs, and deeply analyze data for specific time periods or specific parameters, which reduces the efficiency and accuracy of data analysis. When abnormal situations occur, the system lacks intelligent alarm and suggestion functions, and cannot promptly notify relevant personnel and provide effective optimization strategies, which affects the timely handling and resolution of problems.

[0005] Therefore, how to build a quality monitoring and statistics system that integrates real-time data collection, dynamic trend prediction and analysis, multi-dimensional interactive visualization, and intelligent alarms and optimization suggestions has become a technical problem that needs to be solved urgently. The system should be able to collect equipment operation data in real time and accurately, quickly analyze and predict equipment performance trends, provide flexible visualization tools for users to deeply mine data, and at the same time, trigger alarms in time and provide targeted optimization suggestions when abnormalities are detected, helping enterprises improve production efficiency, reduce operating costs, and enhance market competitiveness. Summary of the invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] Therefore, the technical problem solved by the present invention is that the existing quality monitoring system is inefficient, lacks real-time monitoring and in-depth data analysis, and is unable to timely predict equipment performance trends and discover abnormal problems.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: a visualization method for quality monitoring and statistics, comprising:

[0009] continuously collecting operation data of the first object;

[0010] Predicting an operating state of the first object based on the operating data of the first object, and generating a visual prediction result;

[0011] The visual prediction result is used to determine the operating state of the first object, and the first object is regulated.

[0012] As a preferred solution of the visualization method for quality monitoring and statistics described in the present invention, the operation data of the first object includes real-time parameters and real-time operation status data generated by the first object during operation.

[0013] As a preferred solution of the visualization method for quality monitoring and statistics described in the present invention, the continuously collecting the operation data of the first object includes continuously collecting the operation data of the first object through real-time technology and improving the quality of the operation data.

[0014] As a preferred solution of the visualization method for quality monitoring and statistics described in the present invention, wherein: the predicting the operating state of the first object based on the operating data of the first object includes extracting features based on the operating data of the first object to obtain operating features of the first object;

[0015] Based on the operation characteristics of the first object, a prediction model is constructed to predict the operation status of the first object and generate a prediction result.

[0016] As a preferred solution of the visualization method for quality monitoring and statistics described in the present invention, generating a visualization prediction result includes improving the readability of the prediction result and quantifying the uncertainty of the prediction result;

[0017] Generate a visual prediction representation based on the processed prediction results and quantified results.

[0018] As a preferred solution of the visualization method for quality monitoring and statistics described in the present invention, wherein: the use of the visualization prediction result to determine the operating state of the first object includes performing anomaly detection on the visualization prediction result through a machine learning algorithm;

[0019] performing a deviation analysis on the operating data of the first object using an analysis algorithm;

[0020] The operation status of the first object is determined and regulated by combining the abnormality detection result and the deviation analysis result.

[0021] As a preferred solution of the visualization method applied to quality monitoring and statistics described in the present invention, the machine learning algorithm includes but is not limited to inputting the real-time visualization prediction results into an isolation forest model, calculating the anomaly score for each data point; and marking the data points with anomaly scores higher than a threshold as anomalies.

[0022] The analysis algorithm includes, but is not limited to, inputting the first subject's operating data into a linear regression model to calculate a predicted value of the first subject's operation;

[0023] The predicted value of the first object operation is compared with the actual value of the first object operation, the deviation value and the deviation percentage are calculated, and the deviation degree of the first object operation state is evaluated.

[0024] As a preferred solution of the visualization method for quality monitoring and statistics described in the present invention, wherein: a real-time data stream acquisition module continuously collects operation data of the first object;

[0025] A dynamic trend prediction module, which predicts the operating state of the first object based on the operating data of the first object and generates a visual prediction result;

[0026] The analysis and control module uses the visual prediction result to determine the operating state of the first object and controls the first object.

[0027] A computer device comprises: a memory and a processor; the memory stores a computer program, wherein the processor implements the steps of any one of the methods of the present invention when executing the computer program.

[0028] A computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of any one of the methods of the present invention.

