Digital measuring and calculating method and measuring and calculating system for product maintenance

Abstract

The invention provides a digital measurement and calculation method and system for product maintenance, and relates to the technical field of equipment maintenance, and the method comprises the steps: recording product defect data and product maintenance data, judging whether a current product is maintained for the first time or not, and if the current product is maintained for the first time, uploading the data to a product maintenance management terminal; the management terminal matches a digital measurement and calculation model according to the product defect data and the product maintenance data, and downloads the digital measurement and calculation model to the user mobile terminal; monitoring a state change data set of the maintained product in real time; and the user mobile terminal calls the digital measurement and calculation model to carry out digital measurement and calculation on the state change data set, and outputs a next maintenance measurement and calculation result. The technical problem of low prediction accuracy caused by the fact that the prior art is only based on a general statistical model or a fixed maintenance rule and does not consider historical maintenance data and a defect evolution trend of the product is solved, and the technical effects of improving the maintenance prediction accuracy, improving the product reliability and reducing the maintenance cost are achieved.

Description

Technical Field

[0001] This application relates to the technical field of equipment maintenance, and particularly to a digital calculation method and system for product maintenance. Background Art

[0002] In the modern product usage cycle, maintenance is a key link to ensure stable product performance and extend service life. With the development of technology, product maintenance prediction methods are gradually shifting from traditional experience-based maintenance to data-driven predictive maintenance. Existing maintenance prediction methods mainly rely on general statistical evaluations or fixed rules. By analyzing historical data and combining indicators such as the average failure rate of products, maintenance plans are formulated. However, factors such as the environment and usage intensity experienced by each product during actual use are different, resulting in differences in the types of defects generated and their development trends. After a product undergoes maintenance, its operating state may change, and original defects may be repaired, but new defects may also gradually emerge. Existing general methods cannot fully consider these individual differences, leading to low accuracy in maintenance prediction. This low accuracy may not only result in waste of maintenance resources but also cause equipment failures due to failure to detect potential problems in a timely manner, increasing maintenance costs and downtime and reducing overall operational efficiency. Summary of the Invention

[0003] This application provides a digital calculation method and system for product maintenance, which solves the technical problem of low prediction accuracy in the prior art due to only relying on general statistical models or fixed maintenance rules without considering the historical maintenance data and defect evolution trends of specific products, and achieves the technical effect of improving the accuracy of maintenance prediction, thereby improving product reliability and reducing maintenance costs.

[0004] In view of the above problems, on the one hand, this application provides a digital calculation method for product maintenance. The method includes: recording product defect data and product maintenance data, and determining whether the current product is for the first maintenance. If the current product is for the first maintenance, uploading the product defect data and the product maintenance data to a product maintenance management terminal, where the product maintenance management terminal includes multiple digital calculation models; the product maintenance management terminal performs digital calculation model matching according to the product defect data and the product maintenance data, outputs a matching digital calculation model, and downloads the matching digital calculation model to a user mobile terminal; real-time monitoring the state change data set of the product after maintenance; the user mobile terminal calls the matching digital calculation model to perform digital calculation on the state change data set, and outputs the next maintenance calculation result, where the next maintenance calculation result includes the defect type and the maintenance time.

[0005] On the other hand, the present application also provides a digital measurement system for product maintenance. The system includes: a maintenance data upload module for recording product defect data and product maintenance data, and determining whether the current product is for the first maintenance. If the current product is for the first maintenance, the product defect data and the product maintenance data are uploaded to a product maintenance management terminal, where the product maintenance management terminal includes multiple digital measurement models; a measurement model matching module for digitally matching measurement models through the product maintenance management terminal according to the product defect data and the product maintenance data, outputting a matched digital measurement model, and downloading the matched digital measurement model to a user mobile terminal; a status change monitoring module for real-time monitoring of a status change data set of the product after maintenance; and a maintenance measurement module for digitally measuring the status change data set by invoking the matched digital measurement model through the user mobile terminal, and outputting a next maintenance measurement result, where the next maintenance measurement result includes a defect type and a maintenance time.

[0006] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0007] By recording product defect data and maintenance data, it provides data support for subsequent model matching. Determining whether it is the first maintenance helps to distinguish the maintenance requirements at different stages, provides a basis for personalized measurement, and ensures that the subsequent measurement model can be accurately optimized for the current state of the product. After uploading the data to the management terminal, multiple digital measurement models in the terminal are used for matching. Through data-driven model selection, the most suitable measurement model for the current product state can be found, thereby improving the accuracy of prediction. Downloading the matched measurement model to the user mobile terminal extends the intelligent function of the management terminal to the user side, enabling the user to perform maintenance measurement at any time and anywhere using the mobile terminal, enhancing the user's control ability over the maintenance process, and enabling the user to timely understand the product status and future maintenance requirements. By real-time monitoring the status of the product after maintenance and continuously collecting the status change data of the product after maintenance, the dynamic behavior of the product can be timely captured. This real-time data provides the latest information for subsequent dynamic measurement, ensuring that the maintenance prediction can reflect the actual state of the product. The user mobile terminal invokes the matched model for measurement, combines the real-time monitored data with the matched measurement model, and outputs the measurement result of the next maintenance, including the defect type and the maintenance time. This not only realizes the dynamic and personalized maintenance prediction, but also enables the user to take timely actions and optimize the maintenance plan through the convenience of the mobile terminal.

