Tool management and control method based on Internet of Things and enhanced Apriori algorithm

Abstract

The invention relates to the technical field of tool management and control, and discloses a tool management and control method based on the Internet of Things and an enhanced Apriori algorithm, and the method comprises the steps: collecting tool data through combining the RFID technology of the Internet of Things and a wireless sensor; performing dynamic clustering analysis by using an improved DBSCAN algorithm, screening out a frequent item set by using space-time semantic enhanced Apriori, and generating an association rule; establishing a cost prediction model by using a multivariate adaptive regression spline, constructing a health state prediction model based on an attention mechanism neural network model, and constructing a residual life prediction model by using support vector machine regression; according to the method, the data is deeply mined based on the intelligent algorithm, the association rules are mined in combination with the improved DBSCAN algorithm and the Apriori algorithm fused with the space-time semantics, and decision support is provided for tool management and control. The tool management and control method has the advantages that the data is deeply mined based on the intelligent algorithm, and the association rules are mined in combination with the improved DBSCAN algorithm and the Apriori algorithm fused with the space-time semantics.

Description

Technical Field

[0001] The present invention relates to the technical field of tool management and control, and in particular to a tool management and control method based on the Internet of Things and an enhanced Apriori algorithm. Background Art

[0002] With the advancement of industrial technology and the expansion of production scale, the variety of tools and instruments is increasing, and their application scenarios are becoming increasingly complex, placing tremendous pressure on traditional tool management and control. Accidents caused by improper tool use and management are common. According to statistics, the main problems include: substandard product quality, improper operation by personnel, selection that does not meet actual requirements, and substandard on-site working conditions. Furthermore, during the use of tools, there are many problems such as improper operation by personnel and inadequate personal protective equipment. Therefore, how to properly manage tools and instruments has become a pressing issue. Summary of the Invention

[0003] In view of this, the present invention provides a tool management and control method based on the Internet of Things and an enhanced Apriori algorithm to solve the problem of how to reasonably manage tools.

[0004] In a first aspect, the present invention provides a tool management and control method based on the Internet of Things and an enhanced Apriori algorithm, the method comprising: Combine IoT RFID radio frequency technology and wireless sensors to collect tool data, including tool usage data and tool environment data; The improved DBSCAN algorithm is used to perform dynamic cluster analysis on tool data. The spatiotemporal semantics-enhanced Apriori algorithm is used to filter frequent itemsets from the dynamic cluster analysis results. Association rules are generated based on the frequent itemsets. The spatiotemporal semantics-enhanced Apriori algorithm is used to embed three-dimensional semantic labels of time, space, and working conditions. Using frequent item sets and association rules as data sets, a cost prediction model was established using the multivariate adaptive regression spline algorithm. A health status prediction model was constructed based on a neural network model with an attention mechanism. A remaining life prediction model was constructed using the support vector machine regression algorithm. The cost prediction model was used to predict the procurement and management costs of tools, the health status prediction model was used to predict the health status of tools, and the remaining life prediction model was used to predict tool wear and tear. Make tool management decisions based on the predicted procurement and management costs, health status, and losses of tools and equipment, combined with actual production needs.

[0005] The present invention realizes data collection of tools and instruments by means of Internet of Things RFID and wireless sensor technology, accurately grasps tool usage data and environmental data, deeply mines data based on intelligent algorithms, realizes dynamic data analysis in combination with the improved DBSCAN algorithm, mines association rules by integrating the innovative Apriori algorithm with spatiotemporal semantics in each cluster, and constructs corresponding prediction models using multiple machine learning algorithms to provide data basis and decision support for tool management and control. On this basis, it ensures the effective implementation of tool equipment standards, quality integrity standards and safe operating specifications, realizes a closed-loop management and control from procurement to scrapping, deeply integrates technical means with management requirements, and forms an intelligent management and control system covering the entire life cycle of tools and instruments.

[0006] In a second aspect, the present invention provides a tool control device based on the Internet of Things and an enhanced Apriori algorithm, the device comprising: A data acquisition unit is used to collect tool data using the Internet of Things RFID radio frequency technology and wireless sensors. The tool data includes tool usage data and tool environment data; The data analysis unit is used to perform dynamic cluster analysis on tool data using the improved DBSCAN algorithm, filter frequent item sets from the dynamic cluster analysis results using the spatiotemporal semantics-enhanced Apriori algorithm, and generate association rules based on the frequent item sets. The spatiotemporal semantics-enhanced Apriori algorithm embeds three-dimensional semantic labels of time, space, and working conditions. The prediction unit is used to establish a cost prediction model using a multivariate adaptive regression spline algorithm with frequent item sets and association rules as data sets, a health status prediction model based on a neural network model with an attention mechanism, and a remaining life prediction model using a support vector machine regression algorithm. The cost prediction model is used to predict the procurement and management costs of tools, the health status prediction model is used to predict the health status of tools, and the remaining life prediction model is used to predict tool wear and tear; The decision-making unit is used to make tool management decisions based on the predicted procurement and management costs, health status, and loss of tools, combined with actual production needs.

[0007] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, computer instructions stored in the memory, and the processor executing the computer instructions to execute the tool management method based on the Internet of Things and the enhanced Apriori algorithm of the above-mentioned first aspect or any corresponding embodiment thereof.

[0008] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the tool management method based on the Internet of Things and the enhanced Apriori algorithm of the above-mentioned first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0010] Figure 1 1 is a flow chart of a tool management and control method based on the Internet of Things and an enhanced Apriori algorithm according to an embodiment of the present invention; Figure 2 2. It is a schematic diagram of the whole process control flow of tools according to an embodiment of the present invention; Figure 3 2. It is a schematic diagram of the overall architecture of a tool management and control system based on the Internet of Things and an enhanced Apriori algorithm according to an embodiment of the present invention; Figure 4 1. It is a flowchart of data analysis and decision-making adjustment of the whole-process control system of tools according to an embodiment of the present invention; Figure 5 This is a structural block diagram of a tool control device based on the Internet of Things and an enhanced Apriori algorithm according to an embodiment of the present invention; Figure 6 Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0011] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0012] At present, the main problems in the management of tool use are: unqualified product quality, non-standard operation of personnel, selection does not meet actual requirements, on-site working conditions do not meet the requirements, etc.

[0013] Among them, the causes of product quality problems include substandard factory quality (fake products), unqualified storage conditions, lack of maintenance or improper maintenance, and products that have become ineffective or expired due to long-term use. The reason why the selection of tools and instruments does not meet the actual requirements is that the use and procurement are disconnected. The purchase quantity and model can only be determined based on simple needs. There is a lack of accurate data support, which leads to the purchase of too many or inappropriate tools. The management during use is even more problematic, with many problems such as non-standard personnel operation, substandard on-site working conditions, and improper personal protection. The reasons include the lack of systematic operating standards, unqualified operator training, weak safety awareness and sense of responsibility, and the lack of effective monitoring methods and control processes. It is difficult for companies to track the frequency of use of tools, the use environment, and the standardization of operations, resulting in inadequate on-site control.

[0014] The root cause of this is the lack of comprehensive, systematic, and effective management standards. Enterprises don't pay enough attention to tool management, leading to a relatively casual use of tools, which not only accelerates tool damage but also increases the risk of safety accidents. Tool management involves many steps, and the causes of these problems are complex. To effectively address safety hazards associated with tool use and promote the standardization of safe production, it is imperative to establish a comprehensive and systematic tool management approach, strictly implement tool equipment standards, quality standards, and safe operating procedures (the "three standards"), and achieve intelligent safety management throughout the entire process, from procurement to scrapping (the "one process"). This is of great significance for improving safe production.

