Comprehensive quality management early warning and supervision system

Abstract

The invention discloses a comprehensive quality management early-warning and supervision system, relates to the technical field of quality management, and solves the problems that a defect type classification model is difficult to construct, whether a secondary early-warning signal needs to be triggered is difficult to judge through a comprehensive index obtained by standardization values of various production devices, and the production efficiency is high. And the comparison judgment of the comprehensive quality safety coefficient and the normal comprehensive quality safety index is lacked. Comprising a production data acquisition and transmission management module, a user identity verification module, a fan defect and ledger data analysis module and a quality safety index comparison and early warning module, and is used for acquiring operation parameters, appearance quality and ledger data of production equipment; constructing role-based access control and dual verification; constructing a defect type classification model, a production equipment operation state comprehensive index and a comprehensive risk proportionality coefficient; and comparing the comprehensive quality safety coefficient with a normal comprehensive quality safety index, and judging whether a primary early warning signal is triggered or not.

Description

Technical Field

[0001] The present invention relates to the technical field of quality management, and in particular to a comprehensive quality management early warning and supervision system. Background Art

[0002] With the intensification of market competition and the continuous improvement of consumers' requirements for product quality, companies need to manage quality more efficiently and accurately to ensure that products and services meet standards and meet customer requirements. Therefore, it is particularly important to develop a system that can timely warn and monitor quality problems. By constructing a defect type classification model, the number of fans with different defect types and the total number of fans are counted to calculate the fan defect rate, and a variety of production equipment operating parameters are obtained from sensors to construct a comprehensive index. The operating status of the production equipment is analyzed and judged, and the comprehensive risk ratio coefficient is obtained by combining a variety of ledger data to obtain a comprehensive comprehensive quality safety factor, which is compared and analyzed with the normal safety factor to determine whether it is necessary to trigger an early warning signal.

[0003] Traditional quality management methods often rely on manual inspection and post-processing, making it difficult to discover and solve quality problems in real time, resulting in increased costs and decreased customer satisfaction. It is also difficult to analyze defective fan images to build a defect type classification model to calculate the fan defect rate. It is also difficult to combine the weighted addition of multiple standardized values ​​of production equipment to obtain the numerical range of the comprehensive index of the production equipment operation status to determine whether it is necessary to trigger the second-level warning signal. There is a lack of combining multiple ledger data to obtain a comprehensive risk ratio coefficient to obtain a comprehensive comprehensive quality safety factor, and compare it with the normal safety index obtained according to industry standards and corporate quality reports to determine whether there is an abnormality in the production process and issue a first-level warning signal in a timely manner. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a comprehensive quality management early warning and supervision system for solving the following technical problems:

[0005] Traditional quality management methods often rely on manual inspection and post-processing, making it difficult to discover and solve quality problems in real time, resulting in increased costs and decreased customer satisfaction. It is also difficult to analyze defective fan images to build a defect type classification model to calculate the fan defect rate. It is also difficult to combine the weighted addition of multiple standardized values ​​of production equipment to obtain the numerical range of the comprehensive index of the production equipment operation status to determine whether it is necessary to trigger the second-level warning signal. There is a lack of combining multiple ledger data to obtain a comprehensive risk ratio coefficient to obtain a comprehensive comprehensive quality safety factor, and compare it with the normal safety index obtained according to industry standards and corporate quality reports to determine whether there is an abnormality in the production process and issue a first-level warning signal in a timely manner.

[0006] To solve the above problems, the present invention provides a comprehensive quality management early warning and supervision system, including the following modules:

[0007] Production data collection and transmission management module: During the production process, various sensors and intelligent detection equipment are used to collect the operating parameters of production equipment, the appearance quality data of fans and ledger data in real time, and the collected data is transmitted to the server for storage through the data collection program;

[0008] User identity verification module: defines different roles according to the user's job responsibilities and authority scope, and performs dual authentication by combining the input user name and password with fingerprint recognition. When the authentication succeeds, the user enters the system. When the authentication fails, a three-level warning signal is triggered;

[0009] Fan defect and ledger data analysis module: Build a defect type classification model, identify and classify defective fans in the production process, build a fan defect rate, obtain a comprehensive index of the production equipment operation status by weighting and adding multiple standardized values, and determine whether to trigger a secondary warning signal. Analyze the personnel ledger, fire ledger, and system ledger data respectively, and calculate the personnel risk ratio coefficient, fire accident risk ratio coefficient, and system risk ratio coefficient;

[0010] Quality and safety index comparison and early warning module: The comprehensive quality and safety coefficient obtained by combining the fan defect rate and the comprehensive risk ratio coefficient is compared with the normal comprehensive quality and safety index, and the result of the comparison is used to determine whether to trigger a first-level early warning signal.

[0011] Furthermore, in the production process, various sensors and intelligent detection equipment are used to collect the operating parameters of the production equipment, the appearance quality data of the fan and the ledger data in real time, including the following steps:

[0012] Install a variety of sensors at key locations of production equipment, including temperature sensors, pressure sensors, and vibration sensors. Set the corresponding data collection frequency according to the monitoring requirements of the production equipment, and read data from the sensors at the preset frequency in the data collection program;

[0013] A machine vision inspection system is installed on the production line to obtain the fan appearance image. The data acquisition program obtains the fan appearance quality data from the inspection system, and after associating the fan's production batch number and production time information, the obtained fan image data is pre-processed and analyzed;

[0014] View and obtain the corresponding data in the personnel ledger, fire ledger and system ledger from the ledger data list;

[0015] Configure a wireless network access point to connect the data acquisition terminal to the server. According to the type of data acquisition and the network architecture, determine the data transmission protocol. After encrypting the data using encryption technology, transmit the data to the server for storage.