[0008] In summary, through the data-driven personalized measurement model matching, real-time monitoring and dynamic measurement, and the convenience of mobile terminals, the present application realizes the accurate prediction and dynamic management of product maintenance, significantly improving the accuracy of maintenance prediction. Such accurate prediction can not only detect potential problems in advance, reduce the occurrence of equipment failures, thereby improving product reliability, but also optimize the allocation of maintenance resources and reduce maintenance costs.

[0009] The above description is only an overview of the technical solution of the present application. In order to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically illustrates the specific implementation manners of the present application. Brief Description of the Drawings

[0010] Figure 1 It is a schematic flowchart of the digital measurement method for product maintenance provided by an embodiment of the present application.

[0011] Figure 2 It is a schematic flowchart of the process of training and obtaining multiple digital measurement models in the digital measurement method for product maintenance provided by an embodiment of the present application.

[0012] Figure 3 It is a schematic flowchart of the process of matching digital measurement models in the digital measurement method for product maintenance provided by an embodiment of the present application.

[0013] Figure 4 It is a schematic structural diagram of the digital measurement system for product maintenance provided by an embodiment of the present application.

[0014] Description of the reference numerals: Maintenance data upload module 10, measurement model matching module 20, status change monitoring module 30, maintenance measurement module 40. Detailed Description of the Embodiments

[0015] By providing a digital measurement method and a measurement system for product maintenance in the embodiments of the present application, the technical problem in the prior art that the prediction accuracy is low because only a general statistical model or fixed maintenance rules are used without considering the historical maintenance data and defect evolution trend of specific products is solved, and the technical effect of improving the maintenance prediction accuracy, thereby improving product reliability and reducing maintenance costs is achieved.

[0016] Embodiment 1, as Figure 1 shown, the embodiments of the present application provide a digital measurement method for product maintenance, and the method includes:

[0017] Step S1: Record the product defect data and the product maintenance data, and determine whether the current product is under the first maintenance. If the current product is under the first maintenance, upload the product defect data and the product maintenance data to the product maintenance management terminal, where the product maintenance management terminal includes multiple digital measurement models.

[0018] Specifically, the product defect data refers to the specific records of the faults, abnormalities or performance degradation that occur during the use of the product. For example, an engine may have defects such as overheating and abnormal vibration, and these data can be obtained through sensors or manual inspection records. The product maintenance data refers to the maintenance and repair records of the product, including the repair time, repair method, replaced parts, diagnosis by the repair personnel, etc. When the product fails or undergoes maintenance, the product defect data and the product maintenance data are collected through Internet of Things (IoT) sensors or by technicians manually entering the repair logs. Determine whether the current product is under the first maintenance based on the product maintenance data, that is, whether the current product is the first maintenance in the product life cycle. If it is the first maintenance, then upload the defect data and the maintenance data to the product maintenance management terminal, which is a centralized data processing and analysis platform integrating multiple digital measurement models for storing, calculating and managing maintenance-related data.

[0019] By recording and uploading the product defect data and the product maintenance data, it provides the basic data for the entire maintenance prediction process, enabling the subsequent digital measurement models to obtain accurate initial information.

[0020] Step S2: The product maintenance management terminal performs digital measurement model matching based on the product defect data and the product maintenance data, outputs the matching digital measurement model, and downloads the matching digital measurement model to the user mobile terminal.

[0021] Specifically, the multiple digital measurement models in the product maintenance management terminal are used to analyze the product status under different conditions and predict the maintenance requirements. After receiving the product defect data and the product maintenance data, the product maintenance management terminal performs matching among the multiple digital measurement models it contains based on the data characteristics of these data, compares the applicable ranges and data requirements of each model, and finds the model that is most suitable for the current product's product defect data and product maintenance data, and outputs it as the matching digital measurement model. For example, if the product defect data shows that the wear rate of a certain component of the product is relatively fast, the management terminal will look for a digital measurement model that is good at analyzing the wear of this component. Finally, download this matching digital measurement model to the user mobile terminal, which are devices used by users to query maintenance information and calculate prediction results, such as mobile phones, tablets, industrial handheld terminals, etc.

[0022] Through digital measurement model matching, the most suitable model for the current product situation is found from numerous measurement models and downloaded to the user's mobile terminal, providing a suitable tool for subsequent measurement on the mobile terminal and improving the accuracy and pertinence of the measurement.

[0023] Step S3: Real-time monitor the dataset of the status changes of the product after maintenance.