[0015] The tool management and control method based on the Internet of Things and the enhanced Apriori algorithm provided in the embodiments of the present invention is applicable to various tools and equipment used in various fields such as industrial production and power operation and maintenance, including hand tools (such as wrenches, screwdrivers, etc.), power tools (such as electric drills, electric saws, etc.), pneumatic tools (pneumatic picks, air shovels, etc.), lifting tools (such as lifting pulleys, jacks, etc.) and various safety protection equipment (such as safety helmets, safety belts, etc.).

[0016] According to an embodiment of the present invention, an embodiment of a tool control method based on the Internet of Things and an enhanced Apriori algorithm is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0017] In this embodiment, a tool management and control method based on the Internet of Things and an enhanced Apriori algorithm is provided. Figure 1 FIG. 1 is a flow chart of a tool control method based on the Internet of Things and an enhanced Apriori algorithm according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps: Step S101: Use the Internet of Things RFID radio frequency technology and wireless sensors to collect tool data.

[0018] In the embodiment of the present invention, Figure 2 and Figure 3 As shown, the IoT perception layer is deployed. The IoT perception layer includes on-site IoT devices and multiple sensor nodes, which are used to collect usage data of tools and the environmental data of the tools. By associating with on-site sensors or environmental monitoring IoT devices to collect environmental information, the data collection method can be optimized, costs can be reduced, and efficiency can be improved.

[0019] It is understandable that environmental data can be triggered by tool sensors for synchronous collection and upload. When tool sensors collect specific working condition data, rules can be set to trigger other sensors or devices in the working environment to collect and upload data. At the production site, when the vibration sensor of a power tool (such as an electric drill) detects abnormal vibration, the sensor on the drill can trigger sensors in the surrounding environment through wireless communication technology (such as Wi-Fi, Bluetooth, ZigBee, or 4G / 5G, etc.), which then collects data and uploads it synchronously to the management and control system. This method can achieve accurate and synchronous data collection, avoid unnecessary data redundancy, improve the pertinence and timeliness of data collection, and more accurately reflect the actual environmental conditions when the tool is in use.

[0020] IoT devices refer to IoT devices deployed on-site to monitor environmental data. They use IoT RFID (Radio Frequency Identification) technology to implement functions such as tool entry and exit registration, inventory counting, and inventory warnings. Sensor nodes are installed on tools or their usage environment, and wireless communication is used to form a wireless sensor network. The collected data provides a data basis for determining whether the tools meet quality standards and analyzing whether the operating environment complies with safe operating specifications.

[0021] Specifically, when it comes to sensor selection and installation, for different types of tools, appropriate sensor components must be precisely selected and installed in key locations based on their operating characteristics and management requirements to ensure accurate data collection. Given the varying procurement costs of various sensors, the cumulative cost can be considerable for a large number of tools. Furthermore, sensor installation and energy costs must be considered. Currently, installing multiple sensors on each tool can be costly.

[0022] Taking a handheld electric drill as an example, a temperature sensor can be installed on the motor housing to monitor the temperature of the motor during operation, prevent the motor from being damaged due to overheating, and ensure that the electric drill operates within the normal temperature range; a vibration sensor can be installed at the drill bit connection to collect the vibration frequency in real time. If the vibration is abnormal, it may indicate that the drill bit is worn or there is a problem with the connection components, thereby realizing tool component status monitoring and ensuring operational safety.

[0023] Regarding IoT device data connectivity, in various work scenarios, we fully utilize IoT devices already deployed on-site for environmental monitoring, such as temperature and humidity monitoring stations and noise monitors within workshops and warehouses. By establishing data sharing interfaces, we enable collaborative operation between tool sensors and these devices. This allows tool sensors to collect both their own operating condition data and surrounding environmental data, including temperature, humidity, vibration frequency, current intensity, and pressure. Using wireless communication modules, this collected data is transmitted in real time, eliminating the need for redundant sensor installation, reducing costs and improving data collection efficiency.

[0024] When a tool sensor collects specific working condition data, it triggers other sensors or devices in the environment to collect and upload data according to preset rules. For example, if a temperature sensor on an electric tool (such as an electric drill or electric grinder) detects that the motor temperature exceeds a preset safety threshold (such as 80°C), it will trigger the detection of the surrounding temperature to determine whether the ambient temperature is too high and affecting the motor's heat dissipation, which helps to assess the external factors causing the increase in motor temperature. In actual work scenarios, the above method achieves precise and synchronous data collection, providing an accurate data foundation for subsequent judgment of tool status based on quality integrity standards and assessment of operating environment risks based on safe operating specifications.

[0025] Different types of tools are equipped with various sensor components to address their operational characteristics and management requirements. In real-world work scenarios, a large number of IoT devices are deployed to monitor environmental parameters, such as temperature and humidity monitoring stations and noise monitors installed in workshops and warehouses. Tool sensors establish data sharing interfaces with IoT devices to collect tool status data such as temperature, humidity, vibration frequency, current intensity, and pressure, as well as environmental data such as workplace temperature and humidity. This data is transmitted in real time via wireless communication modules, eliminating the need to repeatedly install the same functional sensor on each tool and providing a basis for analyzing the relationship between tool performance and the environment.

[0026] Step S102 : performing dynamic cluster analysis on tool data using the improved DBSCAN algorithm model, using the spatiotemporal semantics enhanced Apriori algorithm to filter out frequent itemsets from the dynamic cluster analysis results, and generating association rules based on the frequent itemsets.

[0027] In the embodiment of the present invention, Figure 2and Figure 3 As shown, an intelligent analysis layer is deployed. The intelligent analysis layer is used to analyze and process the data in the data storage layer. The intelligent analysis layer includes a data preprocessing module and an association rule mining module. The data preprocessing module is used to clean, extract features and convert data from the original data. The association rule mining module is used to mine frequent item sets and association rules in tool data.

[0028] The association rule mining module includes the improved DBSCAN algorithm and the spatiotemporal semantics-enhanced Apriori algorithm. The improved DBSCAN algorithm incorporates a dynamic operating condition adaptation mechanism, enabling dynamic cluster analysis based on multi-dimensional data features. The spatiotemporal semantics-enhanced Apriori algorithm module embeds three-dimensional semantic labels for time, space, and operating conditions, enhancing the association rule mining mechanism for dynamic spatiotemporal scenarios and achieving a logical closed loop from cluster discovery to rule extraction for fault modes.

[0029] Specifically, the improved Apriori algorithm introduces time segment functions, spatial aggregation functions and working condition mapping functions in the original transaction modeling stage to construct a multi-transaction module with structured semantics. This ensures that each tool operation record not only contains conventional attributes such as tool attributes, usage environment and operating behavior, but also integrates the corresponding time period type, environmental block and operation status, thereby improving the coupling degree between frequent item sets and actual working conditions.

[0030] The intelligent analysis layer is used to establish a model for tool usage control. Through multi-enterprise interconnection and sharing, it collects data on inspection and usage, compiles statistics on these data records, and uses a variety of machine learning algorithms to build analytical models for related issues. Key statistical analyses include tool procurement and management costs, tool performance, quality reliability, operator suitability, and storage environment control. Regular tool supervision, auditing, and statistical analysis are conducted to provide guidance for optimizing tool procurement, strengthening safe use, warehouse management, and adjusting maintenance strategies.