[0016] Further, define different roles according to the user's job responsibilities and scope of authority, and perform dual authentication by combining the input username and password with fingerprint recognition. When the authentication is successful, the user enters the system. When the authentication fails, trigger a three-level warning signal, including the following steps:

[0017] Define different roles according to the user's job responsibilities and scope of authority. Each role has its specific set of permissions. Assign the user to the corresponding role, and the system monitors the user's access behavior in real time. The user first enters the username and password for the first step of identity verification. The system verifies whether the credentials provided by the user match the stored valid credentials in the database. If the match is successful, enter the second step of identity verification. If the match fails, login is not allowed. Perform the second step of identity verification on the user's fingerprint through a fingerprint scanning device. The system compares the scanned fingerprint with the pre-stored fingerprint information to determine the user's identity. If the comparison is successful, the user successfully enters the system. If the comparison fails, trigger a three-level warning signal.

[0018] Further, construct a defect type classification model to identify and classify defective fans during the production process, and count the number of defective fans and the total number of fans to obtain the fan defect rate, including the following steps:

[0019] According to the actual requirements and production situation, determine a production batch, and conduct statistics in units of production batches. Establish a batch traceability system. Determine the sampling quantity of each batch of fans through statistical methods and sampling standards. After obtaining the samples, use the defect type classification model to detect each fan in the samples, and obtain the defect information of each fan from the detection results, including whether there are defects and the types of defects. Record it in the form of a database, count the number of defective fans, use data processing software to separately count the number of fans of different defect types, add up the quantities of each defect type to obtain the total number of fan defects, count the total number of fans within the determined statistical range, and obtain the fan defect rate by dividing the total number of fan defects by the total number of fans;

[0020] Fan defect rate calculation formula:

[0021]

[0022] Among them, F represents the fan defect rate, N pore represents the number of fans with the pore defect type, N crack represents the number of fans with the crack defect type, N trachomaThe number of fans indicating the type of sand hole defect, N total Indicates the total number of fans.

[0023] Furthermore, the defect type classification model includes the following steps:

[0024] Obtain fan images containing different defect types from the production process, use an image annotation tool to annotate the collected images, annotate the defect type or normal category to which each sample belongs, perform preprocessing operations on the images, including data cleaning and image enhancement, select a model framework based on a convolutional neural network to build a defect type classification model, use edge detection algorithms and corner detection algorithms to extract the features in the image data, including the edge and corner features of the image, as the input of the model, use the annotated image data to train the model, update the parameters of the model through multiple iterations, deploy the trained model to the actual production environment, enable it to automatically identify and classify the defect types on the surface of the fan, and in combination with the automated production line, the defect type classification model detects and classifies the defects on the surface of the fan in real time.

[0025] Furthermore, obtaining the comprehensive index of the operating state of the production equipment by weighted summation of multiple obtained standardized values and determining whether to trigger a secondary warning signal includes the following steps:

[0026] Analyze the historical temperature, pressure, and vibration data of the production equipment in the normal operating state, determine the appropriate temperature range, pressure range, and vibration range from the historical data as the normal temperature range, pressure range, and vibration range of the production equipment, and determine the maximum and minimum values of the corresponding production equipment operating parameters; perform standardization processing on the temperature, pressure, and vibration data through the currently measured actual temperature value, actual pressure value, and actual vibration value to obtain the temperature standardization value, pressure standardization value, and vibration standardization value respectively, and obtain the comprehensive index of the operating state of the production equipment by weighted summation of the obtained multiple standardized values;

[0027] Formula for calculating the comprehensive index of the operating state of the production equipment:

[0028]

[0029] Among them, E represents the comprehensive index of the operating state of the production equipment, T represents the current actual temperature value, P represents the current actual pressure value, V represents the current actual vibration value, T max represents the maximum temperature, T min represents the minimum temperature, P max represents the maximum pressure, P min represents the minimum pressure, V max represents the maximum vibration, V minrepresents the minimum vibration value, ω1, ω2 and ω3 represent the corresponding weight coefficients of temperature, pressure and vibration standardization value respectively;

[0030] When 0≤the comprehensive index of the production equipment operation status≤1, it indicates that the operation status of the production equipment is normal and no secondary warning signal is triggered; when the comprehensive index of the production equipment operation status is <0 or the comprehensive index of the production equipment operation status is >1, it indicates that the operation status of the production equipment is abnormal and a secondary warning signal is triggered.

[0031] Furthermore, the data of the personnel ledger, the fire protection ledger and the system ledger are analyzed respectively to calculate the personnel risk ratio coefficient, the fire accident risk ratio coefficient and the system risk ratio coefficient, including the following steps:

[0032] Obtain training record information from the personnel ledger, and track the work performance of employees within a certain period of time after receiving training. Obtain the product input ratio by comparing the total cost of training investment with the total benefits brought by employees after training. Obtain job change information from the personnel ledger, count the number of employees joining and leaving the company in a specific period of time, and obtain the staff turnover rate by the ratio of the number of people leaving and the average number of employees in the company. Add the weighted product input ratio and staff turnover rate to obtain the personnel risk ratio coefficient. Obtain information related to firefighting facilities from the firefighting ledger, regularly check the number and status of firefighting facilities, count the number of intact firefighting facilities and the number of damaged firefighting facilities, and obtain the firefighting facility damage rate by the ratio of the number of damaged firefighting equipment to the total number of firefighting facilities. , obtain relevant information about fire accidents in the fire protection ledger, count the number of fire accidents that occur in enterprises within a certain period of time to obtain the fire incidence rate, and add the fire protection facility damage rate and the fire incidence rate to obtain the fire accident risk ratio coefficient; obtain relevant information in the system ledger, check the implementation of various systems, count the number of matters that meet the system requirements and the number of matters that do not meet the system requirements, and obtain the system implementation non-compliance rate through the number of matters that do not meet the system requirements and the number of matters required by all systems, obtain the training record information in the system system ledger, count the number of people who missed the system training and the total number of people in the enterprise, calculate the system training absence rate, and add the system implementation non-compliance rate and the system training absence rate to obtain the system risk ratio coefficient.

