Escalator multi-source heterogeneous quality detection report generation method

Abstract

The invention discloses an escalator multi-source heterogeneous quality detection report generation method, which comprises the following steps that S1, quality detection information of an escalator is collected from a plurality of heterogeneous data sources, and the quality detection information contains diversity data of different types and formats; s2, compiling the quality detection information from each data source into a uniform data format for data integration, solving the difference and inconsistency among heterogeneous data, and forming a uniform data view; s3, performing analysis and anomaly prediction on the integrated data by using a data analysis and model algorithm to obtain a quality evaluation result; and S4, based on the quality evaluation result, a quality detection report is generated, and the quality detection report comprises various quality index data, a trend chart and an anomaly detection result. According to the method, the efficiency and accuracy of quality detection are improved, the quality detection report can be quickly, accurately and comprehensively generated, and comprehensive monitoring and timely maintenance of the running state of the system are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-source heterogeneous data analysis, and in particular to a method for generating a multi-source heterogeneous quality inspection report for an escalator. Background Art

[0002] Many companies, such as elevator manufacturing, face challenges in quality inspection and maintenance efficiency during actual operation. Companies use multi-source heterogeneous data such as various sensors, IoT devices, and image monitoring. However, due to the dispersion and inconsistent structure of multi-source heterogeneous data, it is difficult to conduct quality inspection and maintenance in a comprehensive and timely manner, and it is difficult to quickly and comprehensively generate quality inspection reports. The efficiency and quality of report generation are poor, and it is difficult to meet actual usage needs. Summary of the invention

[0003] The purpose of the present invention is to provide a method for generating a multi-source heterogeneous quality inspection report for an escalator, which improves the efficiency and accuracy of quality inspection, can generate quality inspection reports quickly, accurately and comprehensively, and realizes comprehensive monitoring and timely maintenance of the system operation status, thereby improving product quality and user satisfaction.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] A method for generating a multi-source heterogeneous quality inspection report for an escalator comprises the following steps:

[0006] S1. Collect escalator quality inspection information from multiple heterogeneous data sources. The quality inspection information covers diverse data of different types and formats.

[0007] S2. Compile the quality inspection information from various data sources into a unified data format for data integration, solve the differences and inconsistencies between heterogeneous data, and form a unified data view;

[0008] The data integration process of step S2 is:

[0009] S21. Data cleaning and preprocessing: Perform data cleaning for each data source, process irrelevant data, missing values ​​and outliers, standardize data and unify formats to ensure data consistency during the integration process;

[0010] Step S21 includes the following steps:

[0011] S211. Delete irrelevant feature data: For each feature, the correlation coefficient and Gini index weighting method are used to determine the relevance of the feature data and the decision. The formula is as follows:

[0012] Score_weighted=w1*Corr+w2*(1-Gini),

[0013] Among them, Score_weighted is the weighted score of the feature; Corr is the correlation coefficient between the feature and the target variable; Gini is the Gini index when the feature is used to divide the data set, w1 is the weight of Corr in the weighted average, and w2 is the weight of Gini in the weighted average; this formula weights the correlation coefficient and the Gini index, and balances the contribution of the two to the importance of the feature by adjusting the weights w1 and w2. It is used to comprehensively consider the linear correlation of the features and the ability to divide the data set, evaluate the relationship between the features and the decision, and sort the correlation scores of each feature, delete the features with correlation less than the threshold, and reduce the amount of data;

[0014] S212, fill missing values: Assume there is a feature X, which has missing values, calculate the mean and median of feature X, and use the mean and median weighted method to fill missing values: X filled =(X mean +X med ) / 2, where X mean is the mean of feature X, X med is the median of feature X, X filled is the filling value of feature X. For all missing values ​​in feature X, X is used filled to fill in;

[0015] S213, abnormal data processing: Assume that there is a feature X, which contains abnormal values, calculate the mean and standard deviation of feature X, use the mean and standard deviation to identify abnormal values, delete abnormal values, and then use the missing value filling method of step S212 to fill the abnormal values. The steps for processing abnormal values ​​are as follows:

[0016] Calculate the mean and standard deviation of feature X:

[0017]

[0018]

[0019] Among them, X i is the i-th value of feature X, n is the number of samples, X std is the standard deviation of feature X; outliers are defined as values ​​outside the range of k times the standard deviation of the mean, where k is an adjustable parameter. The outlier replacement formula is expressed as follows:

[0020]

[0021] Among them, X processed is the processed eigenvalue, k is the multiple of the outlier threshold, and k is 2 or 3; according to the condition, if the data point X iIf the outlier threshold is exceeded, the outlier is replaced with the fill value X. filled , otherwise remain unchanged;

[0022] S22, data mapping and matching: Map and match related data elements in different data sources to establish relationships between data;

[0023] S23. Formulate consistency rules: Determine consistency rules to resolve differences between heterogeneous data and handle conflicts that may arise during the integration process;

[0024] S24. Use ETL tools and data integration middleware to simplify the data integration process;

[0025] S25. Design a unified data model to ensure that the information of each data source is mapped to the same data structure to form a unified data view;

[0026] S26. Real-time data synchronization and update: For data that needs to be synchronized in real time, use real-time data synchronization tools to ensure that the integrated data remains up to date;

[0027] S3. Use data analysis and model algorithms to analyze and predict anomalies on the integrated data and obtain quality assessment results;

[0028] Step S3 includes the following steps:

[0029] S31. Calculation of quality inspection indicators: determine the quality inspection indicators that need to be evaluated, and use the integrated data to calculate the specific values ​​of the quality inspection indicators;

[0030] S32. Trend analysis: Establish a time series analysis model to conduct trend analysis on quality inspection indicators, and use statistical methods, machine learning models or deep learning models to make trend predictions based on historical data;

[0031] S33, Anomaly detection: Establish an anomaly detection model based on machine learning, and use the anomaly detection model to detect abnormal data and identify anomalies in quality inspection indicators;

[0032] S34, Model building and prediction: Establish a prediction model for quality inspection, based on historical data and integrated data, using regression analysis, time series analysis and machine learning models to predict future quality status according to changes in quality inspection indicators;

[0033] S4. Generate a quality inspection report based on the quality assessment results. The quality inspection report includes various quality indicator data, trend charts and abnormal detection results.

[0034] Preferably, the quality inspection information in step S1 comes from sensors, IoT devices, images, document records, inspection reports, component quality, user feedback and maintenance, and each heterogeneous data is converted into JSON format.

[0035] Preferably, in step S1, sensor data is used to obtain the operating status and performance indicators of each component for quality inspection and prediction of maintenance needs; IoT devices are used to realize remote monitoring of the operating status and real-time feedback of data; cameras or video surveillance equipment are installed, and image recognition technology is used to perform quality inspection; information in document records is used as the basis for quality inspection, and the content of the document records includes the specifications, materials and manufacturing processes of the components; the maintenance personnel's observations and records of the operating status are obtained in combination with the information in the inspection report; quality data of components in the supply chain are collected for tracing the quality and source of the components used; user feedback and repair records are collected, and the overall quality status is evaluated in combination with manual inspections.

[0036] Preferably, step S22 includes the following steps:

[0037] S221, data matching: using a similarity-based matching method to calculate similarity scores of data elements, and to associate or match related data elements in different data sources;

[0038] S222. Data mapping: Use metadata management tools to define metadata in different data sources, and map data elements in different data sources to common identifiers or concepts based on the metadata.

[0039] Preferably, the consistency rules in step S23 include conversion of data units and alignment of timestamps.