[0031] It is understandable that through a unified spatiotemporal alignment strategy, multi-dimensional information integration is carried out to construct a data feature map of the entire process of tool management and control. The full-process data of tools from procurement, warehousing, use, maintenance to scrapping are mapped with the working condition characteristics, RFID spatial trajectory information, operation behavior data, and temperature and humidity from environmental monitoring from the perception layer into structured data with spatiotemporal continuity, expanding the data dimension and providing data support for downstream clustering algorithms and association rule mining algorithms.

[0032] Using mined association rules and related data, we establish a tool health status prediction model and conduct specific functional analysis for different tool types. We employ algorithms such as multivariate adaptive regression, support vector machine regression, and neural networks to build tool prediction models. We incorporate association analysis results into the prediction model to improve prediction accuracy. Based on predicted performance degradation trends, we plan tool upgrades in advance. Furthermore, based on the correlation between operating behavior and failures, we develop personalized training plans for operators to reduce failures caused by improper operation.

[0033] In step S103, a cost prediction model is established using a multivariate adaptive regression spline algorithm with frequent item sets and association rules as data sets, a health status prediction model is constructed using a neural network model based on an attention mechanism, and a remaining life prediction model is constructed using a support vector machine regression algorithm.

[0034] In the embodiment of the present invention, Figure 4 As shown, the intelligent analysis layer also includes a prediction model construction module, which is used to build a tool health status prediction model based on the mining results, providing data basis and decision support for the optimization and dynamic adjustment of the "three standards".

[0035] The prediction model building module uses algorithms including, but not limited to, multivariate adaptive regression, support vector machine regression, and neural networks. The prediction model uses historical tool data and real-time monitoring data to predict tool procurement and management costs, as well as health indicators such as failure probability, performance degradation trends, and remaining lifespan. Specifically, the cost prediction model is used to predict tool procurement and management costs, the health prediction model is used to predict tool health, and the remaining lifespan prediction model is used to predict tool wear and tear.

[0036] Step S104 : making tool management decisions based on the predicted tool purchase and management costs, health status, and loss, combined with actual production needs.

[0037] In the embodiment of the present invention, Figure 2 and Figure 3 As shown in the figure, the application service layer is deployed. The application service layer is used to provide users with management and control strategies and decision suggestions based on the results of the intelligent analysis layer. Figure 4 As shown in the figure, the application service layer includes inventory management module, warehouse storage module, use and maintenance module, training assessment module, and supervision and inspection module, which are used to realize functions such as planning procurement and management decision-making, optimizing tool use strategy, standardizing personnel operation, and adjusting maintenance cycle.

[0038] The application service layer includes a tool process control platform and a mobile application. The tool process control platform serves as an interactive interface for enterprise managers and operators, offering a variety of functional modules. Leveraging technology, it conducts process and operational control of tools, facilitates tool usage training, and monitors on-site operations, ensuring compliance with the three standards and promptly identifying any issues. The mobile application provides a more convenient mobile management tool for on-site operators. Using the application on mobile devices (such as smartphones and tablets), operators can query the location and status of tools in real time, enabling quick and easy access to the required tools. Before using a tool, they can scan the RFID tag on the tool to obtain operating instructions and safety precautions, ensuring proper use. Furthermore, operators can directly report any tool anomalies, such as unusual vibration or odor, through the mobile application during use. The system promptly receives and processes this feedback, improving the timeliness of problem detection and resolution.

[0039] Specifically, the tool process control platform includes an inventory management module, a warehouse storage module, a use and maintenance module, a training assessment module, and a supervision and inspection module, which realizes closed-loop management of the entire process of tools from procurement, warehousing, issuance, use, return, inspection, and scrapping. It has functions such as asset entry, asset scheduling, asset scrapping, asset inventory and spot checks, and the status data collected by the sensor is aggregated to the communication node through the wireless network, and transmitted to the intelligent management platform by the communication network for decision analysis, so as to realize the management of tool planning, procurement and management decisions, optimize usage strategies, strengthen safety operations, and adjust maintenance strategies.

[0040] The inventory management module strictly adheres to tool equipment standards, implementing control procedures for tool procurement, acceptance, warehousing, and storage. It utilizes IoT RFID radio frequency technology to manage tool inbound and outbound registration, inventory counts, and inventory alerts. Tools are uniformly categorized and numbered, and comprehensive tool archives and ledgers are established. This robust tool archive and ledger management allows managers to view real-time tool inventory, location distribution, and inventory turnover rates. Using procurement and management cost forecasting models, they can rationally plan purchases and avoid inventory overstocks or stockouts.

[0041] Warehouse Storage Module: Utilizing IoT RFID radio frequency technology and wireless sensor networks, the warehouse storage and management of tools is intelligent, in accordance with quality standards. Comprehensively sensing the storage environment and health status of tools, such as temperature, humidity, and ventilation, collaborates with the RFID system to provide comprehensive monitoring of tools and ensure their quality. Electronic archives are established, clearly marking and numbering storage locations for rapid retrieval, retrieval, and storage, ensuring the implementation of fixed-location management for tools. Real-time monitoring of the storage environment ensures that tools are stored in a favorable environment, ensuring product quality and extending their service life.

[0042] The Operation and Maintenance Module: Based on safe operating procedures, this module leverages RFID technology, wireless sensor networks, tool health prediction models, and image tracking and intelligent recognition technology to conduct tool maintenance and on-site operation monitoring and management, ensuring compliance with the three standards and promptly identifying any issues. Information such as tool receipt time, user, usage duration, and location is recorded to facilitate traceability of tool usage history and accountability. Sensors within the IoT perception layer monitor key parameters such as tool mechanical strength and voltage, as well as the temperature, humidity, and pressure of the tool's operating environment, in real time. Tool function is assessed based on quality standards. In the event of environmental anomalies, environmental control equipment, including but not limited to air conditioners, dehumidifiers, and air purifiers, is activated to ensure that tools are used in an environment that meets quality standards. The tool health prediction model alerts operators to adjust tool usage duration or frequency when tool use approaches risk thresholds. Image recognition technology is used to analyze and identify unsafe behaviors during tool operation, generating real-time alerts and corrective actions. The handheld terminal also has multiple built-in sensors that enable rapid on-site testing of key parameters such as mechanical strength and voltage of tools. It also analyzes and identifies unsafe behaviors during tool operation, providing real-time alerts. Based on intelligent algorithm analysis, it generates maintenance plans, including scheduled maintenance tasks (such as cleaning, lubrication, and calibration) and predictive fault repairs. Maintenance personnel can view maintenance task lists and record maintenance processes and results on the platform, achieving standardized and information-based maintenance work.

[0043] In addition, maintenance cycles are dynamically adjusted based on association rules and intelligent algorithm analysis results. Maintenance strategies are adjusted based on tool type, usage environment, and historical fault data. A reasonable maintenance plan is arranged, including regular maintenance tasks (such as cleaning, lubrication, and calibration), special maintenance tasks, and maintenance tasks based on fault prediction. This improves maintenance efficiency and ensures that tools meet quality standards. Maintenance personnel can view maintenance task lists on the platform, record maintenance processes and results, and achieve standardized and information-based maintenance work, ensuring that tools comply with safe operating regulations and maintain good performance.

[0044] Training and Assessment Module: Based on safe operating procedures, VR (Virtual Reality) technology is used to compile visual training materials, including hazard factors identified, quality standards, and safe operating procedures, and a question bank is established. Personalized training content and assessments are delivered via an app to ensure that users are familiar with the specific tool performance, characteristics, operation, and maintenance requirements before use. For complex or high-risk tools, specialized training is organized to ensure operators are proficient in operating skills and safety precautions. Through training and assessment, operators' safety awareness and standardized operating capabilities are strengthened, reducing safety accidents and tool damage caused by improper operation and ensuring that operators strictly adhere to safe operating procedures.