[0033] Furthermore, the comprehensive overall quality safety coefficient obtained by combining the wind turbine defect rate and the comprehensive risk ratio coefficient is compared with the normal overall quality safety index, and whether a first-level warning signal is triggered is determined according to the comparison result, including the following steps:

[0034] The personnel risk ratio coefficient, the fire accident risk ratio coefficient and the system risk ratio coefficient are added together to obtain the comprehensive risk ratio coefficient, and the fan defect rate and the comprehensive risk ratio coefficient are weighted and added together to obtain the comprehensive overall quality and safety coefficient;

[0035] Comprehensive and overall quality safety factor calculation formula:

[0036] S = [F × 0.1+(P p +P f +P s ) × 0.1] × 100%

[0037] Among them, S represents the comprehensive and overall quality safety factor, F represents the fan defect rate, P p represents the personnel risk proportion coefficient, P f represents the fire accident risk proportion coefficient, P s represents the system and system risk proportion coefficient;

[0038] Determine the normal fan defect rate, as well as the normal personnel risk proportion coefficient, normal fire accident risk proportion coefficient and normal system and system risk proportion coefficient through the statistical analysis of comprehensive industry standards and multiple professional enterprise quality reports, add them up to obtain the normal comprehensive risk proportion coefficient, and obtain the normal overall quality safety index by adding the assignment of the normal defect rate and the normal comprehensive risk proportion coefficient;

[0039] Compare the comprehensive and overall quality safety factor with the normal overall quality safety index. If

[0040] the comprehensive and overall quality safety factor is less than the normal overall quality safety index, it means that the comprehensive ledger data is normal and does not trigger a first-level warning signal. If the comprehensive and overall quality safety factor is equal to or greater than the normal overall quality safety index, it means that the comprehensive ledger data is abnormal and triggers a first-level warning signal.

[0041] Advantages of the present invention:

[0042] The present invention extracts the edge and corner features in the image data through the edge detection algorithm and the corner detection algorithm, constructs a defect type classification model using the model framework based on the convolutional neural network. In the actual production process, the deployed model receives the fan images from the production line in real time, automatically classifies the images, and outputs the prediction results of whether there are defects on the fan and the defect types according to the input image features, realizing the automatic recognition and classification of the fan surface defects, replacing the manual detection, greatly saving human resources and improving the production efficiency;

[0043] By analyzing various ledger data, the present invention constructs a comprehensive risk ratio index by adding the personnel risk ratio coefficient, the fire accident risk ratio coefficient, and the system - based risk ratio coefficient. Combining with the defect type classification model to determine the defect rate of the fan to obtain the comprehensive overall quality and safety coefficient, and judging whether to trigger a first - level warning signal based on the result of comparing it with the normal overall quality and safety index. It realizes the combination of various ledger data and the fan defect rate, forms the whole - process quality control from ledger data to the fan, and judges whether to trigger a first - level warning signal by comparing with the normal overall quality and safety index, providing a clear quantitative standard for quality control. This timely warning mechanism helps to reduce quality risks, improve production efficiency and production quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the module process of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] Next, the technical methods in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0046] Refer to Figure 1 As shown, a total quality management warning and supervision system includes the following modules:

[0047] Production data collection and transmission management module: During the production process, various sensors and intelligent detection devices are used to collect the operating parameters of production equipment, the appearance quality data of the fan, and ledger data in real - time, and the collected data is transmitted to the server for storage through a data collection program;

[0048] User identity authentication module: Different roles are defined according to the job responsibilities and authority scopes of users, and double - authentication is performed by combining the input username and password with fingerprint recognition. When the authentication is successful, the user enters the system; when the authentication fails, a third - level warning signal is triggered;

[0049] Fan defect and ledger data analysis module: Construct a defect type classification model to identify and classify defective fans during the production process, construct a fan defect rate, obtain a comprehensive index of the operating state of production equipment by weighted addition of various obtained standardized values, and judge whether to trigger a second - level warning signal. Analyze the personnel ledger, fire ledger, and system - based ledger data respectively, and calculate the personnel risk ratio coefficient, the fire accident risk ratio coefficient, and the system - based risk ratio coefficient;

[0050] Quality and safety index comparison and early warning module: The comprehensive quality and safety coefficient obtained by combining the fan defect rate and the comprehensive risk ratio coefficient is compared with the normal comprehensive quality and safety index, and the result of the comparison is used to determine whether to trigger a first-level early warning signal.

[0051] Specifically, the operating parameters, appearance quality and various ledger data of the production equipment are collected; a defect type classification model is constructed, and according to the numerical range of the comprehensive index of the production equipment operation status, it is determined whether it is necessary to trigger the secondary warning signal and construct a comprehensive index of raw material quality indicators; the comprehensive comprehensive quality safety factor is compared with the normal comprehensive quality safety index to determine whether the first-level warning signal is triggered.

[0052] In one embodiment of the present invention, in the production process, various sensors and intelligent detection equipment are used to collect the operating parameters of the production equipment, the appearance quality data of the fan and the ledger data in real time, including the following steps:

[0053] Install a variety of sensors at key locations of production equipment, including temperature sensors, pressure sensors, and vibration sensors. Set the corresponding data collection frequency according to the monitoring requirements of the production equipment, and read data from the sensors at the preset frequency in the data collection program;

[0054] A machine vision inspection system is installed on the production line to obtain the fan appearance image. The data acquisition program obtains the fan appearance quality data from the inspection system, and after associating the fan's production batch number and production time information, the obtained fan image data is pre-processed and analyzed;

[0055] View and obtain the corresponding data in the personnel ledger, fire ledger and system ledger from the ledger data list;

[0056] Configure wireless network access points to connect data acquisition terminals to the server. Determine the data transmission protocol based on the type of data acquisition and network architecture. Use encryption technology to encrypt the data before transmitting it to the server for storage.