[0029] Beneficial effects of the present invention: The visualization method for quality monitoring and statistics provided by the present invention realizes comprehensive and real-time monitoring of production equipment by integrating four modules: real-time data collection, dynamic trend prediction and analysis, multi-dimensional interactive visualization, and alarm and suggestion. The system can timely and accurately identify the performance change trend of equipment, predict equipment status changes in advance, and effectively prevent failures. The multi-dimensional interactive visualization interface improves the flexibility and efficiency of users in data analysis and abnormality troubleshooting, making the decision-making process more intuitive and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0031] Figure 1 An overall flow chart of a visualization method for quality monitoring and statistics provided by the first embodiment of the present invention; DETAILED DESCRIPTION

[0032] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0033] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides a visualization method for quality monitoring and statistics, including:

[0034] S1: Continuously collect operating data of a first object.

[0035] In one implementation of the present application, the first object refers to a specific device, machine, production line or system in industrial production, such as a machine tool, an engine, a sensor network, an assembly line, etc.

[0036] The operating parameters and status information of the first object are continuously and in real time collected through the real-time data stream acquisition module. The operating data of the first object is continuously collected through real-time technology, and the quality of the operating data is improved.

[0037] The operating data of the first object refers to the real-time collection of various parameters and status information generated by the monitored equipment during operation, including but not limited to sensor data, performance indicators, operating status parameters, status information, operating status, alarm information, maintenance records, environmental parameters, external environment data, working conditions and historical data.

[0038] In an optional embodiment, the first object may also be an environmental monitoring station, and the system collects environmental parameters through a real-time data stream acquisition module, including air quality indicators (such as PM2.5, PM10, SO 2 、NO 2 , O 3, CO) and meteorological data (such as temperature, humidity, wind speed, wind direction, and rainfall). Using the dynamic trend prediction and analysis module, time series models (such as ARIMA and Prophet) are used to predict environmental parameters and identify changing trends in pollutant concentrations and potential abnormal events (such as sudden pollution). The multi-dimensional interactive visualization interface provides a geographic information system (GIS) display and supports spatial hierarchical data drilling from the national, provincial, and municipal levels to specific monitoring sites. Users can customize the type and time range of pollutants of concern and view real-time and historical data. The alarm and suggestion module sets warning thresholds according to national or local environmental standards, monitors pollutant concentrations in real time, triggers warnings when the threshold is exceeded, and notifies relevant personnel and the public through channels such as SMS, email, and mobile applications, provides health protection advice and emergency measures, and supports scientific decision-making and rapid response of environmental governance departments.

[0039] In an optional embodiment, the first object may also be a water treatment facility. The system collects water quality parameters (such as pH value, dissolved oxygen, turbidity, chemical oxygen demand COD, biological oxygen demand BOD, ammonia nitrogen, total phosphorus, total nitrogen) and equipment operation parameters (such as water pump pressure, flow meter data, valve opening, agitator speed) through a real-time data stream acquisition module. Using the dynamic trend prediction and analysis module, a time series model (such as ARIMA, LSTM) is used to predict water quality and equipment performance, and identify water quality exceeding the standard risk and equipment performance decline trend. The multi-dimensional interactive visualization interface supports multi-layer data drilling from the overall plant water quality overview to specific processing units and single equipment. Users can customize chart types and color coding to monitor water quality changes and equipment operation status in real time. The alarm and suggestion module sets thresholds according to water quality safety standards and equipment operating specifications, and performs real-time monitoring. When water quality parameters exceed the threshold or equipment operation is abnormal, an alarm is triggered to notify relevant personnel, and optimization suggestions are provided based on the historical case library, such as adjusting the amount of chemical dosage, arranging equipment maintenance, etc., to ensure the safe and efficient operation of water treatment facilities and improve the water quality compliance rate and equipment service life.

[0040] Furthermore, through the real-time data stream acquisition module, the operating parameters, status information and environmental data of the equipment are continuously collected to ensure the integrity and reliability of the data, thereby providing sufficient basis for real-time monitoring, fault warning, performance evaluation and optimization decision-making of the equipment. The real-time collected data not only supports adjustment and optimization in dynamic scenarios, but also provides raw data input for anomaly detection, deviation analysis, trend prediction, etc., helping the system to promptly identify and respond to equipment deviations from normal status. In addition, the long-term accumulation of data lays the foundation for historical analysis and life evaluation of equipment performance, thereby realizing comprehensive monitoring and intelligent regulation of equipment operation status.