[0024] Specifically, the status changes of the product after maintenance are monitored in real time by installing sensors on the product or using the software monitoring function of the product itself, and the dataset of the status changes of the product after maintenance is collected. This dataset of status changes is a set of data reflecting the changes in various status indicators of the product after maintenance during use. For industrial equipment, dedicated monitoring sensors such as pressure sensors and flow sensors can be installed to monitor the equipment status in real time. These sensors will obtain the operation status data of the equipment in real time and organize these data into a dataset of status changes. For consumer electronics products such as smart watches, the internal chips and software systems can automatically monitor changes in data such as heart rate monitoring data and movement step count data, and use these data as part of the dataset of status changes.

[0025] By real-time monitoring the dataset of the status changes of the product after maintenance, the actual status information of the product after maintenance can be obtained in a timely manner, providing an accurate data basis for the next measurement and helping to more accurately predict the next maintenance situation.

[0026] Step S4: The user's mobile terminal calls the matched digital measurement model to perform digital measurement on the dataset of the status changes and outputs the measurement result of the next maintenance, where the measurement result of the next maintenance includes the defect type and the maintenance time.

[0027] Specifically, after receiving the dataset of the status changes, the user's mobile terminal calls the previously downloaded matched digital measurement model to perform digital measurement on it and outputs the measurement result of the next maintenance. This measurement result of the next maintenance is the relevant result regarding the next product maintenance, including the defect type (such as a certain component of a mechanical product being worn or a certain functional module of an electronic product malfunctioning) and the maintenance time (for example, it is estimated that the next maintenance is required after how many days or after the equipment has run for how many hours). For example, if it is a digital measurement model based on a neural network, the mobile terminal will input the dataset of the status changes into the model, and the model will calculate the result through the calculation of multiple layers of neurons. For a simple linear regression model, the relevant result of the next maintenance will be calculated based on the relationship between the variables in the dataset.

[0028] By calling the matching digital measurement model to measure the status change data set, the next maintenance measurement result is obtained, thus realizing the accurate prediction of the next maintenance situation of the product, which helps to arrange the maintenance work in advance, improve the product reliability and reduce the maintenance cost.

[0029] Further, if the current product is not for the first maintenance, the product defect data and the product maintenance data are uploaded to the matching digital measurement model for incremental learning to obtain an updated matching digital measurement model; the status change data set of the product after maintenance is monitored in real time; the user mobile terminal calls the updated matching digital measurement model to perform digital measurement on the status change data set and outputs the next maintenance measurement result.

[0030] Specifically, if the current product is not for the first maintenance, the product defect data and the maintenance data are uploaded to the previously matched digital measurement model. Based on the existing model, new data is used for incremental learning to further optimize the model. The digital measurement model can use the optimizer and learning rate adjustment strategy provided by the deep learning framework (such as TensorFlow or PyTorch) to achieve incremental learning. By learning the new data, the parameters of the model are updated, thus obtaining an updated matching digital measurement model. Then, the status change data set of the product after maintenance is monitored in real time. The implementation method is similar to the previous steps, relying on the sensors or software monitoring functions on the product to obtain data and form a status change data set. The user mobile terminal calls the updated matching digital measurement model to perform digital measurement on the status change data set, thereby outputting the next maintenance measurement result, including possible defect types and recommended maintenance times.

[0031] For products that are not for the first maintenance, incremental learning is carried out through new data, so that the matching digital measurement model can continuously adapt to the actual situation changes of the product, better capture different status change patterns generated as the number of maintenance increases, so as to improve the accuracy and generalization ability of the model and more accurately predict the next maintenance situation. Monitoring the status change data set in real time and using the updated model for measurement can further improve the timeliness and accuracy of the prediction, thus better arranging the maintenance plan and reducing the maintenance cost.

[0032] Further, downloading the matching digital measurement model to the user mobile terminal in step S2 further includes:

[0033] An edge computing network between the product maintenance management terminal and multiple user mobile terminals is established, where each user mobile terminal serves as an edge computing node to establish an edge transmission channel with the product maintenance management terminal; according to the edge computing network, when the product maintenance management terminal obtains the matching digital measurement model, it is downloaded to the user mobile terminal based on the corresponding edge transmission channel.

[0034] Specifically, the edge computing network is a network architecture composed of multiple edge computing nodes (i.e., user mobile terminals) and a central product maintenance management terminal. To establish an edge computing network, it is first necessary to determine the user mobile terminals participating in the edge computing network, which requires identifying and registering the user mobile terminals. For example, a unique identifier (such as a device ID) is assigned to each mobile terminal. Then, network connection parameters are configured on the product maintenance management terminal to identify and connect these mobile terminals, and a separate transmission channel, namely the edge transmission channel, is established between each user mobile terminal and the product maintenance management terminal. For example, the Wi-Fi direct connection protocol or a custom dedicated network protocol is used to ensure a stable connection can be established between each mobile terminal (as an edge computing node) and the product maintenance management terminal, thereby constructing the edge computing network. When the product maintenance management terminal obtains a matching digital measurement model, the model is downloaded to the user mobile terminal through the established corresponding edge transmission channel.