[0057] Specifically, according to the key parts of the production equipment and monitoring requirements, select appropriate temperature sensors, pressure sensors, and vibration sensors. For example, select a high-temperature-resistant temperature sensor for high-temperature environments; according to the characteristics of the fans on the production line and detection requirements, select an appropriate machine vision detection system, develop a data acquisition program to enable it to communicate with the machine vision detection system, and obtain the appearance quality data of the fans. In the data acquisition program, associate the obtained fan image data with the production batch number and production time information of the fan to establish a one-to-one correspondence; view the personnel ledger, fire ledger, and system ledger from the ledger data list, and formulate a wireless network coverage plan based on the production site layout of the enterprise, the number and distribution of data acquisition terminals, determine the number, location, and signal strength requirements of the wireless access points to be deployed, use a mobile terminal to log in to the management interface to configure the basic parameters of the wireless access points, select an appropriate data transmission protocol according to the type of data acquisition and the enterprise network architecture, encrypt the data using a symmetric encryption algorithm, and at the server end, receive the encrypted data from the data acquisition terminal and decrypt it using the corresponding decryption algorithm, and then store the decrypted data in the database.

[0058] In one embodiment of the present invention, different roles are defined according to the job responsibilities and authority scopes of users, and dual authentication is performed by combining the input of the user name and password with fingerprint recognition. When the authentication is successful, the user enters the system. When the authentication fails, a three-level warning signal is triggered, including the following steps:

[0059] Define different roles according to the job responsibilities and authority scopes of users. Each role has its specific set of permissions. Assign users to the corresponding roles, and the system monitors the user's access behavior in real time; the user first enters the user name and password for the first step of identity verification. The system verifies whether the credentials provided by the user match the stored valid credentials in the database. If the match is successful, it enters the second step of identity verification. If the match fails, the user is not allowed to log in; perform the second step of identity verification on the user's fingerprint through a fingerprint scanning device. The system compares the scanned fingerprint with the pre-stored fingerprint information to determine the user's identity. If the comparison is successful, the user successfully enters the system. If the comparison fails, a three-level warning signal is triggered.

[0060] Specifically, the user enters the username and password on the login interface. After the system receives the credentials entered by the user, it searches for the record corresponding to the username in the database and compares the password entered by the user with the password stored in the database. If the password matches successfully, the second-step authentication is entered; if the password match fails, the system prohibits the user from logging in and displays the corresponding error message, such as "Username or password is incorrect"; the system prompts the user to place the finger on the fingerprint scanning device. The fingerprint scanning device collects the user's fingerprint image and transmits it to the system. The system compares the scanned fingerprint image with the fingerprint template pre-stored in the database. If the fingerprint comparison is successful, the system confirms the user's identity, allows the user to successfully log in to the system, and jumps to the corresponding working page. If the fingerprint comparison fails, the system triggers a three-level warning signal, simultaneously prohibits the user from accessing the system, and displays information such as "Fingerprint verification failed" on the interface.

[0061] In one embodiment of the present invention, a defect type classification model is constructed to identify and classify defective fans during the production process, and the fan defect rate is obtained by counting the number of defective fans and the total number of fans, including the following steps:

[0062] According to the actual requirements and production situation, a production batch is determined. Statistics are carried out in units of production batches, a batch traceability system is established, the sampling quantity of each batch of fans is determined through statistical methods and sampling standards. After obtaining the samples, the defect type classification model is used to detect each fan in the samples, and the defect information of each fan is obtained from the detection results, including whether there are defects and the types of defects present. Records are made in the form of a database, the number of defective fans is counted, the data processing software is used to separately count the number of fans of different defect types, the numbers of each defect type are added up to obtain the total number of fan defects, the total number of fans within the determined statistical range is counted, and the fan defect rate is obtained by dividing the total number of fan defects by the total number of fans;

[0063] Fan defect rate calculation formula:

[0064]

[0065] Among them, F represents the fan defect rate, N pore represents the number of fans of the pore defect type, N crack represents the number of fans of the crack defect type, N trachoma represents the number of fans of the sand hole defect type, N total represents the total number of fans.

[0066] Specifically, determine an appropriate production batch division method according to factors such as the actual production plan, production process characteristics, and fan quality control requirements. For example, for an assembly line, a production batch can be divided every hour or every half day. Establish a batch traceability system using an information management system. During the fan production process, various information related to the fan is recorded in real time. According to the characteristics of the fan, the stability of the production process, and historical quality data, select an appropriate statistical method to determine the sampling quantity. According to the characteristics of the fan and common quality problems, collect historical quality problem data, including the type, manifestation form, and cause of defects. Train the model to ensure that it can accurately identify and classify various possible defect types of the fan. Record the test results of the sample fans in the form of a database, including details such as the serial number, production batch, test items, test data, whether there are defects, and defect types of the sample fans. For the fans with defects, mark and count them separately for subsequent statistical analysis of the number of defective fans. Use data processing software to statistically analyze the defect information recorded in the database. Classify according to the defect type and count the number of fans of each defect type separately. Add up the number of fans of each defect type to obtain the total number of defective fans in this production batch. Statistically count the total number of fans within the determined statistical range. Obtain the fan defect rate of this production batch according to the ratio of the total number of defective fans to the total number of fans.