[0041] S2: Predicting an operating state of the first object based on the operating data of the first object, and generating a visualized prediction result.

[0042] Feature extraction is performed based on the operation data of the first object to obtain the operation feature of the first object.

[0043] Based on the operation characteristics of the first object, a prediction model is constructed to predict the operation status of the first object and generate a prediction result.

[0044] The prediction results are processed to improve readability and the uncertainty of the prediction results is quantified.

[0045] Generate a visual prediction representation based on the processed prediction results and quantified results.

[0046] Feature extraction is performed based on the operation data of the first object to obtain the operation characteristics of the first object. The operation data of the first object is preprocessed, including denoising, smoothing, normalization and other operations, and key characteristic parameters that can reflect the operation status of the equipment are extracted, including mean, variance, spectrum characteristics and trend indicators. These characteristic parameters constitute the operation characteristics of the first object and provide a basis for subsequent model construction and prediction.

[0047] Construct an ARIMA forecasting model to forecast the operating status of the first object and generate forecast results. Select an ARIMA model suitable for time series data and determine the model parameters (p, d, q) based on the operating characteristic data. Use historical operating characteristic data to train the model so that it can accurately fit the time series pattern of the data. Input the latest operating characteristic data into the model to generate the forecast value of the operating status of the first object at a future time and form a forecast result.

[0048] The forecast results are processed to improve readability and the uncertainty of the forecast results is quantified. First, the forecast results are smoothed to eliminate short-term fluctuations and highlight long-term trends; outliers are detected and processed to prevent them from affecting the overall trend judgment; data units and scales are adjusted according to user habits to make the results easier to understand. At the same time, the confidence interval of the forecast results is calculated to quantify the uncertainty of the forecast. The error indicators of the evaluation model include mean square error (MSE) and mean absolute error (MAE), providing an objective evaluation of the model accuracy.

[0049] Generate a visual forecast representation based on the processed forecast results and quantitative results. Draw the processed forecast trend curve and add confidence interval bands on the curve to intuitively display the uncertainty range of the forecast. Through a multi-dimensional interactive visualization interface, the forecast results are updated in real time, allowing users to view forecast data for different time periods. Provide interactive functions, users can zoom, drag, and click on charts to deeply analyze data; mark abnormal points in the forecast, add detailed information and explanations, and assist users in understanding and decision-making.

[0050] Furthermore, by conducting in-depth analysis and feature extraction of the operating data of the first object, an effective prediction model is constructed to foresee the future operating status of the equipment in advance. This process converts historical data into valuable operating features, enabling the system to identify potential trends and patterns, thereby accurately predicting the operating status of the equipment. By generating visual prediction results, the health status and performance change trends of the equipment can be intuitively displayed, providing timely decision-making basis for operation and maintenance personnel. This step not only helps prevent equipment failures, but also optimizes the operating efficiency of the equipment, reduces maintenance costs, and improves the stability and reliability of the production line. Through forward-looking predictions of the equipment's operating status, intelligent scheduling and risk warning can be achieved, providing guarantees for the safe and efficient operation of the entire system.

[0051] S3: Using the visualized prediction result to determine the operating state of the first object, and regulating the first object.

[0052] Anomalies are detected on the visualized prediction results through a machine learning algorithm; deviation analysis is performed on the operating data of the first object using an analysis algorithm; the operating state of the first object is determined and regulated by combining the anomaly detection results and the deviation analysis results.

[0053] In one implementation of the present application, the machine learning algorithm is an isolation forest model. The real-time visual prediction results are input into the isolation forest model, and an anomaly score is calculated for each data point; this score reflects the "degree of isolation" of the data point in the data space. That is, if a data point is significantly different from other points, the isolation forest model will give a higher anomaly score.

[0054] Based on the anomaly score calculated by the Isolation Forest Model, data points with scores higher than a preset threshold are marked as outliers. This threshold setting can be adjusted according to specific business needs. Generally, the higher the score, the more likely the data point is an outlier, that is, it exhibits characteristics that are significantly different from the normal data distribution.