[0035] Through the edge computing network, the model is directly deployed to the user mobile terminal, reducing data transmission latency, improving the real-time performance and accuracy of maintenance prediction. At the same time, the edge computing architecture can effectively reduce the burden on the central server, optimize system resource utilization, and thus quickly perform subsequent maintenance measurement work.

[0036] Furthermore, as Figure 2 shown, the product maintenance management terminal includes multiple digital measurement models, and each digital measurement model is obtained through training. The training methods include:

[0037] Step S1-1: Collect historical record data of products of the same type as the current product, including product defect historical data, product maintenance historical data, and product status changes corresponding to each product. The product defect historical data includes defect type, location, severity, and occurrence frequency. The product maintenance historical data includes maintenance operation content, maintenance cycle, resource consumption, and maintenance effect.

[0038] Step S1-2: According to the historical record data, identify the initial product defect data and initial product maintenance data of the first maintenance of each product.

[0039] Step S1-3: Classify the historical record data according to the initial product defect data and initial product maintenance data of the first maintenance of each product to obtain multiple groups of historical record data.

[0040] Step S1-4: Perform model training respectively according to the multiple groups of historical record data to obtain multiple digital measurement models.

[0041] Specifically, historical record data of products of the same type as the current product are collected from multiple data sources, including product defect historical data, product maintenance historical data, and product status changes. These data reflect various situations during the use of the product and are the basis for model training. For example, for a manufacturing enterprise, relevant data during the product production process can be obtained from the production management system, and product repair records can be obtained from the after-sales repair department. Among them, the product defect historical data records the defect type, location, severity, and occurrence frequency. For example, for a certain device, the motor overheating fault, the location is the motor, the severity is medium, and the occurrence frequency is once a quarter. The product maintenance historical data records the operation content of maintenance, maintenance cycle, resource consumption, and maintenance effect. For example, a certain device has its filter replaced once a quarter, the maintenance time is 2 hours, the resource consumption is 1 filter replacement, and the maintenance effect is good. The product status changes record the changes in the operating state of the device after maintenance, such as changes in parameters such as temperature and pressure. By collecting comprehensive historical record data of the same type of products, various situations from product production to use can be covered, providing rich materials for model training.

[0042] Analyze the historical record data of each collected product. Based on the historical record data, identify the initial product defect data and initial product maintenance data for each product, that is, the defect data and maintenance data recorded during the first maintenance of the product. These data usually reflect the state of the product in the initial use stage. For example, according to the time sequence or the usage sequence of the product, filter out the data of the first defect and the first maintenance of each product. According to the similarities of the initial product defect data and initial product maintenance data of each product, classify the historical record data, group the data with similar characteristics, and obtain multiple groups of historical record data for targeted model training. Identifying and classifying the initial product defect data and initial product maintenance data makes the training data more targeted and can better reflect the impact of different initial situations on the subsequent maintenance of the product.

[0043] Use machine learning algorithms (such as linear regression, decision tree, neural network, etc.) to train each set of historical record data to obtain multiple digital measurement models. For example, if there is more numerical data to be processed, regression algorithms (such as linear regression, decision tree regression, etc.) can be used; if there is more categorical data to be processed (such as the defect type being categorical data), classification algorithms (such as support vector machine, naive Bayes, etc.) can be used. Take the product defect history data, product maintenance history data, and product status change data in each set of historical record data as inputs, and use a machine learning framework (such as Scikit-learn) to implement the model training process, and train a digital measurement model that can predict the maintenance situation. For example, for a group of products with overheating motors, train a dedicated fault prediction model that can predict future fault types and maintenance times based on real-time data.

[0044] By collecting and analyzing historical data, identifying the initial maintenance data, classifying the data, and training the model, multiple digital measurement models for different initial states and change trends can be generated, which not only improves the personalization and accuracy of the models, but also better adapts to the maintenance needs of different types of products.

[0045] Further, step S1-3 includes:

[0046] Step S1-31: Calculate the defect similarity and maintenance similarity between products based on the initial product defect data and initial product maintenance data of the initial maintenance of each product.

[0047] Step S1-32: Use hierarchical clustering to cluster the products with both the defect similarity and the maintenance similarity greater than the preset similarity threshold to obtain multiple groups of clustered products, and output multiple groups of historical record data corresponding to the multiple groups of clustered products.

[0048] Specifically, calculate the defect similarity and maintenance similarity between products based on the initial product defect data and initial product maintenance data of the initial maintenance of each product. Among them, the defect similarity is used to measure the similarity degree of the initial product defects between different products, and is calculated by comparing indicators such as defect type, location, severity, and occurrence frequency. The maintenance similarity is used to measure the similarity degree of the initial maintenance between different products, and is calculated by comparing indicators such as maintenance operation content, maintenance cycle, resource consumption, and maintenance effect. The preset similarity threshold is a pre-set minimum similarity value used to determine whether two products are similar enough to be grouped into the same cluster.