[0067] In one embodiment of the present invention, the defect type classification model includes the following steps:

[0068] Obtain fan images containing different defect types from the production process. Use an image annotation tool to annotate the collected images, marking the defect type or normal category to which each sample belongs. Perform preprocessing operations on the images, including data cleaning and image enhancement. Select a model framework based on a convolutional neural network to construct a defect type classification model. Use edge detection algorithms and corner detection algorithms to extract features in the image data, including the edge and corner features of the image, as the input of the model. Use the annotated image data to train the model, and update the model parameters through multiple iterations. Deploy the trained model to the actual production environment to enable it to automatically identify and classify the defect types on the surface of the fan. Combined with the automated production line, the defect type classification model detects and classifies the defects on the surface of the fan in real time.

[0069] Specifically, install image acquisition devices in the production process to continuously collect fan images containing different defect types. At the same time, collect some normal fan images as comparison samples, store these images in a specified folder on the computer according to certain naming rules, use image annotation tools to annotate the collected images, perform cleaning and enhancement operations on the images for the defect areas or normal areas in each image. After selecting an edge detection algorithm to extract the edge features in the image data, use a corner detection algorithm to extract the corner features in the image data. Select a model framework based on a convolutional neural network to build a defect type classification model. Use the edge and corner features extracted from the image and the original image data as the input of the model. During the training process of the model, these features will jointly participate in the learning and updating of the model, enabling the model to make accurate defect classifications by fully utilizing various information features. Deploy the trained defect type classification model to the actual production environment, receive fan images from the production line in real time, and automatically classify the images. The model outputs the prediction results of whether there are defects in the fan and the defect types according to the input image features.

[0070] In one embodiment of the present invention, the step of obtaining the comprehensive index of the production equipment operation state by weighted addition of various obtained standardized values and determining whether to trigger a secondary warning signal includes the following steps:

[0071] Analyze the historical temperature, pressure, and vibration data of the production equipment in the normal operation state, determine the appropriate temperature range, pressure range, and vibration range from the historical data as the normal temperature range, pressure range, and vibration range of the production equipment, and determine the maximum and minimum values of the corresponding production equipment operation parameters; perform standardization processing on the temperature, pressure, and vibration data through the currently measured actual temperature value, actual pressure value, and actual vibration value to obtain the temperature standardization value, pressure standardization value, and vibration standardization value respectively, and obtain the comprehensive index of the production equipment operation state by weighted addition of the obtained various standardized values;

[0072] Comprehensive index calculation formula of production equipment operation state:

[0073]

[0074] where, E represents the comprehensive index of the production equipment operation state, T represents the current actual temperature value, P represents the current actual pressure value, V represents the current actual vibration value, T max represents the temperature maximum value, T min represents the temperature minimum value, P max represents the pressure maximum value, P min represents the pressure minimum value, V max represents the vibration maximum value, V minIt represents the minimum value of vibration, and ω1, ω2, and ω3 respectively represent the corresponding weight coefficients of temperature, pressure, and vibration standardized values;

[0075] When 0 ≤ the comprehensive index of the operating state of the production equipment ≤ 1, it indicates that the operating state of the production equipment is normal and no secondary warning signal is triggered; when the comprehensive index of the operating state of the production equipment < 0 or the comprehensive index of the operating state of the production equipment > 1, it indicates that the operating state of the production equipment is abnormal and a secondary warning signal is triggered.

[0076] Specifically, export the historical temperature, pressure, and vibration data of the equipment under normal operating conditions within a past period of time from the monitoring system of the production equipment, perform preliminary processing on the collected data by adding them up, remove obvious error or abnormal data points, calculate the mean and standard deviation of the historical temperature data. Generally speaking, the normal temperature range can be determined by floating a certain multiple of the standard deviation above and below the mean. For example, if it is considered that the normal temperature should fluctuate around the mean and usually does not exceed ±2 standard deviations, calculate the mean and standard deviation of the historical pressure data using a method similar to that of temperature, and determine the normal pressure range according to the characteristics of the equipment and process requirements. Analyze the historical vibration data, including characteristics such as the amplitude and frequency of vibration. Usually, the root mean square value can be used to represent the magnitude of vibration. Calculate the RMS mean and standard deviation of the historical vibration data, and determine the normal vibration range according to the operating requirements of the equipment; find the maximum and minimum values of temperature, pressure, and vibration in the historical data, measure the current temperature value, pressure value, and vibration value in real time through sensors installed on the production equipment, obtain the corresponding temperature, pressure, and vibration standardized values through the maximum-minimum normalization method, add the standardized values after multiplying them by the corresponding weights to obtain the comprehensive index of the operating state of the production equipment, and judge the operating state of the production equipment according to the calculated value of the comprehensive index and its change trend. Incorporate the above analysis process into the continuous detection system of the production equipment, and regularly update the historical data and recalculate the ranges and weights of various parameters to adapt to changes in equipment performance and operating environment; among them, the value of ω1 is 0.4, and the values of ω2 and ω3 are both 0.3.