[0045] Supervision and Inspection Module: Based on tool equipment standards, quality standards, and safe operating procedures, this module is used to supervise, inspect, and manage the use and maintenance of tools. Leveraging RFID radio frequency technology and apps, it comprehensively supervises, inspects, and manages the use and maintenance of tools. It keeps track of tool status information and clarifies relevant inspection content and standards, including pre-use inspections, regular inspections, preventive inspections, and scrapping procedures. Based on specific tool regulations, a professional inspection agency is selected, or the company conducts regular annual quality and performance inspections. Inspection content is clearly defined, and inspection results are entered into a regular inspection record sheet. Through supervision and inspection, problems with tool equipment, quality, and operation can be promptly identified, and relevant personnel can be urged to make corrections to ensure the effective implementation of the "three standards."

[0046] The tool management and control method based on the Internet of Things and the enhanced Apriori algorithm provided in this embodiment realizes data collection of tools and tools by leveraging the Internet of Things RFID and wireless sensor technology, accurately grasps tool usage data and environmental data, deeply mines data based on intelligent algorithms, and realizes dynamic data analysis in combination with the improved DBSCAN algorithm. The innovative Apriori algorithm that integrates spatiotemporal semantics in each cluster is used to mine association rules, and a variety of machine learning algorithms are used to build corresponding prediction models to provide data basis and decision support for tool management and control. On this basis, the effective implementation of tool equipment standards, quality integrity standards and safety operation specifications is ensured, and a closed-loop management and control system from procurement to scrapping is realized, so that technical means and management requirements are deeply integrated to form an intelligent management and control system covering the entire life cycle of tools and tools.

[0047] In this embodiment, a tool management and control method based on the Internet of Things and an enhanced Apriori algorithm is provided. The process includes the following steps: Step S501: Use the Internet of Things RFID radio frequency technology and wireless sensors to collect tool data.

[0048] For details, please see Figure 1Step S101 of the illustrated embodiment will not be described in detail here.

[0049] In some optional embodiments, the method further comprises: Step S502: According to the application scenario and data transmission requirements, a corresponding transmission method is selected, a data transmission gateway is set, and the tool data is transmitted.

[0050] In an embodiment of the present invention, a data transmission layer is deployed, and the data collected by the IoT perception layer is converted into a protocol and transmitted to the data storage layer with the help of a data transmission gateway, thereby ensuring the stable transmission of data related to tool equipment standards, quality integrity standards, and safe operating specifications.

[0051] The data transmission gateway supports a variety of wired or wireless communication methods, including Ethernet, 4G / 5G networks, Wi-Fi, etc. The data transmission gateway transmits tool data to the database management system of the cloud server or local data center for storage.

[0052] The data storage layer uses a database management system to store basic tool information, usage records, sensor data, and maintenance and fault records, providing data query and storage support for the implementation of the "three standards." The data storage layer is used to store basic tool information, usage records, sensor data, and maintenance and fault records. It includes modules such as basic standards, tool archives, and process control. The database is categorized to facilitate query and is linked to related modules such as job document compilation to achieve data information sharing. The database's main contents include: basic standards module, inventory storage module, inventory management module, on-site usage information module (including existing problems and hidden dangers), and supervision and inspection module.

[0053] Specifically, the data is transmitted to the data storage layer through the gateway via the data transmission layer, and the appropriate transmission method is selected according to different application scenarios and data volume requirements.

[0054] In local areas within the factory, low-power, short-range ZigBee or Bluetooth transmission is used; within large construction sites or corporate campuses, Wi-Fi or 4G / 5G networks are used for data transmission. During the data transmission process, a data transmission gateway is set up to connect the aggregation nodes of the wireless sensor network, convert the received data into different protocols, and transmit it. Encryption technologies such as SSL (Secure Sockets Layer) and TLS (Transport Layer Security) are strictly used to ensure data integrity and availability, prevent data theft, tampering, or loss during transmission, and ensure that data related to the implementation of the "three standards and one process" tool control is transmitted securely and stably to subsequent links.

[0055] The data storage layer utilizes a relational database (such as MySQL) or a non-relational database (such as MongoDB) to build a data storage system. A rational data table structure is designed to categorize and store basic tool information, usage records, sensor data, and maintenance and fault records. Databases are established for basic standards, inventory management, warehouse storage, field usage information, and supervision and inspection, effectively integrating data from each module for easy query and management. Technical measures such as data backup and redundant storage ensure the security and reliability of data storage, providing data support for the implementation of tool equipment standards, quality integrity standards, and safe operating procedures, enabling relevant personnel to quickly access the required data for analysis and decision-making.

[0056] The Basic Standards module primarily stores basic standards for tool management, serving as the regulatory basis for the entire management system. These standards cover tool configuration standards, quality standards, safe operating procedures, and more.

[0057] The Inventory Management Module primarily stores information related to tool inventory, enabling effective management of tool procurement, warehousing, outbound delivery, and inventory counts. This includes basic tool information such as name, model, and specifications; procurement information such as purchase date, quantity, and price; and information such as the inventory count of various tools and inbound and outbound records.

[0058] Inventory Management Module: This module primarily stores information about tools during their storage in the warehouse, focusing on the storage environment and maintenance status of the tools. This includes warehouse environment information, such as temperature, humidity, and ventilation; tool storage location and fixed management information for quick location and access; tool quality status and maintenance information, tracking in real time whether tools have quality issues such as damage and aging, and recording the time and content of each maintenance. For example, in a warehouse storing insulating tools, the warehouse management module will record the warehouse's temperature and humidity data, the storage location and quality status of insulating gloves and boots, and regular inspection and maintenance records.

[0059] On-site usage information: This module primarily stores actual information about tools used on-site, as well as any problems and potential hazards discovered during use. This includes tool acquisition information, such as the time and person who acquired the tool; usage information, including the time and duration of use, and operating parameters; and fault records, such as tool damage, abnormal performance, and improper operation. For example, records of a particular electric drill include the person who acquired it, the location and duration of use, operating voltage and temperature, and any issues such as motor overheating and unstable speed that occurred during use.

[0060] The Supervision and Inspection Module primarily stores information related to the supervision and inspection of tool use and maintenance to ensure the effective implementation of various tool management standards and regulations. This includes inspection plan information, such as inspection cycle, inspection content, and inspection personnel arrangements; inspection results, which record the actual condition of tools during each inspection and whether they meet standard requirements; and rectification records, which record rectification measures, responsible persons, rectification deadlines, and rectification results for any issues discovered during the inspection.

[0061] In some optional embodiments, the method further comprises: Step S503: Delete abnormal values ​​and duplicate values ​​in the tool data, and fill in missing data using linear interpolation.

[0062] Step S504 : extracting features from the tool data. The extracted features include tool usage frequency, usage duration, and fault interval duration.

[0063] Step S505 , normalizing and standardizing the tool data.

[0064] In an embodiment of the present invention, the tool data is preprocessed, including data cleaning, data denoising, data interpolation, feature extraction, data conversion and other operations, so as to facilitate the subsequent use of intelligent algorithms to analyze the data and explore the coupling relationship between different types of data.

[0065] In addition to collecting routine working condition data such as tool performance, current, temperature, and environmental data such as temperature, humidity, and pressure, the collection of dynamic working condition data such as tool operation behavior data, energy consumption fluctuations, and fault information is increased. Multi-sensor and IoT device fusion technology is used to perform time synchronization and spatial calibration of different types of sensor data.