[0040] Preferably, step S4 comprises the following steps:

[0041] S41. Report structure design: define the structure of the quality inspection report, determine the content to be included in the report and the order of arrangement of each part, and consider the needs of users so that the report can meet the information needs of different user groups;

[0042] S42, Quality Index Summary: Statistics and calculations are performed on the integrated data, and detailed data of each quality inspection index are summarized;

[0043] S43, Trend chart display: Use data visualization tools to generate trend charts to display the changing trends of quality inspection indicators, so that users can intuitively understand the evolution of quality conditions. Trend charts include line charts, bar charts and pie charts;

[0044] S44, abnormal detection result report: the abnormal detection results are organized as part of the quality inspection report. The abnormal detection results include the detected abnormal conditions and the specific cause analysis of the abnormal conditions, and provide abnormal condition descriptions to help users understand the special conditions in quality inspection;

[0045] S45, prediction results and suggestions: In the analysis phase, the prediction of quality inspection is carried out, and the prediction results of future quality status are integrated into the quality inspection report. The prediction results include the predicted value and credibility assessment of future quality status;

[0046] S46. Visual presentation: Use charts, graphs and tables to present the quality inspection report visually.

[0047] After adopting the above technical solution, the present invention has the following beneficial effects: The present invention aims to deal with quality inspection information from different sources and heterogeneous data, and proposes a method for generating a multi-source heterogeneous quality inspection report for escalators. The method integrates multiple data sources, uses advanced data integration and analysis technology, generates a detailed quality inspection report, and provides users with comprehensive and accurate quality assessment. First, quality inspection information is collected from multiple heterogeneous data sources, covering diverse data of different types and formats, including sensor data, image data, text data, etc. Secondly, data integration technology is used to compile information from various data sources into a unified data format for fusion, solve the differences and inconsistencies between heterogeneous data, and form a unified data view. Thirdly, data analysis and model algorithms are used to conduct in-depth analysis of the integrated data, involving the calculation of quality inspection indicators, trend analysis, anomaly detection, etc., to obtain more comprehensive and accurate quality assessment results. Finally, based on the results of the analysis, a quality inspection report is generated to provide users with clear quality assessment information. The report content includes detailed data of various quality indicators, trend charts, anomaly detection results, etc., so that users can have a deep understanding of the quality status of the inspected object. This method provides more comprehensive and accurate information support for quality inspection through the comprehensive utilization and efficient integration of multi-source heterogeneous data. It can be applied to various fields, such as manufacturing, healthcare, etc., to improve the level of quality management and optimize the decision-making process. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0050] like Figure 1As shown, a method for generating a multi-source heterogeneous quality inspection report for an escalator includes the following steps:

[0051] S1. Collect escalator quality inspection information from multiple heterogeneous data sources. The quality inspection information covers diverse data of different types and formats.

[0052] The quality inspection information in step S1 comes from sensors, IoT devices, images, document records, inspection reports, component quality, user feedback and maintenance, and each heterogeneous data is converted into JSON format;

[0053] In step S1, sensor data is used to obtain the operating status and performance indicators of each component for quality inspection and prediction of maintenance needs; IoT devices are used to achieve remote monitoring of the operating status and real-time feedback of data; cameras or video surveillance equipment are installed, and image recognition technology is used for quality inspection; information in document records is used as the basis for quality inspection, and the content of the document records includes the specifications, materials and manufacturing processes of the components; combined with the information in the inspection report, the maintenance personnel's observations and records of the operating status are obtained; quality data of components in the supply chain are collected to trace the quality and source of the components used; user feedback and repair records are collected, and combined with manual inspections, the overall quality status is evaluated.

[0054] S2. Compile the quality inspection information from various data sources into a unified data format for data integration, solve the differences and inconsistencies between heterogeneous data, and form a unified data view;

[0055] The data integration process of step S2 is:

[0056] S21. Data cleaning and preprocessing: Perform data cleaning for each data source, process irrelevant data, missing values ​​and outliers, standardize data and unify formats to ensure data consistency during the integration process;

[0057] Step S21 includes the following steps:

[0058] S211. Delete irrelevant feature data: For each feature, the correlation coefficient and Gini index weighting method are used to determine the relevance of the feature data and the decision. The formula is as follows:

[0059] Score_weighted=w1*Corr+w2*(1-Gini),

[0060] Among them, Score_weighted is the weighted score of the feature; Corr is the correlation coefficient between the feature and the target variable; Gini is the Gini index when the feature is used to divide the data set, w1 is the weight of Corr in the weighted average, and w2 is the weight of Gini in the weighted average; this formula weights the correlation coefficient and the Gini index, and balances the contribution of the two to the importance of the feature by adjusting the weights w1 and w2. It is used to comprehensively consider the linear correlation of the features and the ability to divide the data set, evaluate the relationship between the features and the decision, and sort the correlation scores of each feature, delete the features with correlation less than the threshold, and reduce the amount of data;

[0061] S212, fill missing values: Assume there is a feature X, which has missing values, calculate the mean and median of feature X, and use the mean and median weighted method to fill missing values: X filled =(X mean +X med ) / 2, where X mean is the mean of feature X, X med is the median of feature X, X filled is the filling value of feature X. For all missing values ​​in feature X, X is used filled to fill in;

[0062] S213, abnormal data processing: Assume that there is a feature X, which contains abnormal values, calculate the mean and standard deviation of feature X, use the mean and standard deviation to identify abnormal values, delete abnormal values, and then use the missing value filling method of step S212 to fill the abnormal values. The steps for processing abnormal values ​​are as follows:

[0063] Calculate the mean and standard deviation of feature X:

[0064]

[0065]

[0066] Among them, X i is the i-th value of feature X, n is the number of samples, X std is the standard deviation of feature X; outliers are defined as values ​​outside the range of k times the standard deviation of the mean, where k is an adjustable parameter. The outlier replacement formula is expressed as follows:

[0067]

[0068] Among them, X processed is the processed eigenvalue, k is the multiple of the outlier threshold, and k is 2 or 3; according to the condition, if the data point X iIf the outlier threshold is exceeded, the outlier is replaced with the fill value X. filled , otherwise remain unchanged;

[0069] S22, data mapping and matching: Map and match related data elements in different data sources to establish relationships between data;

[0070] Step S22 includes the following steps:

[0071] S221, data matching: using a similarity-based matching method to calculate similarity scores of data elements, and to associate or match related data elements in different data sources;

[0072] S222, Data Mapping: Use metadata management tools to define metadata in different data sources, and map data elements in different data sources to common identifiers or concepts based on the metadata;

[0073] S23. Formulate consistency rules: Determine consistency rules to resolve differences between heterogeneous data and handle conflicts that may arise during the integration process;

[0074] The consistency rules in step S23 include conversion of data units and alignment of timestamps;

[0075] S24. Use ETL tools and data integration middleware to simplify the data integration process;

[0076] S25. Design a unified data model to ensure that the information of each data source is mapped to the same data structure to form a unified data view;

[0077] S26. Real-time data synchronization and update: For data that needs to be synchronized in real time, use real-time data synchronization tools to ensure that the integrated data remains up to date;

[0078] S3. Use data analysis and model algorithms to analyze and predict anomalies on the integrated data and obtain quality assessment results;

[0079] Step S3 includes the following steps:

[0080] S31. Calculation of quality inspection indicators: determine the quality inspection indicators that need to be evaluated, and use the integrated data to calculate the specific values ​​of the quality inspection indicators;

[0081] S32. Trend analysis: Establish a time series analysis model to conduct trend analysis on quality inspection indicators, and use statistical methods, machine learning models or deep learning models to make trend predictions based on historical data;

[0082] S33, Anomaly detection: Establish an anomaly detection model based on machine learning, and use the anomaly detection model to detect abnormal data and identify anomalies in quality inspection indicators;

[0083] S34, Model building and prediction: Establish a prediction model for quality inspection, based on historical data and integrated data, using regression analysis, time series analysis and machine learning models to predict future quality status according to changes in quality inspection indicators;

[0084] S4. Generate a quality inspection report based on the quality assessment results. The quality inspection report includes various quality indicator data, trend charts and abnormal detection results;

[0085] Step S4 includes the following steps:

[0086] S41. Report structure design: define the structure of the quality inspection report, determine the content to be included in the report and the order of arrangement of each part, and consider the needs of users so that the report can meet the information needs of different user groups;

[0087] S42, Quality Index Summary: Statistics and calculations are performed on the integrated data, and detailed data of each quality inspection index are summarized;

[0088] S43, Trend chart display: Use data visualization tools to generate trend charts to display the changing trends of quality inspection indicators, so that users can intuitively understand the evolution of quality conditions. Trend charts include line charts, bar charts and pie charts;

[0089] S44, abnormal detection result report: the abnormal detection results are organized as part of the quality inspection report. The abnormal detection results include the detected abnormal conditions and the specific cause analysis of the abnormal conditions, and provide abnormal condition descriptions to help users understand the special conditions in quality inspection;

[0090] S45, prediction results and suggestions: In the analysis phase, the prediction of quality inspection is carried out, and the prediction results of future quality status are integrated into the quality inspection report. The prediction results include the predicted value and credibility assessment of future quality status;

[0091] S46. Visual presentation: Use charts, graphs and tables to present the quality inspection report visually.

[0092] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for generating a multi-source heterogeneous quality inspection report for an escalator, characterized in that: The following steps are involved: S1. Collect escalator quality inspection information from multiple heterogeneous data sources. The quality inspection information covers diverse data of different types and formats. S2. Compile the quality inspection information from various data sources into a unified data format for data integration, solve the differences and inconsistencies between heterogeneous data, and form a unified data view; The data integration process of step S2 is: S21. Data cleaning and preprocessing: Perform data cleaning for each data source, process irrelevant data, missing values ​​and outliers, standardize data and unify formats to ensure data consistency during the integration process; Step S21 includes the following steps: S211. Delete irrelevant feature data: For each feature, the correlation coefficient and Gini index weighting method are used to determine the relevance of the feature data and the decision. The formula is as follows: Score_weighted=w1*Corr+w2*(1-Gini), Among them, Score_weighted is the weighted score of the feature; Corr is the correlation coefficient between the feature and the target variable; Gini is the Gini index when the feature is used to divide the data set, w1 is the weight of Corr in the weighted average, and w2 is the weight of Gini in the weighted average; this formula weights the correlation coefficient and the Gini index, and balances the contribution of the two to the importance of the feature by adjusting the weights w1 and w2. It is used to comprehensively consider the linear correlation of the features and the ability to divide the data set, evaluate the relationship between the features and the decision, and sort the correlation scores of each feature, delete the features with correlation less than the threshold, and reduce the amount of data; S212, fill missing values: Assume there is a feature X, which has missing values, calculate the mean and median of feature X, and use the mean and median weighted method to fill missing values: X filled =(X mean +X med ) / 2, where X mean is the mean of feature X, X med is the median of feature X, X filled is the filling value of feature X. For all missing values ​​in feature X, X is used filled to fill in; S213, abnormal data processing: Assume that there is a feature X, which contains abnormal values, calculate the mean and standard deviation of feature X, use the mean and standard deviation to identify abnormal values, delete abnormal values, and then use the missing value filling method of step S212 to fill the abnormal values. The steps for processing abnormal values ​​are as follows: Calculate the mean and standard deviation of feature X: Among them, X i is the i-th value of feature X, n is the number of samples, X std is the standard deviation of feature X; outliers are defined as values ​​outside the range of k times the standard deviation of the mean, where k is an adjustable parameter. The outlier replacement formula is expressed as follows: Among them, X processed is the processed eigenvalue, k is the multiple of the outlier threshold, and k is 2 or 3; according to the condition, if the data point X i If the outlier threshold is exceeded, the outlier is replaced with the fill value X. filled , otherwise remain unchanged; S22. Data mapping and matching: Map and match related data elements in different data sources to establish relationships between data; S23. Formulate consistency rules: Determine consistency rules to resolve differences between heterogeneous data and handle conflicts that may arise during the integration process; S24. Use ETL tools and data integration middleware to simplify the data integration process; S25. Design a unified data model to ensure that the information of each data source is mapped to the same data structure to form a unified data view; S26. Real-time data synchronization and update: For data that needs to be synchronized in real time, use real-time data synchronization tools to ensure that the integrated data remains up to date; S3. Use data analysis and model algorithms to analyze and predict anomalies on the integrated data and obtain quality assessment results; Step S3 includes the following steps: S31. Calculation of quality inspection indicators: determine the quality inspection indicators that need to be evaluated, and use the integrated data to calculate the specific values ​​of the quality inspection indicators; S32. Trend analysis: Establish a time series analysis model to conduct trend analysis on quality inspection indicators, and use statistical methods, machine learning models or deep learning models to make trend predictions based on historical data; S33, Anomaly detection: Establish an anomaly detection model based on machine learning, and use the anomaly detection model to detect abnormal data and identify anomalies in quality inspection indicators; S34, Model building and prediction: Establish a prediction model for quality inspection, based on historical data and integrated data, using regression analysis, time series analysis and machine learning models to predict future quality status according to changes in quality inspection indicators; S4. Generate a quality inspection report based on the quality assessment results. The quality inspection report includes various quality indicator data, trend charts and abnormal detection results.