[0049] For each product, extract the defect data (such as defect type, location, severity, and occurrence frequency) at the first maintenance and perform quantification processing, and calculate the similarity with other products. For example, for defect types, different numerical values can be assigned according to the severity or occurrence frequency of different types. Then, according to the selected similarity calculation method (such as Euclidean distance, cosine similarity, etc.), calculate the initial product defect data of different products. Taking Euclidean distance as an example, each factor in the initial product defect data of two products is regarded as a dimension of a vector, and the distance between the vectors is calculated. The smaller the distance, the higher the defect similarity. For the calculation of maintenance similarity, similarly, quantify or encode factors such as maintenance operation content, maintenance cycle, resource consumption, and maintenance effect in the initial product maintenance data. Then calculate the maintenance similarity between different products according to the selected similarity calculation method. For example, the maintenance content of two devices is to replace the filter element, the maintenance cycles are "once a quarter" and "once every six months" respectively, and the resource consumptions are "1 filter element replacement" and "2 filter element replacements" respectively. By calculating the similarity of these features, the maintenance similarity value is obtained.

[0050] According to the calculated defect similarity and maintenance similarity, use the hierarchical clustering algorithm to cluster the products. Hierarchical clustering will gradually merge products with similarity higher than the preset threshold to form multiple clusters. During the clustering process, each product is regarded as a data point, and its corresponding defect similarity and maintenance similarity are used as the features of the data point. Then, starting from each product as a separate class, calculate the distance between classes according to the defect similarity and maintenance similarity (such as using distance calculation methods such as single-linkage, complete-linkage, or average-linkage). Continuously merge the classes with the closest distance (i.e., the highest similarity) and meeting the preset similarity threshold until all products are clustered into appropriate groups. For example, set the defect similarity threshold to 0.8 and the maintenance similarity threshold to 0.7. Through the hierarchical clustering algorithm, products meeting these conditions are grouped into the same cluster. After hierarchical clustering is completed, multiple groups of clustered products are obtained, and each group of clustered products corresponds to a set of historical record data. These data will be used for subsequent model training. For example, after clustering, three groups of products are obtained, each group of products has similar defect and maintenance characteristics, and the corresponding three sets of historical record data will be used to train three different digital measurement models respectively.

[0051] By calculating the defect similarity and maintenance similarity, the potential similarity between products can be mined from the initial product defect data and the initial product maintenance data. This similarity reflects the common characteristics of products in terms of the first maintenance. Using hierarchical clustering to cluster products according to these similarities can group products with similar first maintenance characteristics into one group. This makes the historical record data of products within the same group have higher relevance and consistency in subsequent model training, and can improve the efficiency and accuracy of model training.

[0052] Further, as Figure 3 shown, the product maintenance management terminal in step S2 performs digital measurement model matching according to the product defect data and the product maintenance data, including:

[0053] Step S21: The product maintenance management terminal obtains multiple product defect features and multiple product maintenance features corresponding to the multiple digital measurement models.

[0054] Step S22: Calculate multiple defect similarities and multiple maintenance similarities between the product defect data and the product maintenance data and the multiple product defect features and multiple product maintenance features.

[0055] Step S23: Obtain multiple comprehensive similarities according to the multiple defect similarities and the multiple maintenance similarities.

[0056] Step S24: Based on the multiple comprehensive similarities, select a matching digital measurement model from the multiple digital measurement models, and the comprehensive similarity of the matching digital measurement model is the highest among the multiple comprehensive similarities.

[0057] Specifically, when performing model matching, the product maintenance management terminal extracts the product defect features and product maintenance features corresponding to each model from the stored multiple digital measurement models. These features are the key attributes extracted from historical data during model training. Among them, the product defect feature is the typical feature of the product defect aspect targeted by each digital measurement model; the product maintenance feature is the feature related to the product maintenance aspect of each digital measurement model. For example, the defect feature of model A is "motor overheating, location: motor, severity: medium, occurrence frequency: once per quarter", and the maintenance feature is "maintenance content: replace coolant, maintenance cycle: once per quarter, resource consumption: 1 liter of coolant, maintenance effect: good".

[0058] For the defect data and maintenance data of the current product, use a similarity calculation method (such as cosine similarity or Euclidean distance) to calculate the similarity between the current data and the defect features and maintenance features of each model respectively, and determine multiple groups of defect similarities and maintenance similarities. Among them, each group of defect similarities and maintenance similarities corresponds to a digital measurement model.

[0059] Calculate the comprehensive similarity of each model based on the defect similarity and the maintenance similarity. The calculation of the comprehensive similarity can adopt the method of weighted average. Determine the weights of the defect similarity and the maintenance similarity according to the business requirements. For example, if it is considered that the defect similarity is more important when matching the model, the weight of the defect similarity can be set to 0.6, and the weight of the maintenance similarity can be set to 0.4. Then, for each digital measurement model, calculate its comprehensive similarity according to the weights, that is, comprehensive similarity = defect similarity × 0.6 + maintenance similarity × 0.4.