[0066] Specifically, data cleaning operations include removing outliers, duplicate data, and incomplete data; feature extraction operations include extracting features related to tool performance and health status, such as tool usage frequency, usage duration, and fault interval duration; and data conversion operations include normalizing and standardizing data.

[0067] During the data cleaning phase, outliers are identified and removed by setting reasonable thresholds. Duplicate data is identified and deleted using the data's timestamps and unique identifiers. Incomplete data is supplemented by using mean filling or linear interpolation based on the trends of previous and subsequent data.

[0068] For example, tool management requires a deep understanding of the impact of environmental parameters and operator information on power tool use. To this end, we collected real-world data related to power tool management, covering aspects such as ambient temperature and humidity, current intensity, vibration frequency, workplace ventilation, the number of operations, and the duration of each operation.

[0069] The collected data needs to be cleaned. First, the 3σ principle is used to remove outliers. Assume that the data set of the usage time of a power tool is { X 1, X 2,…, X n}, calculate the mean μ And standard deviation σ, if the data point X i Satisfaction | X i - μ If the value of |>3σ is greater than 3σ, it is considered an outlier and removed. Based on the data's inherent volatility within different time windows, the outlier threshold is dynamically adjusted to more accurately identify anomalies. Duplicate data is identified and deleted by comparing tool numbers and timestamps. Missing data is supplemented using weighted interpolation or linear interpolation based on the trends of previous and subsequent data.

[0070] In the feature extraction phase, key feature information is extracted from collected data, environmental parameters, and human operation information, based on system requirements. Cost information such as tool purchase price and repair costs is extracted through the inventory management module. User operation information such as frequency and duration of use is collected from tool operation logs. Data on tools and environmental parameters is collected through the IoT perception layer. The number of faults and the time between faults are extracted from fault records. In the data conversion phase, the key feature data of the required tools is standardized and normalized according to the algorithm requirements. The Z-Score normalization formula is used to process the data to obtain the standardized data.

[0071] For example, when building a health prediction model for an electric drill based on collected environmental parameters and operator operation data, it's necessary to extract key characteristic data such as usage frequency, operating temperature, load intensity, and number of failures. Because the dimensions and numerical ranges differ, direct comparison and analysis are impossible, requiring the use of a standardized Z-Score formula.

[0072] The Z-Score normalization formula is: ,in, X is the original data, is the mean, is the standard deviation. Taking ambient temperature as an example, we calculate the mean and standard deviation of all temperature data and use the standardization formula to obtain standardized temperature data. This puts all types of data on the same comparable scale, eliminating the influence of dimension.

[0073] The preprocessed data can improve the accuracy of frequent item set mining, enhance the training speed and accuracy of the prediction model, and help to more accurately predict the procurement and management costs of tools, as well as health status indicators such as failure probability, performance degradation trend, and remaining life. It provides strong data support for the implementation and optimization of tool equipment standards, quality integrity standards, and safety operating specifications.

[0074] Step S506 , using the improved DBSCAN algorithm to perform dynamic cluster analysis on the tool data, using the spatiotemporal semantics enhanced Apriori algorithm to filter out frequent itemsets from the dynamic cluster analysis results, and generating association rules based on the frequent itemsets.

[0075] Specifically, the above step S506 includes: Step S5061: introduce a dynamic working condition adaptation mechanism and use the improved DBSCAN algorithm to perform dynamic cluster analysis on tool data.

[0076] Step S5062: introduce the spatiotemporal operating condition importance factor, construct a weighted support function, and filter out frequent itemsets from the dynamic cluster analysis results.

[0077] Step S5063: integrate the spatiotemporal working condition semantics and mine association rules that meet the confidence requirements.

[0078] In an embodiment of the present invention, an improved DBSCAN algorithm is used to perform cluster analysis based on different features of tool data. The DBSCAN algorithm is a density-based clustering algorithm that can form clusters based on the density of data points without pre-specifying the number of clusters.

[0079] The improved DBSCAN algorithm incorporates a dynamic working condition adaptation mechanism to dynamically adjust tool control throughout the entire process. Tool category labels are embedded in the data feature map. Through supervised training on historical fault label samples, the algorithm extracts the coupling relationship between specific tool types and their optimal parameters. For example, when analyzing usage data for different tool types, the DBSCAN algorithm categorizes tool data into distinct cluster labels, such as power tools, hand tools, arc welding tools, and high-frequency electrical tools.

[0080] A semantically enhanced Apriori algorithm is used to mine frequent itemsets and association rules. Its core mechanism is based on the a priori property that "all subsets of a frequent itemset must also be frequent itemsets," generating association rules based on frequent itemsets. This improved algorithm is no longer limited to mining frequent item combinations within static transaction sets. Instead, it embeds three-dimensional semantic labels across time, space, and working conditions, constructing a context-enhanced association rule discovery mechanism for dynamic scenarios. By setting weighted support and confidence, it mines meaningful frequent itemsets and association rules from datasets.

[0081] When mining frequent item sets and association rules in each cluster, in order to support frequent item mining of high-dimensional semantic transactions, the algorithm constructs a weighted support function and introduces the importance factor of spatiotemporal conditions. The function is defined as follows:

[0082] in, 、 、 Respectively represent i The importance weights of the semantic dimensions of time, space and working conditions in each transaction, is the indicator function, when the item set A Included in transaction The value is 1 when it is in the middle, otherwise it is 0. N is the total number of transactions.

[0083] Through a dynamic weighting mechanism, item sets located in high-risk time periods, critical environmental blocks, or high-stress working conditions are given higher weights in the frequency calculation, thereby tilting the identification of frequent item sets toward potential accident association patterns.

[0084] Furthermore, based on the frequent itemsets, a weighted confidence function is defined to perform rule screening and sorting:

[0085] in, A and B are sets of attributes that meet specific conditions, A → B It means "if A If it appears, B association rules that may appear, is the weighted confidence of the rule, which is used to measure the A → B reliability, For item sets A ∪ B The weighted support of For item sets A The weighted support of .

[0086] The semantically enhanced Apriori algorithm can extract representative transactions based on high-risk clusters, annotate them with three-dimensional labels, and then enter the transaction pool to participate in weighted mining, thus achieving a logical closed loop from cluster discovery to rule extraction of failure modes.

[0087] Association rules are generated based on frequent sets, and confidence levels are calculated and filtered. For example, the association rule {midday operation, high-temperature location, abnormal vibration} → {bearing failure} is generated, and the calculated confidence level is 0.8, which is greater than the minimum confidence threshold (minconfidence = 0.7). This means that in the cluster of electric drilling tools, when such tools are operated in a high-temperature location during the midday hours and experience abnormal vibration, there is an 80% probability of a bearing failure. By integrating association rule mining with spatiotemporal working condition semantics, high-risk operating scenarios can be accurately identified, providing data support for enterprise tool maintenance and management. Based on the results, tool procurement, inventory management, and maintenance strategies can be optimized to ensure the effective implementation of the "three standards."

[0088] The improved Apriori algorithm supports frequent item mining for high-dimensional semantic transactions by constructing a weighted support function and expanding the traditional support formula by introducing a spatiotemporal working condition importance factor. Based on frequent item sets, the algorithm inherits the traditional Apriori rule generation logic and defines a weighted confidence function for rule screening and sorting. By combining these functions, valuable strong association rules are mined from the tool management process dataset and the key factors affecting the tool health status are identified. The generated strong association rules, combined with real-time tool data, are used to predict the changing trends of key indicators, providing data support for the subsequent adjustment of tool usage management strategies and the implementation of the "three standards and one process" policy.