2. The method for generating a multi-source heterogeneous quality inspection report for an escalator according to claim 1, characterized in that: The quality inspection information in step S1 comes from sensors, IoT devices, images, document records, inspection reports, component quality, user feedback and maintenance, and each heterogeneous data is converted into JSON format.

3. The method for generating a multi-source heterogeneous quality inspection report for an escalator according to claim 2, characterized in that: In step S1, the operating status and performance indicators of each component are obtained by using sensor data for quality inspection and predicting maintenance needs; the operating status is remotely monitored and the data is fed back in real time by using IoT devices; cameras or video surveillance equipment are installed and image recognition technology is used for quality inspection; the information in the document records is used as the basis for quality inspection, and the content of the document records includes the specifications, materials and manufacturing processes of the components; the maintenance personnel's observations and records of the operating status are obtained in combination with the information in the inspection report; the quality data of the components in the supply chain are collected to trace the quality and source of the components used; Collect user feedback and repair records, combine with manual inspections, and evaluate the overall quality status.

4. The method for generating a multi-source heterogeneous quality inspection report for an escalator according to claim 1, characterized in that: Step S22 includes the following steps: S221, data matching: using a similarity-based matching method to calculate similarity scores of data elements, and to associate or match related data elements in different data sources; S222. Data mapping: Use metadata management tools to define metadata in different data sources, and map data elements in different data sources to common identifiers or concepts based on the metadata.

5. The method for generating a multi-source heterogeneous quality inspection report for an escalator according to claim 1, characterized in that: The consistency rules in step S23 include conversion of data units and alignment of timestamps.

6. The method for generating a multi-source heterogeneous quality inspection report for an escalator according to claim 1, characterized in that: Step S4 includes the following steps: S41. Report structure design: define the structure of the quality inspection report, determine the content to be included in the report and the order of arrangement of each part, and consider the needs of users so that the report can meet the information needs of different user groups; S42, Quality Index Summary: Statistics and calculations are performed on the integrated data, and detailed data of each quality inspection index are summarized; S43, Trend chart display: Use data visualization tools to generate trend charts to display the changing trends of quality inspection indicators, so that users can intuitively understand the evolution of quality conditions. Trend charts include line charts, bar charts and pie charts; S44, abnormal detection result report: the abnormal detection results are organized as part of the quality inspection report. The abnormal detection results include the detected abnormal conditions and the specific cause analysis of the abnormal conditions, and provide abnormal condition descriptions to help users understand the special conditions in quality inspection; S45, prediction results and suggestions: In the analysis phase, the prediction of quality inspection is carried out, and the prediction results of future quality status are integrated into the quality inspection report. The prediction results include the predicted value and credibility assessment of future quality status; S46. Visual presentation: Use charts, graphs and tables to present the quality inspection report visually.

Images