[0055] Data points marked as abnormal will attract the attention of the system. Further abnormal pattern recognition and deviation analysis can be performed, and alerts can be generated based on these abnormal information for managers or control systems to adjust or optimize.

[0056] In an optional embodiment of the present application, the machine learning algorithm can use a support vector machine (SVM). Using the real-time visual prediction results as input data, first use the historical normal operation data to train the SVM model, learn the distribution characteristics of normal data and build a decision boundary. In real-time operation, the SVM model classifies each newly input predicted data point to determine whether it is within the decision boundary. If the data point falls outside the decision boundary, it means that its characteristics are significantly different from normal data and are marked as abnormal points. The setting of the threshold can be adjusted according to the classification interval of the model, and points far from the decision boundary are usually selected as abnormal data. Data points marked as abnormal will attract the attention of the system, and abnormal pattern recognition and deviation analysis can be performed later, and alarms are generated based on these abnormal information for management personnel or control systems to adjust or optimize.

[0057] In an optional embodiment of the present application, the machine learning algorithm may be a K-means clustering algorithm. The real-time visual prediction results are input into the K-means algorithm, the data is clustered and analyzed, and the data points are divided into multiple clusters, each cluster representing a similar normal operating state. For each predicted data point in the real-time data, the distance from the center point of each cluster is calculated. If the distance between the data point and the center of its nearest cluster is greater than a preset threshold, the data point is regarded as an abnormal point, which may indicate an abnormality or change in the operating state of the equipment. The setting of the threshold can be adjusted according to the average distance and standard deviation of the data within the cluster. Data points marked as abnormal will attract the attention of the system, and further abnormal analysis and deviation assessment can be performed later, and alarms are generated based on the results to guide managers or control systems to take corresponding regulatory measures.

[0058] In one implementation of the present application, the analysis algorithm is a linear regression model, which takes the historical operation data of the first object as input, establishes a linear regression model, and generates a predicted value of the operation state of the first object. The predicted value is calculated based on the input feature data and represents the normal operation state that the device should achieve under specific conditions.

[0059] The actual operation data of the first object is compared with the predicted value obtained by the linear regression model, and the deviation value between the two is calculated. This deviation value reflects the difference between the actual operation and the predicted value. At the same time, the deviation percentage is calculated, that is, the percentage of the difference between the actual value and the predicted value to the predicted value, to quantify the degree of deviation of the operating state. The deviation percentage can help quickly identify situations that deviate from the normal operating range and evaluate whether the operating state of the equipment is abnormal or the performance is degraded.

[0060] In an optional embodiment of the present application, the analysis algorithm can also be a decision tree regression model method, and the historical operating data of the first object is input into the model for training. Decision tree regression can capture nonlinearities and complex patterns in the data by recursively dividing the data set and establishing a mapping relationship between features and target variables. After the training is completed, the model can generate a predicted value of the operating state of the first object based on the input feature data, representing the normal operating level that the equipment should achieve under specific conditions. Then, the actual operating data is compared with the predicted value, the deviation value and deviation percentage are calculated, and the degree of deviation of the equipment operating state is evaluated. Deviation analysis helps to quickly identify abnormalities or declines in equipment performance, making it easier to take optimization measures in a timely manner.

[0061] In an optional embodiment of the present application, the analysis algorithm can also be a support vector regression (SVR) model. SVR is an extension of the support vector machine (SVM), is suitable for regression problems, and can handle linear and nonlinear relationships. In implementation, the historical operating data of the first object is first input into the SVR model for training to learn the relationship between the feature data and the operating state. The trained SVR model can predict the normal operating state of the first object based on the input feature data. Then, the actual operating data is compared with the predicted value of the SVR model, the deviation value and the deviation percentage are calculated, and the degree of deviation of the operating state is quantified. The SVR model is insensitive to outliers, has good generalization ability, and is suitable for performance evaluation and anomaly detection of complex systems.