[0089] By combining the improved DBSCAN algorithm with the semantically enhanced Apriori algorithm, the usage scenarios, operation data, environmental information and fault correlation of tools are analyzed, the key factors affecting the health status of tools are determined, the mining efficiency and targeting are improved, and a prediction model is built based on historical data to dynamically optimize the procurement and management decisions, usage strategies, personnel operations and maintenance cycles of tools.

[0090] Specifically, the steps to improve the DBSCAN algorithm include: In step Sa, working condition perception is introduced to analyze working condition data collected by wireless sensors and calculate characteristic volatility index.

[0091] In step Sb, nonlinear mapping is performed based on the characteristic volatility index, a density threshold adjustment function is constructed, and a reference neighborhood radius and minimum point number parameter combination adapted to the working condition data is output.

[0092] In step Sc, the DBSCAN algorithm is improved based on the output parameter combination of the benchmark neighborhood radius and the minimum number of points.

[0093] In an embodiment of the present invention, the improved DBSCAN algorithm introduces a dynamic working condition adaptation mechanism and adopts a working condition perception module. The working condition perception module dynamically adjusts the density threshold and clustering sensitivity in real time according to the operating status of the tool, reduces the clustering misjudgment rate, and improves the recognition accuracy of collected data and fault information.

[0094] The working condition perception module collects working condition data by analyzing wireless sensors. The working condition data includes current fluctuations, temperature gradients, vibration frequencies and other data.

[0095] Calculating characteristic volatility θ As the basis for judgment, the dynamic adjustment process of the density threshold is driven by this. The characteristic volatility index measures the intensity of the tool operation by the rate of change of the working condition parameters per unit time. The formula is as follows:

[0096] in, X i Indicates the i The characteristic parameters of each working condition, such as temperature T, vibration amplitude V, current intensity I, etc., dX i / d t is the time derivative of the parameter, n is the number of characteristic parameters being monitored.

[0097] Specifically, the characteristic volatility θ Reflects the load complexity and dynamic disturbance degree of the tool. The higher the value, the more severe the working or impact load condition is. The lower the value, the more light load or stable the operation is.

[0098] Based on characteristic volatility θ Perform nonlinear mapping to construct a density threshold adjustment function and output a dynamic adaptation for different working conditions. Combined with the MinPts parameter, the density threshold adjustment function uses an exponential decay form to construct an adaptive neighborhood radius function, as shown below:

[0099] in, is the adaptive neighborhood radius function, is the base neighborhood radius, k is the adjustment coefficient, and its value can be obtained by the system through offline fitting of the training set.

[0100] The core logic of this function is that when the characteristic volatility θ Increased, indicating that the tool is in a high-load and high-risk state. Automatically reduce and limit the clustering range to improve the sensitivity of identifying outliers, otherwise the characteristic volatility θ It is also necessary to avoid normal fluctuations in low-risk scenarios being misidentified as fault noise.

[0101] At the same time, according to Based on the output of the system, the system adjusts MinPts synchronously, raising it to a higher value in high-risk working conditions to enhance the robustness of local density judgment. This dynamic adjustment mechanism embeds tool category labels in the feature map and extracts the coupling relationship between specific tool types and their optimal parameters through supervised training on historical fault label samples. The minimum number of points parameter MinPts is an important parameter of the DBSCAN algorithm, based on the adjusted benchmark neighborhood radius and the minimum number of points parameter MinPts to improve the DBSCAN algorithm.

[0102] For example, data from 100 different power tools can be divided into different clusters based on their usage, such as drilling, cutting, grinding, and fastening. Two dynamic working condition features, temperature (°C) and vibration frequency (Hz), are selected for drilling power tools. The data under this cluster is normalized, and each feature value is mapped to the interval [0, 1]. If the reference neighborhood radius is set =0.5 and the minimum number of points MinPts=5. For high load conditions, dynamic adjustment is performed based on the characteristic fluctuation rate. The MinPts value is reduced, and the MinPts value is increased, focusing on local high-density areas and enhancing interference resistance in dense areas. The number of points in the neighborhood of each point is then calculated, and the drilling power tool data is divided into four clusters. Cluster 1 contains normal operating condition data, Cluster 2 contains temperature anomaly data, Cluster 3 contains vibration anomaly data, and Cluster 4 contains noise. Association rule mining is performed based on the different clustering results and labels, and corresponding predictive models are constructed to formulate different usage, maintenance, and management strategies.

[0103] Specifically, the frequent itemsets obtained by screening from the dynamic clustering analysis results in step S5062 include: Step S50621: embed three-dimensional semantic labels into the dynamic clustering analysis results to obtain a candidate item set.

[0104] Step S50622: Calculate the support of candidate item sets, and select item sets with support greater than a preset support threshold as frequent item sets.

[0105] In an embodiment of the present invention, after dynamic cluster analysis is performed on tool data, the dynamic cluster analysis results are embedded in three-dimensional semantic tags.

[0106] For example, clustering drilling power tools contains 1,000 fault records, each of which is embedded with a 3D semantic label (e.g., "nighttime operation," "high temperature location," or "abnormal vibration"). The clustered data includes various attributes, such as tool type, usage scenario, vibration frequency, ambient temperature, and fault condition. This data is organized into a transaction dataset.

[0107] Set minsupport=0.15 and minconfidence=0.7 (minsupport is the minimum weighted support threshold, minconfidence is the minimum weighted confidence threshold). Generate candidate item sets and calculate support, and filter out frequent item sets with support greater than 0.15.

[0108] For example, if the support of {high temperature location, abnormal vibration, bearing failure} is 0.4, it is a frequent itemset. Itemsets with support less than the minimum support threshold are deleted, and higher-order candidate itemsets are continuously generated. After multiple iterations, all frequent itemsets are determined.

[0109] Step S507, using frequent item sets and association rules as data sets, using the multivariate adaptive regression spline algorithm to establish a cost prediction model, building a health status prediction model based on the neural network model of the attention mechanism, and using the support vector machine regression algorithm to build a remaining life prediction model.

[0110] Specifically, the construction of the health status prediction model based on the attention mechanism neural network model in step S507 includes: Step S5071, constructing a neural network model.

[0111] In step S5072, the feature vector processed by the attention mechanism is input into the neural network model for model training. Based on L1 regularization pruning, the weight of the neural network model is constrained to obtain a health status prediction model.

[0112] In this embodiment of the present invention, in a practical application scenario for predicting the health status of an electric drill, association rules mined using an improved algorithm were used to obtain data from the IoT perception layer and other database modules, including current, temperature, usage time, vibration frequency, and historical fault records. 1,000 data sets were collected, with 800 used as training sets and 200 as test sets. Data preprocessing was performed first, and a neural network model based on the attention mechanism was constructed.

[0113] Specifically, the neural network model includes an input layer, a hidden layer, and an output layer. The input layer receives a 4-dimensional preprocessed feature vector. Assume that the processed feature vector , set up two hidden layers with 32 neurons each and using ReLU activation function and one output layer with Sigmoid activation function neurons to output the predicted probability of the health status of the electric drill.

[0114] Among them, the ReLU activation function is as follows:

[0115] The Sigmoid function is as follows:

[0116] The training set data is input into the model in sequence to calculate the predicted value, and the binary cross entropy loss function is used to measure the difference between the predicted value and the true value (historical fault record).

[0117] The binary cross entropy loss function is shown below:

[0118] in, L is the loss function value, N is the sample size, y i It is i The true label of each sample (value is 0 or 1), For the i The predicted probability of the sample (the probability value output by the model, range [0, 1]).

[0119] The Adam optimizer is used to update the model parameters (including attention weights and neural network weights) to minimize the loss, and the training is repeated for 100 iterations.