[0077] In one embodiment of the present invention, analyzing the data of the personnel ledger, fire protection ledger, and system system ledger respectively, and calculating the personnel risk ratio coefficient, fire accident risk ratio coefficient, and system system risk ratio coefficient includes the following steps:

[0078] Obtain the training record information in the personnel ledger, and track the work performance of employees within a certain period after receiving training. Obtain the product input ratio by calculating the ratio of the total cost invested in training to the total benefits brought by employees after training. Obtain the job change information in the personnel ledger, count the number of new hires and departures within a specific period of the enterprise, and obtain the personnel turnover rate by calculating the ratio of the number of departures to the average number of employees. Add the product input ratio and the personnel turnover rate with weights to obtain the personnel risk ratio coefficient. Obtain the information related to fire protection facilities in the fire protection ledger, regularly check the quantity and status of fire protection facilities, count the number of intact fire protection facilities and the number of damaged fire protection facilities, and obtain the damage rate of fire protection facilities by calculating the ratio of the number of damaged fire protection facilities to the total number of fire protection facilities. Obtain the information related to fire accidents in the fire protection ledger, count the number of fire accidents occurred in the enterprise within a certain period to obtain the fire accident rate, and add the damage rate of fire protection facilities and the fire accident rate with weights to obtain the fire accident risk ratio coefficient. Obtain the relevant information in the system ledger, check the implementation of various systems, count the number of items that meet the system requirements and the number of items that do not meet the system requirements, and obtain the non-compliance rate of system implementation by calculating the ratio of the number of items that do not meet the system requirements to the total number of items required by the system. Obtain the training record information in the system ledger, count the number of people absent from system training and the total number of employees in the enterprise, calculate the absence rate of system training, and add the non-compliance rate of system implementation and the absence rate of system training with weights to obtain the system risk ratio coefficient.

[0079] Specifically, extract detailed information such as the training course name, training time, training duration, and training costs of employees from the personnel ledger. Within a specific period after employees receive training, for example, a month or a quarter, collect the employees' work performance data, including the completion of performance indicators, the results of work quality evaluations, and the improvement rate of work efficiency. First, calculate the total cost of training investment, that is, the total cost generated by all employees participating in training, including training course fees, training material fees, training instructor fees, etc. Then, calculate the total benefits brought by employees after training. For example, if an employee's work efficiency increases by 20% after training, a task that originally took 100 hours a week to complete now only takes 80 hours. Assuming the value of the employee's work per hour is 50 yuan, then the increase in the employee's weekly benefits is 20×50 = 1000 yuan. Add up the increased benefits of all employees to get the total benefits. Finally, divide the total cost of training investment by the total benefits brought by employees after training to obtain the product input ratio. Screen out the employee onboarding records and departure records within a specific period from the personnel ledger, including information such as the onboarding time, departure time, and job title. Through statistical analysis of job movement information, calculate the total number of new hires and the total number of departures within a specific period respectively. Divide the number of departures by the average number of employees on the job to obtain the staff turnover rate, where the average number of employees on the job is the average of the number of employees on the job at the beginning and the number of employees on the job at the end. Extract detailed information such as the name of fire protection facilities, installation location, purchase time, inspection time, inspectors, and inspection results from the fire protection ledger. According to the set inspection cycle, such as once a month, arrange special personnel to conduct on-site inspections of fire protection facilities and update the inspection information in the fire protection ledger. During the inspection process, record the status of each fire protection facility in detail, such as whether it is operating normally and whether it needs repair. Through statistical analysis of the inspection results in the fire protection ledger, calculate the number of intact fire protection facilities and the number of damaged fire protection facilities. Divide the number of damaged fire protection facilities by the total number of fire protection facilities to obtain the fire protection facility damage rate. Obtain the fire accident records in the fire protection ledger, including information such as the accident occurrence time, accident cause, and accident losses. Statistically analyze the number of fire accidents that occurred in the enterprise within a certain period and calculate the fire accident incidence rate. For example, if the enterprise had a total of 3 fire accidents in a year, then the fire accident incidence rate for that year is 3 times / year. Extract information such as the specific content, implementation requirements, and inspection standards of each system from the system system ledger, and at the same time obtain the corresponding implementation record, including whether it meets the system requirements and the specific situation of non-compliance. During the inspection process, statistically analyze the number of items that meet the system requirements and the number of items that do not meet the system requirements respectively. Divide the number of items that do not meet the system requirements by the total number of items required by the system to obtain the non-compliance rate of system implementation. Extract relevant information on system training from the system system ledger, including the training course name, training time, number of people who should participate, and actual number of participants. Divide the number of people absent from system training by the total number of employees in the enterprise to obtain the system training absenteeism rate.

[0080] In one embodiment of the present invention, the comprehensive overall quality and safety coefficient obtained by combining the fan defect rate and the comprehensive risk ratio coefficient is compared with the normal overall quality and safety index, and whether to trigger a first-level warning signal is judged according to the comparison result, including the following steps:

[0081] Add the personnel risk ratio coefficient, the fire accident risk ratio coefficient, and the system risk ratio coefficient to obtain the comprehensive risk ratio coefficient, and assign weights to and add the fan defect rate and the comprehensive risk ratio coefficient to obtain the comprehensive overall quality and safety coefficient;

[0082] Comprehensive overall quality and safety coefficient calculation formula:

[0083] S = [F×0.1 + (P p + P f + P s )×0.1]×100%

[0084] Wherein, S represents the comprehensive overall quality and safety coefficient, F represents the fan defect rate, P p represents the personnel risk ratio coefficient, P f represents the fire accident risk ratio coefficient, P s represents the system risk ratio coefficient;

[0085] Determine the normal fan defect rate, as well as the normal personnel risk ratio coefficient, the normal fire accident risk ratio coefficient, and the normal system risk ratio coefficient through the statistical analysis of the comprehensive industry standards and the quality reports of multiple professional enterprises, add them to obtain the normal comprehensive risk ratio coefficient, and assign weights to and add the normal defect rate and the normal comprehensive risk ratio coefficient to obtain the normal overall quality and safety index;

[0086] Compare the comprehensive overall quality and safety coefficient with the normal overall quality and safety index. If the comprehensive overall quality and safety coefficient is less than the normal overall quality and safety index, it means that the comprehensive ledger data is normal and the first-level warning signal is not triggered. If the comprehensive overall quality and safety coefficient is equal to or greater than the normal overall quality and safety index, it means that the comprehensive ledger data is abnormal and the first-level warning signal is triggered.