[0060] After calculating the comprehensive similarities of all digital measurement models, compare the values of these comprehensive similarities, find the digital measurement model corresponding to the highest comprehensive similarity, and output it as the matching digital measurement model.

[0061] By calculating the defect similarity and the maintenance similarity, and selecting the most matching digital measurement model in combination with the comprehensive similarity, it is possible to select the most suitable model for maintenance prediction according to the specific characteristics of the current product, thereby improving the accuracy and reliability of the prediction.

[0062] Furthermore, step S4 includes:

[0063] Step S41: Extract features from the state change data set to obtain a feature vector set, where the state change data set includes changes in the operating state of mechanical equipment, changes in electrical equipment parameters, and changes in software systems.

[0064] Step S42: Perform digital measurement on the feature vector set based on the matching digital measurement model, and output the next maintenance measurement result.

[0065] Specifically, extract features from various information included in the state change data set (such as changes in the operating state of mechanical equipment, changes in electrical equipment parameters, and changes in software systems, etc.), represent them in the form of vectors, and form a feature vector set. Each vector in the feature vector set contains numerical values or symbols that can represent the corresponding state change characteristics, and these feature vectors can be used as the input of the subsequent digital measurement model. For example, for changes in the operating state of mechanical equipment, the feature vector includes numerical features such as the vibration frequency and temperature change of the equipment; for changes in electrical equipment parameters, the feature vector includes the change values of parameters such as voltage and current; for changes in software systems, the feature vector includes features such as the response time and memory occupancy change of the software. By extracting features from the state change data set containing multiple devices and systems, the complex state change information can be transformed into a feature vector set that is convenient for model processing, effectively integrating information from different sources, removing redundant information, and retaining the features valuable for maintenance measurement.

[0066] Exemplarily, for the changes in the operating state of mechanical equipment, the data collected by sensors can be used to obtain relevant features. For example, vibration sensors installed on mechanical equipment can obtain data such as vibration frequency and amplitude, and temperature sensors can obtain the temperature data of the equipment. After these data are processed (such as filtering, normalization, etc.), they form a part of the feature vector representing the changes in the operating state of mechanical equipment. For some complex operating states of mechanical equipment, time-domain analysis or frequency-domain analysis may also be required. For example, by performing a Fourier transform on the vibration signal, features in the frequency domain can be obtained, such as the energy distribution of different frequency components, and these features are added to the feature vector. For the changes in electrical equipment parameters, the change data of parameters such as voltage, current, and resistance are directly obtained from the monitoring system of the electrical equipment. Then, according to specific requirements, these data are processed, such as calculating the power factor (calculated through voltage, current, and phase difference), and these processed parameters are used as the feature vector of the part of the changes in electrical equipment parameters. For the changes in software systems, data are obtained from the software's log files or performance monitoring tools. For example, the response time of the software can be extracted from the log file, and the change in memory occupancy can be obtained through the performance monitoring tool. These data are quantified and sorted to form the feature vector of the part of the changes in software systems. Finally, the feature vectors from mechanical equipment, electrical equipment, and software systems are combined together to form a complete set of feature vectors.

[0067] Taking the obtained set of feature vectors as input and inputting it into the matching digital measurement and calculation model, the model processes the set of feature vectors according to the internal algorithms and parameters, and calculates the next maintenance measurement result. For example, for a digital measurement and calculation model based on machine learning algorithms (such as decision trees, neural networks, etc.), it will calculate based on the structure of the model and the weights obtained through training for the input feature vectors, and finally obtain results such as the time of the next maintenance, the possible types of defects, and the maintenance resources required. Digital measurement and calculation based on the matching digital measurement and calculation model for the set of feature vectors can make full use of the prediction ability of the model, accurately output the next maintenance measurement result, help plan maintenance work in advance, improve the reliability and service life of the equipment, and reduce the operation and maintenance costs.

[0068] Through feature extraction and digital measurement and calculation, key information can be extracted from complex real-time data, and accurate prediction can be made using the matching model, which not only improves the accuracy of maintenance prediction but also can reflect the operating state of the equipment in real time, helping users take measures in advance to avoid equipment failures.

[0069] In summary, the digital measurement and calculation method for product maintenance provided by the embodiments of the present application has the following technical effects:

[0070] By collecting product defect data and maintenance data, and judging whether it is the first maintenance, the foundation for subsequent personalized prediction is laid. For products that are undergoing maintenance for the first time, the data is uploaded to the management terminal and matched using multiple digital measurement models to ensure the pertinence of the prediction model. For products that are not undergoing maintenance for the first time, the prediction accuracy is further optimized by updating the model through incremental learning. By establishing an edge computing network, the matching model is efficiently downloaded to the user's mobile terminal, and the advantages of edge computing are used to reduce data transmission delays and improve real-time performance. In the model training stage, by collecting historical data, identifying initial maintenance data, classifying data, and performing hierarchical clustering, multiple measurement models for different initial states and change trends are generated, further improving the personalization and accuracy of the model. In the prediction stage, through feature extraction and digital measurement, the equipment status is monitored in real time and the measurement results of the next maintenance are output, helping users to arrange maintenance work in advance and avoid equipment failures.