[0120] In the network structure, for different input feature data, the attention mechanism module will assign different weights to each feature, highlighting the features that have a greater impact on the health status of the tool. Suppose the input feature vector is , the attention weight vector is , the feature vector after attention mechanism processing is , and then Input to subsequent neural network layers for training and prediction.

[0121] During the training process, the weights of the neural network are constrained by combining a pruning method based on L1 regularization. When the absolute value of a weight is less than a certain threshold (such as 0.01), it is set to 0, removing unimportant connections and neurons, reducing overfitting and improving the accuracy of the prediction model.

[0122] Specifically, the cost prediction model established by using the multivariate adaptive regression spline algorithm in step S507 includes: In step S5073, the tool purchase price, maintenance cost, energy consumption cost, and storage cost are used as input features, and the purchase cost and management cost are used as output labels. The relationship between the input features and the output labels is learned using a multivariate adaptive regression spline algorithm model, and combined with time series analysis, a cost prediction model is established.

[0123] In step S5074, the cost prediction model is trained, and the regression coefficient is solved using the least square method, with the objective function being to minimize the sum of squares of the error between the predicted value output by the cost prediction model and the actual cost value, until the cost prediction model converges.

[0124] In an embodiment of the present invention, procurement and management costs are analyzed and predicted to determine key factors affecting procurement and management costs, including tool purchase price, service life, maintenance cost, energy consumption cost, storage cost and other characteristics.

[0125] A cost prediction model was constructed using the Multivariate Adaptive Regression Spline (MARS) algorithm to identify key factors influencing a purchaser's management costs, including tool purchase price, expected service life, annual maintenance cost, annual energy cost, and annual storage cost. Using tool purchase price, expected service life, annual maintenance cost, annual energy cost, and annual storage cost as input features, and total purchase and management cost as the output label, the model learned the complex relationship between these features and costs. Combined with time series analysis, a cost prediction model was established to predict tool purchase and management costs.

[0126] According to the actual needs of the enterprise, it helps the inventory management module to reasonably formulate procurement plans, provide decision-making basis for the procurement department, optimize procurement strategies, reduce management costs, and ensure that the procurement process complies with tool equipment standards and cost control requirements.

[0127] Specifically, the multivariate adaptive regression spline algorithm model predicts the total purchase and management cost of electric drills based on cost data, assuming that the influencing factors include tool purchase price P , service life L life , maintenance costs C repair , energy consumption cost C energy , storage costs C storage etc., the total output procurement and management cost is C , the MARS model is expressed as:

[0128] in, is the regression coefficient, It is composed of data features X=[ P , Llife , C repair , C energy , C storage ]The basis functions generated by the adaptive selection of basis functions and their combinations can more flexibly and accurately fit the complex nonlinear relationship between cost and various influencing factors.

[0129] For example, we collected data on 100 different models of electric drills, including information such as purchase price, expected service life, annual maintenance cost, and annual storage cost. We first cleaned the data to remove obvious errors or anomalies. We then normalized the data using the Z-Score normalization formula, for example, to calculate the purchase price range of 200-500 yuan and the service life of 1-5 years. We then divided the data into 70% of the training set and 30% of the test set.

[0130] The model is trained using the training set data, and the regression coefficient is solved by the least squares method. The objective function is to minimize the sum of squares of the errors between the predicted value output by the cost prediction model and the actual cost value. The parameters are continuously adjusted during the training process until the model converges.

[0131] Specifically, the support vector machine regression algorithm is used in the above step S507 to construct the remaining life prediction model, which includes: Step S5075: Divide the historical data set into a training set and a test set according to a preset ratio.

[0132] Step S5076: Use the SVR model for training, map the four-dimensional features through the RBF kernel function, and filter the support vectors through the Lagrange multiplier method.

[0133] Step S5077: Iteratively learn and optimize the parameters of the model through the training set, and adjust the penalty parameters and the parameters of the RBF kernel function.

[0134] Step S5078, calculate the loss function value of the model on the training set.

[0135] Step S5079: Use the test set to evaluate the trained model to obtain a remaining life prediction model.

[0136] In this embodiment of the present invention, for consumable tools, in addition to collecting conventional data such as operating current, temperature, environmental conditions, and operating conditions, a focus is placed on collecting usage data directly related to wear and tear. Using a support vector regression (SVR) algorithm, this algorithm uses data such as wear, usage time, surface damage, and load impact as input labels and the tool's remaining life as the output label. By training and learning from a large amount of historical data, a remaining life prediction model for the tool is constructed to predict future wear and tear.

[0137] Taking electric drill bits as an example, a remaining life prediction model was trained based on the support vector machine regression algorithm. The input features were material hardness H, usage time t, load impact strength F, and drill bit temperature T. The output label was the drill bit's remaining life. Relevant feature data was collected and preprocessed, and standardized data was integrated to form a historical dataset. This dataset was randomly divided into a training set (70 items) and a test set (30 items) in a 7:3 ratio. The SVR (Support Vector Machine Regression) model was used for training.

[0138] The RBF kernel function (Radial Basis Function Kernel) is used to perform nonlinear mapping on the four-dimensional features (material hardness H, usage time t, load impact strength F, and drill bit temperature T) to make them linearly separable in high-order space, and the Lagrange multiplier method is used to screen the decisive support vectors.

[0139] The parameters are iteratively learned and optimized through the training set, and the parameters C (penalty parameter) and γ (parameter of the RBF kernel function) are adjusted.

[0140] Calculate the loss function value of the model on the training set. The loss function is as follows:

[0141] in, MSE is the loss function value, is the true label, is the predicted value.

[0142] The trained model is evaluated using the test set to form a remaining life prediction model that conforms to the actual situation.

[0143] The remaining life prediction model is:

[0144] in, Tpred is the remaining life, X is the input feature vector, 、 is the Lagrange multiplier, is the RBF kernel function,b is the bias term.

[0145] Based on the model's predictions, operators can be provided with optimized usage recommendations, revised maintenance and replacement plans, and implemented quality standards and safe operating procedures. During procurement, a rational purchasing plan can be developed based on the predicted tool lifespan, avoiding inventory overstocks or shortages and ensuring compliance with tool allocation standards.

[0146] The cost prediction model constructed through the multivariate adaptive regression spline algorithm can accurately fit the nonlinear relationship between cost and multiple factors, providing strong support for enterprise cost control. Combining the improved DBSCAN algorithm with the enhanced Apriori association rule mining algorithm, it can achieve rapid classification of tools and equipment, and mine the complex coupling relationship between various types of data in the operation of tools and equipment, such as the relationship between midday scenes, high temperatures, high vibration frequencies and faults. The health status prediction model of tools and equipment is constructed based on the neural network algorithm of the attention mechanism, combined with pruning technology to reduce overfitting, helping to deeply understand the operating status of tools and equipment. The support vector regression algorithm remaining life prediction model for consumable tools accurately predicts wear and tear and optimizes usage, maintenance and procurement plans. Based on the predicted indicators such as tool procurement management cost, performance status, and remaining life, a reliable basis is provided for the adjustment of tool procurement, usage and management strategies to ensure the full implementation of the "three standards".

[0147] Step S508: Make tool management decisions based on the predicted tool purchase and management costs, health status, and tool wear and tear, combined with actual production needs.

[0148] Specifically, the above step S508 includes: Step S5081: Formulate a tool procurement management strategy based on the predicted tool procurement and management costs, combined with tool failure frequency and inventory information.

[0149] Step S5082: Based on the predicted health status of the tool and in combination with the tool usage and maintenance data, formulate tool usage strategy optimization suggestions.