[0087] Specifically, add the fan defect rate and the comprehensive risk ratio coefficient to obtain the comprehensive overall quality and safety coefficient. Collect the regulations and data on the fan defect rate in relevant comprehensive industry standards, and use statistical methods such as the mean, median, and mode to analyze the central tendency of the data. At the same time, calculate the degree of dispersion of the data. According to the results of the statistical analysis, comprehensively consider the industry requirements and the actual situation of the enterprise to determine the normal defect rate of the fan. According to the method of obtaining the normal defect rate, collect the data on the personnel ledger, fire protection ledger, and system system ledger in the industry standards and enterprise quality reports, determine the normal personnel risk ratio coefficient, fire accident risk ratio coefficient, and system system risk ratio coefficient, add them up to obtain the normal risk ratio coefficient, and add it to the normal defect rate to obtain the normal overall quality and safety index. Compare the comprehensive overall quality and safety coefficient with it. If the comprehensive overall quality and safety coefficient is less than the normal overall quality and safety index, it means that the comprehensive ledger data is normal and does not trigger a first-level warning signal. If the comprehensive overall quality and safety coefficient is equal to or greater than the normal overall quality and safety index, it means that the comprehensive ledger data is abnormal and triggers a first-level warning signal.

[0088] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A total quality management warning and supervision system, characterized in that, Includes the following modules: Production data collection and transmission management module: During the production process, various sensors and intelligent detection equipment are used to collect the operating parameters of production equipment, the appearance quality data of fans and ledger data in real time, and the collected data is transmitted to the server for storage through the data collection program; User identity verification module: defines different roles according to the user's job responsibilities and authority scope, and performs dual authentication by combining the input user name and password with fingerprint recognition. When the authentication succeeds, the user enters the system. When the authentication fails, a three-level warning signal is triggered; Fan defect and ledger data analysis module: Build a defect type classification model, identify and classify defective fans in the production process, build a fan defect rate, obtain a comprehensive index of the production equipment operation status by weighting and adding multiple standardized values, and determine whether to trigger a secondary warning signal. Analyze the personnel ledger, fire ledger, and system ledger data respectively, and calculate the personnel risk ratio coefficient, fire accident risk ratio coefficient, and system risk ratio coefficient; Quality and safety index comparison and early warning module: The comprehensive quality and safety coefficient obtained by combining the fan defect rate and the comprehensive risk ratio coefficient is compared with the normal comprehensive quality and safety index, and the result of the comparison is used to determine whether to trigger a first-level early warning signal.

2. The total quality management warning and supervision system according to claim 1, characterized in that, In the production process, various sensors and intelligent detection equipment are used to collect the operating parameters of the production equipment, the appearance quality data of the fan and the ledger data in real time, including the following steps: Install a variety of sensors at key locations of production equipment, including temperature sensors, pressure sensors, and vibration sensors. Set the corresponding data collection frequency according to the monitoring requirements of the production equipment, and read data from the sensors at the preset frequency in the data collection program; A machine vision inspection system is installed on the production line to obtain the fan appearance image. The data acquisition program obtains the fan appearance quality data from the inspection system, and after associating the fan's production batch number and production time information, the obtained fan image data is pre-processed and analyzed; View and obtain the corresponding data in the personnel ledger, fire ledger and system ledger from the ledger data list; Configure wireless network access points to connect data acquisition terminals to the server. Determine the data transmission protocol based on the type of data acquisition and network architecture. Use encryption technology to encrypt the data before transmitting it to the server for storage.

3. The total quality management early warning and supervision system according to claim 1, characterized in that, The system defines different roles according to the user's job responsibilities and authority scope, and performs dual authentication by inputting the user name and password in combination with fingerprint recognition. When the authentication succeeds, the user enters the system. When the authentication fails, a three-level warning signal is triggered, including the following steps: Define different roles according to the user's job responsibilities and scope of authority. Each role has its specific set of permissions. Assign users to the corresponding roles, and the system monitors the user's access behavior in real time. The user first enters the username and password for the first step of identity verification. The system verifies whether the credentials provided by the user match the stored valid credentials in the database. If the match is successful, it enters the second step of identity verification. If the match fails, the user is not allowed to log in. For the second step of identity verification, the user's fingerprint is scanned by a fingerprint scanning device. The system compares the scanned fingerprint with the pre-stored fingerprint information to determine the user's identity. If the comparison is successful, the user successfully enters the system. If the comparison fails, a third-level warning signal is triggered.

4. A total quality management warning and supervision system according to claim 1, characterized in that, The construction of the defect type classification model to identify and classify defective fans during the production process, and count the number of defective fans and the total number of fans to obtain the fan defect rate, includes the following steps: According to the actual requirements and production situation, determine a production batch. Conduct statistics in units of production batches, establish a batch traceability system, determine the sampling quantity of each batch of fans through statistical methods and sampling standards. After obtaining the samples, use the defect type classification model to detect each fan in the samples, and obtain the defect information of each fan from the detection results, including whether there are defects and the types of defects existing. Record it in the form of a database, count the number of defective fans, use data processing software to separately count the number of fans of different defect types, add up the numbers of each defect type to obtain the total number of fan defects, count the total number of fans within the determined statistical range, and obtain the fan defect rate by dividing the total number of fan defects by the total number of fans; Fan defect rate calculation formula: Among them, F represents the fan defect rate, N pore represents the number of fans with pore defect type, N crack represents the number of fans with crack defect type, N trachoma represents the number of fans with sand hole defect type, N total represents the total number of fans.