[0071] Overall, the embodiments of the present application achieve accurate prediction and dynamic management of product maintenance through data-driven personalized measurement model matching, real-time monitoring and dynamic measurement, and the convenience of mobile terminals, significantly improving the accuracy of maintenance predictions, thereby helping to optimize the allocation of maintenance resources, reduce maintenance costs, and extend product life.

[0072] Embodiment 2, as Figure 4 As shown, based on the same inventive concept as the above-mentioned embodiment 1, the embodiment of the present application provides a digital measurement system for product maintenance, the system comprising:

[0073] The maintenance data uploading module 10 is used to record product defect data and product maintenance data, and determine whether the current product is undergoing maintenance for the first time. If the current product is undergoing maintenance for the first time, the product defect data and the product maintenance data are uploaded to the product maintenance management terminal, wherein the product maintenance management terminal includes multiple digital measurement models.

[0074] The calculation model matching module 20 is used to perform digital calculation model matching according to the product defect data and the product maintenance data through the product maintenance management terminal, output a matching digital calculation model, and download the matching digital calculation model to the user mobile terminal.

[0075] The state change monitoring module 30 is used to monitor the state change data set of the product after maintenance in real time.

[0076] The maintenance calculation module 40 is used to call the matching digital calculation model through the user mobile terminal to perform digital calculation on the state change data set and output the next maintenance calculation result, wherein the next maintenance calculation result includes the defect type and maintenance time.

[0077] Further, the maintenance data upload module 10 in the embodiment of the present application is further configured to perform the following steps:

[0078] If the current product is not for the first maintenance, upload the product defect data and the product maintenance data to the matching digital measurement model for incremental learning to obtain an updated matching digital measurement model; real-time monitor the state change data set of the product after maintenance; the user mobile terminal calls the updated matching digital measurement model to perform digital measurement on the state change data set and outputs the next maintenance measurement result.

[0079] Further, the measurement model matching module 20 in the embodiment of the present application is further configured to perform the following steps:

[0080] Establish an edge computing network between the product maintenance management terminal and multiple user mobile terminals, where each user mobile terminal serves as an edge computing node to establish an edge transmission channel with the product maintenance management terminal; according to the edge computing network, when the product maintenance management terminal obtains the matching digital measurement model, download it to the user mobile terminal based on the corresponding edge transmission channel.

[0081] Further, the system in the embodiment of the present application further includes a measurement model training module, and the measurement model training module is used to perform the following steps:

[0082] Collect historical record data of products of the same type as the current product, including product defect historical data, product maintenance historical data, and product state changes corresponding to each product. The product defect historical data includes defect type, location, severity, and occurrence frequency. The product maintenance historical data includes maintenance operation content, maintenance cycle, resource consumption, and maintenance effect; according to the historical record data, identify the initial product defect data and initial product maintenance data for the first maintenance of each product; classify the historical record data according to the initial product defect data and initial product maintenance data for the first maintenance of each product to obtain multiple groups of historical record data; perform model training on the multiple groups of historical record data respectively to obtain multiple digital measurement models.

[0083] Further, the measurement model training module is further configured to perform the following steps:

[0084] According to the initial product defect data and initial product maintenance data for the first maintenance of each product, calculate the defect similarity and maintenance similarity between products; use hierarchical clustering to cluster products with both the defect similarity and the maintenance similarity greater than a preset similarity threshold to obtain multiple groups of clustered products, and output the multiple groups of historical record data corresponding to the multiple groups of clustered products.

[0085] Further, the measurement model matching module 20 in the embodiments of the present application is further configured to perform the following steps:

[0086] The product maintenance management terminal obtains a plurality of product defect features and a plurality of product maintenance features corresponding to the plurality of digital measurement models; calculates a plurality of defect similarities and a plurality of maintenance similarities between the product defect data and the product maintenance data and the plurality of product defect features and the plurality of product maintenance features; obtains a plurality of comprehensive similarities according to the plurality of defect similarities and the plurality of maintenance similarities; and selects a matching digital measurement model from the plurality of digital measurement models based on the plurality of comprehensive similarities, where the comprehensive similarity of the matching digital measurement model is the highest among the plurality of comprehensive similarities.

[0087] Further, the maintenance measurement module 40 in the embodiments of the present application is further configured to perform the following steps:

[0088] Extract features from the state change data set to obtain a feature vector set, where the state change data set includes changes in the operating state of mechanical equipment, changes in electrical equipment parameters, and changes in software systems; perform digital measurement on the feature vector set based on the matching digital measurement model, and output the next maintenance measurement result.

[0089] Through the foregoing detailed description of the digital measurement method for product maintenance in this specification, those skilled in the art can clearly know the digital measurement system for product maintenance in this embodiment. For the system disclosed in Embodiment 2, since it corresponds to the method disclosed in Embodiment 1, it has corresponding functional modules and beneficial effects. For the relevant parts, reference can be made to the description in the method part.