[0062] When an abnormal score or deviation is detected that exceeds the preset threshold, the alarm mechanism will be triggered to notify relevant personnel of the abnormal information through the system interface, SMS and email. The alarm content includes the type, location, degree and possible cause of the abnormality, and provides solution suggestions to assist operators in making quick decisions. At the same time, based on the prediction model, the system automatically adjusts the equipment operating parameters such as temperature, pressure, load or flow to restore the normal operation of the equipment. After the adjustment, the adjustment effect is evaluated in real time through the feedback mechanism. If the adjustment fails to effectively solve the problem, the system will prompt further manual intervention to ensure the safe and efficient operation of the equipment.

[0063] Furthermore, by comprehensively using machine learning and analysis algorithms, the ability to intelligently judge and dynamically control the operating status of the first object is improved. Anomaly detection and deviation analysis help to promptly detect anomalies or deviations in equipment operation, ensuring that the system can make warnings or adjustments before failures occur. By combining the results of these two analyses, the system can accurately judge the health status, performance deviations, and potential risks of the equipment, and thus take appropriate control measures, such as adjusting operating parameters or triggering alarms, to prevent equipment damage or performance degradation. This process not only improves the operational safety of the equipment, but also optimizes production efficiency and avoids downtime and economic losses caused by equipment failure or abnormal operation. Through continuous intelligent monitoring and control, the equipment can achieve more stable, reliable, and efficient operation.

[0064] Embodiment 2 is an embodiment of the present invention, which provides a visualization system for quality monitoring and statistics, including:

[0065] The real-time data stream acquisition module continuously and in real time collects equipment operating parameters and status information to obtain equipment operating data.

[0066] Equipment operation data includes sensor data, performance indicators, operating status parameters, status information, operating status, alarm information, maintenance records, environmental parameters, external environment data, working conditions and historical data.

[0067] The dynamic trend prediction and analysis module performs real-time analysis on the continuously collected equipment operation data, identifies and builds a trend model of equipment performance and status changes over time, and dynamically generates a prediction curve reflecting the trend direction through the trend model.

[0068] Feature extraction is performed based on the operation data of the first object to obtain the operation characteristics of the first object. The operation data of the first object is preprocessed, including denoising, smoothing, normalization and other operations, and key characteristic parameters that can reflect the operation status of the equipment are extracted, including mean, variance, spectrum characteristics and trend indicators. These characteristic parameters constitute the operation characteristics of the first object and provide a basis for subsequent model construction and prediction.

[0069] Construct an ARIMA forecasting model to forecast the operating status of the first object and generate forecast results. Select an ARIMA model suitable for time series data and determine the model parameters (p, d, q) based on the operating characteristic data. Use historical operating characteristic data to train the model so that it can accurately fit the time series pattern of the data. Input the latest operating characteristic data into the model to generate the forecast value of the operating status of the first object at a future time and form a forecast result.

[0070] The forecast results are processed to improve readability and the uncertainty of the forecast results is quantified. First, the forecast results are smoothed to eliminate short-term fluctuations and highlight long-term trends; outliers are detected and processed to prevent them from affecting the overall trend judgment; data units and scales are adjusted according to user habits to make the results easier to understand. At the same time, the confidence interval of the forecast results is calculated to quantify the uncertainty of the forecast. The error indicators of the evaluation model include mean square error (MSE) and mean absolute error (MAE), providing an objective evaluation of the model accuracy.

[0071] Launch a multi-dimensional interactive visualization interface to dynamically display charts including time series analysis, key indicator comparison, and anomaly point annotations based on the output of the dynamic trend prediction and analysis module. Users can freely select display dimensions, adjust the timeline, and drill down to detailed information for a specific time period.

[0072] The analysis results are monitored through the Alert and Suggestion module. When it is detected that the trend deviates from the preset standard or an abnormal pattern occurs, an alert notification is triggered and optimization suggestions are made based on historical data.

[0073] (2) The real-time data stream acquisition module includes:

[0074] The data interface management submodule is responsible for connecting with different device protocols for universal data collection.

[0075] The real-time transport protocol submodule adopts the MQTT low-latency protocol to ensure the continuity and real-time performance of data flow.

[0076] The data quality control submodule is used to perform data integrity verification and packet loss processing.

[0077] The high-concurrency processing submodule includes a data cache and queue mechanism, which is used to process the high-concurrency data stream generated when multiple devices are connected at the same time.