5. The total quality management warning and supervision system according to claim 4, characterized in that, The defect type classification model includes the following steps: Obtain fan images containing different defect types from the production process, use image annotation tools to annotate the collected images, mark the defect type or normal category to which each sample belongs, perform preprocessing operations on the images, including data cleaning and image enhancement. Select a model framework based on a convolutional neural network to construct the defect type classification model. Use edge detection algorithms and corner detection algorithms to extract the features in the image data, including the edge and corner features of the image, as the input of the model. Use the annotated image data to train the model, update the model's parameters through multiple iterations, and deploy the trained model to the actual production environment to enable it to automatically identify and classify the defect types on the surface of the fans. Combined with the automated production line, the defect type classification model detects the defects on the surface of the fans in real time for classification.

6. The total quality management early warning and supervision system according to claim 1, characterized in that The step of obtaining the comprehensive index of the operation status of the production equipment by weighted summation of the obtained multiple standardized values and determining whether to trigger a second-level warning signal includes the following steps: Analyze the historical temperature, pressure and vibration data of the production equipment under normal operating conditions, determine the appropriate range of temperature intervals, pressure intervals and vibration intervals from the historical data as the normal temperature range, pressure range and vibration range of the production equipment, and determine the maximum and minimum values ​​of the corresponding production equipment operating parameters; standardize the temperature, pressure and vibration data through the actual temperature value, actual pressure value and actual vibration value obtained by current measurement, and obtain the temperature standardized value, pressure standardized value and vibration standardized value respectively, and obtain the comprehensive index of the production equipment operating status by weighted addition of the multiple standardized values ​​obtained; The calculation formula of the comprehensive index of production equipment operation status is: Among them, E represents the comprehensive index of the operating state of the production equipment, T represents the current actual temperature value, P represents the current actual pressure value, V represents the current actual vibration value, T max represents the maximum temperature value, T min represents the minimum temperature value, P max represents the maximum pressure value, P min represents the minimum pressure value, V max represents the maximum vibration value, V min represents the minimum vibration value, and ω1, ω2, and ω3 respectively represent the corresponding weight coefficients of the standardized values of temperature, pressure, and vibration; When 0≤the comprehensive index of the production equipment operation status≤1, it indicates that the operation status of the production equipment is normal and no secondary warning signal is triggered; when the comprehensive index of the production equipment operation status is <0 or the comprehensive index of the production equipment operation status is >1, it indicates that the operation status of the production equipment is abnormal and a secondary warning signal is triggered.

7. The total quality management early warning and supervision system according to claim 1, characterized in that The method of analyzing the personnel records, fire records and system records data respectively, and calculating the personnel risk ratio coefficient, the fire accident risk ratio coefficient and the system risk ratio coefficient, comprises the following steps: Obtain training record information from the personnel ledger, and track the work performance of employees within a certain period of time after receiving training. Obtain the product input ratio by comparing the total cost of training investment with the total benefits brought by employees after training. Obtain job change information from the personnel ledger, count the number of employees joining and leaving the company in a specific period of time, and obtain the staff turnover rate by the ratio of the number of people leaving and the average number of employees in the company. Add the weighted product input ratio and staff turnover rate to obtain the personnel risk ratio coefficient. Obtain information related to firefighting facilities from the firefighting ledger, regularly check the number and status of firefighting facilities, count the number of intact firefighting facilities and the number of damaged firefighting facilities, and obtain the firefighting facility damage rate by the ratio of the number of damaged firefighting equipment to the total number of firefighting facilities. , obtain relevant information about fire accidents in the fire protection ledger, count the number of fire accidents that occur in enterprises within a certain period of time to obtain the fire incidence rate, and add the fire protection facility damage rate and the fire incidence rate to obtain the fire accident risk ratio coefficient; obtain relevant information in the system ledger, check the implementation of various systems, count the number of matters that meet the system requirements and the number of matters that do not meet the system requirements, and obtain the system implementation non-compliance rate through the number of matters that do not meet the system requirements and the number of matters required by all systems, obtain the training record information in the system system ledger, count the number of people who missed the system training and the total number of people in the enterprise, calculate the system training absence rate, and add the system implementation non-compliance rate and the system training absence rate to obtain the system risk ratio coefficient.

8. A total quality management early warning and supervision system according to claim 1, characterized in that The comprehensive overall quality safety coefficient obtained by combining the fan defect rate and the comprehensive risk ratio coefficient is compared with the normal overall quality safety index, and judging whether to trigger a first-level warning signal according to the comparison result includes the following steps: Add the personnel risk ratio coefficient, the fire accident risk ratio coefficient, and the institutional system risk ratio coefficient to obtain the comprehensive risk ratio coefficient, and assign weights to the fan defect rate and the comprehensive risk ratio coefficient and add them to obtain the comprehensive overall quality and safety coefficient; Comprehensive overall quality and safety coefficient calculation formula: S = [F × 0.1+(P p +P f +P s ) × 0.1] × 100% Among them, S represents the comprehensive overall quality and safety coefficient, F represents the fan defect rate, P p represents the personnel risk proportion coefficient, P f represents the fire accident risk proportion coefficient, P s represents the system risk proportion coefficient; Determine the normal fan defect rate, as well as the normal personnel risk ratio coefficient, the normal fire accident risk ratio coefficient, and the normal institutional system risk ratio coefficient through the comprehensive industry standards and the statistical analysis of multiple professional enterprise quality reports, add them to obtain the normal comprehensive risk ratio coefficient, and assign values to the normal defect rate and the normal comprehensive risk ratio coefficient and add them to obtain the normal overall quality and safety index; Compare the comprehensive overall quality and safety coefficient with the normal overall quality and safety index. If the comprehensive overall quality and safety coefficient is less than the normal overall quality and safety index, it means that the comprehensive ledger data is normal and does not trigger a first-level warning signal. If the comprehensive overall quality and safety coefficient is equal to or greater than the normal overall quality and safety index, it means that the comprehensive ledger data is abnormal and triggers a first-level warning signal.

Images