[0090] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A digital measurement method for product maintenance, characterized in that: The method comprises: Record product defect data and product maintenance data, and determine whether the current product is undergoing initial maintenance. If the current product is undergoing initial maintenance, upload the product defect data and product maintenance data to a product maintenance management terminal, wherein the product maintenance management terminal includes multiple digital measurement models; The product maintenance management terminal matches the digital measurement model according to the product defect data and the product maintenance data, outputs the matching digital measurement model, and downloads the matching digital measurement model to the user mobile terminal; Real-time monitoring of the status change data set of the product after maintenance; The user mobile terminal calls the matching digital measurement model to perform digital measurement on the state change data set, and outputs the next maintenance measurement result, wherein the next maintenance measurement result includes the defect type and the maintenance time.

2. The digital measurement method for product maintenance according to claim 1, characterized in that: If the current product is not under initial maintenance, the product defect data and the product maintenance data are uploaded to the matching digital measurement model for incremental learning to obtain an updated matching digital measurement model; Real-time monitoring of the status change data set of the product after maintenance; The user mobile terminal calls the updated matching digital calculation model to perform digital calculation on the state change data set and outputs the next maintenance calculation result.

3. The digital measurement method for product maintenance according to claim 1, characterized in that: Establishing an edge computing network of the product maintenance management terminal and multiple user mobile terminals, wherein each user mobile terminal serves as an edge computing node to establish an edge transmission channel with the product maintenance management terminal; According to the edge computing network, after the product maintenance management terminal obtains the matching digital measurement model, it is downloaded to the user mobile terminal based on the corresponding edge transmission channel.

4. The digital measurement method for product maintenance according to claim 1, characterized in that: The product maintenance management terminal includes multiple digital calculation models, each of which is obtained through training, and the method includes: Collect historical records of products of the same type as the current product, including product defect history data, product maintenance history data, and product status changes corresponding to each product. The product defect history data includes defect type, location, severity, and frequency of occurrence. The product maintenance history data includes maintenance operation content, maintenance cycle, resource consumption, and maintenance effect. Identifying initial product defect data and initial product maintenance data of initial maintenance of each product based on the historical record data; Classifying the historical record data according to the initial product defect data and the initial product maintenance data of the initial maintenance of each product to obtain multiple groups of historical record data; Model training is performed respectively according to the multiple groups of historical record data to obtain multiple digital measurement models.

5. The digital measurement method for product maintenance as claimed in claim 4, characterized in that: Get multiple sets of historical records, including: Calculate the defect similarity and maintenance similarity between products based on the initial product defect data and initial product maintenance data of the initial maintenance of each product; Use hierarchical clustering to cluster the products whose defect similarity and maintenance similarity are both greater than a preset similarity threshold to obtain multiple groups of clustered products, and output multiple groups of historical record data corresponding to the multiple groups of clustered products.

6. The digital measurement method for product maintenance as claimed in claim 5, characterized in that: The product maintenance management terminal performs digital measurement model matching according to the product defect data and the product maintenance data, and the method includes: The product maintenance management terminal obtains a plurality of product defect features and a plurality of product maintenance features corresponding to the plurality of digital measurement models; Calculating multiple defect similarities and multiple maintenance similarities between the product defect data and the product maintenance data and the multiple product defect features and the multiple product maintenance features; Acquire multiple comprehensive similarities according to the multiple defect similarities and the multiple maintenance similarities; Based on the multiple comprehensive similarities, a matching digital measurement model is selected from the multiple digital measurement models, and the comprehensive similarity of the matching digital measurement model is the highest among the multiple comprehensive similarities.

7. The digital measurement method for product maintenance according to claim 1, characterized in that: The user mobile terminal calls the matching digital calculation model to perform digital calculation on the state change data set and outputs the next maintenance calculation result. include: The state change data set includes changes in the operating state of mechanical equipment, changes in electrical equipment parameters, and changes in software systems; Performing feature extraction on the state change data set to obtain a feature vector set; The feature vector set is digitally calculated based on the matching digital calculation model, and the next maintenance calculation result is output.

8. A digital measurement system for product maintenance, characterized in that: The system is used to execute the digital measurement method for product maintenance according to any one of claims 1 to 7, comprising: A maintenance data uploading module is used to record product defect data and product maintenance data, and determine whether the current product is undergoing maintenance for the first time. If the current product is undergoing maintenance for the first time, the product defect data and the product maintenance data are uploaded to a product maintenance management terminal, wherein the product maintenance management terminal includes multiple digital measurement models; A calculation model matching module, used to perform digital calculation model matching according to the product defect data and the product maintenance data through the product maintenance management terminal, output a matching digital calculation model, and download the matching digital calculation model to a user mobile terminal; The status change monitoring module is used to monitor the status change data set of the product after maintenance in real time; The maintenance calculation module is used to call the matching digital calculation model through the user mobile terminal to perform digital calculation on the state change data set and output the next maintenance calculation result, wherein the next maintenance calculation result includes the defect type and maintenance time.

Images