[0078] (3) The dynamic trend prediction includes:

[0079] The feature extraction and selection submodule is used to identify and select equipment operating parameters.

[0080] The time series model training submodule builds a prediction model based on the identified and selected equipment operating parameters.

[0081] The online learning and updating submodule is used for the prediction model to continuously optimize itself based on the newly identified and selected data.

[0082] The prediction result evaluation submodule uses the cross-validation method to evaluate the model prediction performance.

[0083] (4) The analysis module includes:

[0084] The anomaly detection submodule is used to identify anomalies in the equipment operating parameters.

[0085] The trend analysis submodule is used to analyze the change patterns of equipment operating parameters over time and determine the trend type.

[0086] The key data association analysis submodule is used to analyze the correlation between various data and reveal potential causal relationships.

[0087] The performance degradation prediction submodule is used to predict the possible future performance degradation trend of the device based on historical data.

[0088] Anomalies are detected on the visualized prediction results through the isolation forest model; deviation analysis is performed on the operating data of the first object using the analysis algorithm; the operating state of the first object is determined and regulated by combining the anomaly detection results and the deviation analysis results.

[0089] The real-time visual prediction results are input into the isolation forest model to calculate an anomaly score for each data point; this score reflects the "degree of isolation" of the data point in the data space. That is, if a data point is very different from other points, the isolation forest model will give a higher anomaly score.

[0090] Based on the anomaly score calculated by the Isolation Forest Model, data points with scores higher than a preset threshold are marked as outliers. This threshold setting can be adjusted according to specific business needs. Generally, the higher the score, the more likely the data point is an outlier, that is, it exhibits characteristics that are significantly different from the normal data distribution.

[0091] Data points marked as abnormal will attract the attention of the system. Further abnormal pattern recognition and deviation analysis can be performed, and alerts can be generated based on these abnormal information for managers or control systems to adjust or optimize.

[0092] The historical operation data of the first object is used as input to establish a linear regression model to generate a predicted value of the operation state of the first object. The predicted value is calculated based on the input feature data and represents the normal operation state that the device should achieve under specific conditions.

[0093] Next, the actual operation data of the first object is compared with the predicted value obtained by the linear regression model, and the deviation value between the two is calculated. This deviation value reflects the difference between the actual operation and the predicted value. At the same time, the deviation percentage is calculated, that is, the percentage of the difference between the actual value and the predicted value to the predicted value, to quantify the degree of deviation of the operating state. The deviation percentage can help quickly identify situations that deviate from the normal operating range and evaluate whether the operating state of the equipment is abnormal or the performance is degraded.

[0094] (5) The method of dynamically generating a prediction curve reflecting the trend trend through the trend model includes:

[0095] The model prediction output submodule outputs the predicted value according to the set time step.

[0096] The curve smoothing submodule applies moving average or exponential smoothing method to reduce the noise of the prediction curve.

[0097] The confidence interval calculation submodule is used to add confidence intervals to the prediction curve to reflect the uncertainty of the prediction.

[0098] The dynamic update display submodule is used to ensure that the prediction curve is updated in real time as new data is input.

[0099] Generate a visual forecast representation based on the processed forecast results and quantitative results. Draw the processed forecast trend curve and add confidence interval bands on the curve to intuitively display the uncertainty range of the forecast. Through a multi-dimensional interactive visualization interface, the forecast results are updated in real time, allowing users to view forecast data for different time periods. Provide interactive functions, users can zoom, drag, and click on charts to deeply analyze data; mark abnormal points in the forecast, add detailed information and explanations, and assist users in understanding and decision-making.

[0100] (6) The multi-dimensional interactive visualization interface includes:

[0101] Multi-layer data drilling submodule, allowing users to view data in depth from overview to details;

[0102] Custom view configuration submodule, which allows users to select chart types and color coding according to their needs;

[0103] Real-time interactive feedback submodule, used to respond to user operations instantly;

[0104] The anomaly labeling and annotation submodule is used to label abnormal data points.

[0105] Preferably, the user can freely select the display dimension and adjust the time axis including:

[0106] The time range selector submodule is used for user-defined time windows;

[0107] The dimension switching panel submodule is used by users to switch between different dimensions;

[0108] Dynamic timeline zoom submodule, used to provide mouse wheel zooming and dragging timeline functions;

[0109] The data refresh frequency setting submodule is used by users to adjust the frequency of chart data refresh according to their needs.

[0110] (7) The alarm and suggestion module includes:

[0111] A real-time monitoring submodule is used to continuously track the deviation of analysis results from the preset standards;

[0112] An abnormal pattern recognition submodule, used to identify specific abnormal behavior patterns based on machine learning;

[0113] The rule engine submodule is used to determine whether to trigger an alarm based on preset rules;

[0114] The notification channel management submodule is used to configure and manage various channels for alert push.

[0115] (8) The trigger alarm notification includes:

[0116] The urgency classification submodule is used to automatically classify the alarm level according to the severity of the anomaly;

[0117] The notification content customization submodule is used to generate notification text containing anomaly description, impact scope, and recommended measures;

[0118] The hierarchical response mechanism submodule is used to automatically dispatch response teams at different levels according to the alarm level.

[0119] (9) The optimization suggestions made based on historical data include:

[0120] The historical case library maintenance submodule is used to store and manage historical fault resolution cases;

[0121] Similarity matching submodule, used to match current anomalies with historical cases;

[0122] The optimization strategy generation submodule is used to generate optimization suggestions based on the matching results.

[0123] Embodiment 3, an embodiment of the present invention, is different from the first two embodiments in that:

[0124] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0125] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0126] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0127] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0128] Example 4, taking the machine tool equipment in a production workshop as the first object, the operating data of the equipment is used for quality monitoring and statistics. The machine tool equipment monitors its operating status through sensors, including multiple parameters such as temperature, vibration, pressure, and speed. Through the real-time data stream acquisition module, these operating data are continuously collected. During the test, the collected data will be improved through signal processing and cleaning processes to remove noise and outliers.

[0129] Based on the collected operation data, the operation status of the equipment is predicted. First, the historical operation data of the equipment is feature extracted, including the equipment's duty cycle, load conditions, environmental factors, etc., to extract the features that have a greater impact on the equipment's operation status. Then, the ARIMA forecasting model is used to predict the future status of the equipment. The model is fitted based on historical data to predict the future operation status.

[0130] In order to improve the readability of the prediction results, the generated prediction results are visualized, and the prediction trends are displayed in the form of charts, curves, etc. The uncertainty of the prediction results is quantified, and the error interval (such as 95% confidence interval) is used to indicate the reliability of the prediction.

[0131] Next, the machine learning algorithm is used to detect anomalies in the visualization prediction results. Taking the isolation forest model as an example, the visualization results are input into the model in real time, and the anomaly score of each data point is calculated to determine whether the equipment operation is abnormal. If the anomaly score is higher than the preset threshold, it is marked as an anomaly point.

[0132] At the same time, a linear regression algorithm is used to perform deviation analysis on the equipment's operating data to calculate the deviation between the equipment's predicted value and the actual value. The deviation value and deviation percentage are used to evaluate the degree of deviation of the equipment's current operating state. Combining the abnormality detection results and the deviation analysis results, the equipment's operating state is finally determined, and regulation is performed based on this, such as adjusting the machine tool's operating parameters, or issuing an alarm notification for manual intervention. The experimental results are shown in Table 1.

[0133] Table 1 Experimental data table

[0134]

[0135]

[0136] According to Table 1, the deviation value and deviation percentage between the predicted value and the actual value reflect the current operating status of the equipment. For example, the actual temperature of device A experiment 1 is 65.2℃, while the predicted temperature is 66.0℃, with a deviation value of 1.2% and a low deviation percentage, indicating that the equipment is relatively stable. The deviation percentage of device B experiment 2 is 0.7%, which also shows a relatively small deviation, and the equipment is in good operating condition.

[0137] However, in experiment 2 of device C, although the deviation percentage is 0.6%, there is still a slight fluctuation between the upper and lower limits of the deviation range. This small fluctuation may be caused by changes in the external environment or uneven distribution inside the device. Through deviation analysis, combined with the historical data of the device, the control strategy can be further refined to optimize the device operation status and reduce instability.

[0138] In addition, through the anomaly detection of the isolation forest model, the anomaly score for each data point can help determine in real time whether there is a potential problem with the equipment. In the experimental data, the vibration frequency of device B in experiment 2 decreased slightly. The isolation forest model can detect this slight change and warn of possible equipment anomalies to prevent potential failures. Compared with traditional methods, the combination of anomaly detection and deviation analysis can provide more accurate and real-time quality monitoring for equipment.

[0139] From the data analysis results, the invented visualization prediction method can accurately evaluate the operating status of the equipment and determine whether the equipment needs to be regulated through the visualization results. Compared with the existing technology, this method not only has higher prediction accuracy, but also combines machine learning and linear regression analysis to more effectively cope with dynamically changing production environments, thereby providing real-time and dynamic quality monitoring and regulation capabilities.

[0140] Compared with traditional monitoring methods based on static rules, the use of the isolation forest algorithm for anomaly detection can significantly improve the accuracy of fault warnings. Traditional methods usually rely on manually set thresholds, while the isolation forest model can dynamically evaluate the degree of abnormality of data points, adjust the model based on training data, and avoid human intervention and incorrect threshold settings.

[0141] The application of linear regression model in deviation analysis can not only quantify the operation deviation of the equipment, but also more intuitively present the performance changes of the equipment through the deviation percentage, providing a basis for subsequent regulation and optimization. In all experiments, the results of the combination of linear regression algorithm and isolation forest model significantly improved the real-time monitoring and the accuracy of early warning, reflecting the innovation and advantages of the invented method.

[0142] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A visualization method for quality monitoring and statistics, characterized in that: include: continuously collecting operation data of the first object; Predicting an operating state of the first object based on the operating data of the first object, and generating a visual prediction result; The visual prediction result is used to determine the operating state of the first object, and the first object is regulated.

2. The visualization method for quality monitoring and statistics according to claim 1, characterized in that: The operation data of the first object includes real-time parameters and real-time operation status data generated during the operation of the first object.

3. The visualization method for quality monitoring and statistics according to claim 2, characterized in that: The continuously collecting the operation data of the first object includes continuously collecting the operation data of the first object through real-time technology and improving the quality of the operation data.

4. The visualization method for quality monitoring and statistics according to claim 3, characterized in that: Predicting the operating state of the first object based on the operating data of the first object includes extracting features based on the operating data of the first object to obtain operating features of the first object; Based on the operation characteristics of the first object, a prediction model is constructed to predict the operation status of the first object and generate a prediction result.

5. The visualization method for quality monitoring and statistics according to claim 4, characterized in that: Generating visual prediction results includes improving the readability of the prediction results and quantifying the uncertainty of the prediction results; Generate a visual prediction representation based on the processed prediction results and quantified results.

6. The visualization method for quality monitoring and statistics according to claim 5, characterized in that: The determining the operating state of the first object by using the visual prediction result includes performing anomaly detection on the visual prediction result by using a machine learning algorithm; performing a deviation analysis on the operating data of the first object using an analysis algorithm; The operation status of the first object is determined and regulated by combining the abnormality detection result and the deviation analysis result.

7. The visualization method for quality monitoring and statistics according to claim 6, characterized in that: The machine learning algorithm includes, but is not limited to, inputting the real-time visual prediction results into an isolation forest model, calculating an anomaly score for each data point; and marking data points with an anomaly score higher than a threshold as anomalies. The analysis algorithm includes, but is not limited to, inputting the first subject's operating data into a linear regression model to calculate a predicted value of the first subject's operation; The predicted value of the first object operation is compared with the actual value of the first object operation, the deviation value and the deviation percentage are calculated, and the deviation degree of the first object operation state is evaluated.

8. A visualization system for quality monitoring and statistics, characterized by: A real-time data stream acquisition module continuously collects operation data of the first object; A dynamic trend prediction module, which predicts the operating state of the first object based on the operating data of the first object and generates a visual prediction result; The analysis and control module uses the visual prediction result to determine the operating state of the first object and controls the first object.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the visualization method for quality monitoring and statistics described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the visualization method for quality monitoring and statistics described in any one of claims 1 to 7 are implemented.

Images