[0150] Step S5083: formulate a tool maintenance plan based on the predicted tool wear and tear and the tool maintenance cycle.

[0151] In an embodiment of the present invention, the procurement and management decision-making planning function is utilized, with the help of the inventory management module and the cost forecasting model, the failure frequency of tools, inventory status and information on new tools on the market are comprehensively considered to formulate scientific and reasonable procurement and management strategies, including decisions on procurement quantity, procurement model, supplier selection, maintenance plan and storage strategy, to avoid inventory backlogs or out-of-stock, to meet tool equipment standards and optimize cost management.

[0152] By utilizing the tool usage strategy optimization function, combined with the health status obtained from the usage and maintenance module and the health status prediction model, operators are provided with optimization suggestions for tool usage strategies to ensure that tools are used in an environment that meets quality standards.

[0153] Leveraging the standardized operation function and the training and assessment module, operators are trained and assessed before using tools and equipment, ensuring they are proficient in operating skills and safety precautions. Leveraging image recognition technology in the use and maintenance module, operator behavior is monitored and recorded in real time, strengthening their awareness of safety regulations and ensuring strict implementation of safety operating procedures.

[0154] By utilizing the maintenance cycle adjustment function, we can rationally arrange maintenance plans based on association rules and prediction models when the risk of tool failure increases, ensuring that tools meet quality standards.

[0155] The system also includes a monitoring and maintenance module, which monitors performance indicators and data quality at all levels of the system. Performance indicators include sensor node battery charge, signal strength, data transmission success rate, data transmission latency, algorithm runtime, memory usage, and user operation response time. Data quality includes data integrity, accuracy, and consistency. When performance indicators are abnormal or data quality issues arise, appropriate optimization and repair measures are implemented to ensure stable system operation and the continued effective implementation of the "Three Standards and One Process."

[0156] The tool management and control method based on the Internet of Things and the enhanced Apriori algorithm provided in this embodiment has the following beneficial effects: (1) With the help of sensors and RFID systems in the perception layer of the Internet of Things, real-time and accurate monitoring and precise positioning and tracking of tools can be achieved. Sensors collect working condition data and accurately grasp the operating status of tools; RFID tags facilitate rapid positioning, optimize inventory management, prevent asset loss, and improve resource allocation efficiency. By analyzing tool usage and inventory data, the system can optimize resource allocation based on usage frequency, idle status, etc., and implement tool equipment standards; at the same time, it can reasonably arrange upgrades based on performance predictions to ensure the applicability of tools and dynamically optimize resource utilization; (2) Ensure that intelligent algorithms can deeply mine data, predict the health status trend of tools, and warn of potential faults in advance, prevent them from happening in time, and ensure the safety of personnel and property. According to the requirements of safety operation specifications, systematically regulate the behavior of operators, train and assess operators before use, push operation guides and safety specifications when using, and record behavior data during use to facilitate the correction of bad habits and illegal operations, reduce human errors, and reduce the risk of safety accidents; (3) Based on intelligent algorithm-based fault prediction and diagnosis, maintenance can be arranged in advance to reduce tool downtime, ensure that tools always meet quality standards, and guarantee production continuity; quickly locate the cause of a fault when it occurs, assist maintenance personnel in efficient repairs, shorten maintenance time, and improve tool utilization; grasp the location and status of tools in real time, and quickly deploy them according to production needs; at the same time, share tool information, enhance inter-departmental collaboration, improve work synergy, and improve overall production efficiency; (4) With standardized use management as the core, a method for intelligent control of the entire tool process is proposed that integrates dynamic working condition adaptation and spatiotemporal semantic mining algorithms. This solution significantly improves the accuracy of fault prediction and maintenance strategy in the tool use management process by dynamically adjusting the algorithm threshold to adapt to real-time working conditions and combining spatiotemporal semantics to enhance association rule mining. It provides a data basis for the dynamic optimization of the "three standards and one process" and provides an effective technical solution for tool control. (5) After analyzing the massive amount of tool data accumulated, it provides a comprehensive and accurate basis for corporate decision-making. For example, when making purchasing decisions, reference can be made to data such as failure rate and cost-performance ratio to select the best supplier and model, thereby reducing procurement costs. Regular data review can identify management problems, such as tools with high failure rates. After analyzing the causes, targeted improvements can be made to optimize management strategies. Strategies can also be adjusted based on industry trends and new technologies, combined with data feedback, to ensure the implementation of the "three standards and one process" and maintain corporate competitiveness.

[0157] like Figure 3 As shown, Figure 3 This is the overall architecture diagram of the tool management and control system based on the Internet of Things and the enhanced Apriori algorithm. A system performance monitoring module is established to monitor the performance indicators of the Internet of Things perception layer, data transmission layer, data storage layer, intelligent analysis layer, and application service layer in real time. In the Internet of Things perception layer, the battery level, signal strength, and data transmission success rate of sensor nodes are monitored. In the data transmission layer, data transmission latency and bandwidth utilization are monitored. In the intelligent analysis layer, attention is paid to the algorithm's runtime and memory usage. In the data storage layer, the stability and read and write speed of data storage are monitored. In the application service layer, statistics are collected on the response time of user operations and the system's throughput. When abnormal performance indicators are detected, timely alarms are issued and corresponding optimization measures are taken, such as replacing sensor batteries, adjusting network configurations, and optimizing algorithm parameters, to ensure the stable operation of the system and the continuous and effective implementation of the "three standards and one process."

[0158] Data quality maintenance: Regularly perform quality checks and maintenance on the data in the data storage system. Check the integrity of the data to ensure that all necessary data fields have values; check the accuracy of the data to ensure that the data conforms to the actual situation; check the consistency of the data to ensure that the data between different tables matches each other. Compare the basic information table and the usage record table of the tools to ensure that the tool number in the usage record is consistent with the number in the basic information table. For any data quality problems found, repair them in a timely manner, such as supplementing missing data, correcting erroneous data, and cleaning up duplicate data. Establish a data quality assessment mechanism, regularly evaluate data quality, continuously improve data collection and management processes, improve data quality, and provide reliable data guarantees for the effective implementation of the "three standards and one process".

[0159] In this embodiment, a tool control device based on the Internet of Things and an enhanced Apriori algorithm is also provided. The device is used to implement the above-mentioned embodiments and preferred implementation methods, and the details that have been explained will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0160] This embodiment provides a tool control device based on the Internet of Things and the enhanced Apriori algorithm, such as Figure 5 Shown, including: The data collection unit 501 is used to collect tool data by combining the Internet of Things RFID radio frequency technology and wireless sensors. The tool data includes tool usage data and tool environment data.

[0161] The data analysis unit 502 is used to perform dynamic cluster analysis on tool data using the improved DBSCAN algorithm, filter out frequent item sets from the dynamic cluster analysis results using the spatiotemporal semantic enhancement Apriori algorithm, generate association rules based on the frequent item sets, and embed the spatiotemporal semantic enhancement Apriori algorithm into the three-dimensional semantic labels of time, space, and working conditions.

[0162] The prediction unit 503 is used to establish a cost prediction model using a multivariate adaptive regression spline algorithm with frequent item sets and association rules as data sets, to build a health status prediction model based on a neural network model with an attention mechanism, and to build a remaining life prediction model using a support vector machine regression algorithm. The cost prediction model is used to predict the procurement and management costs of tools, the health status prediction model is used to predict the health status of tools, and the remaining life prediction model is used to predict tool wear and tear.

[0163] The decision-making unit 504 is used to make tool management decisions based on the predicted tool procurement and management costs, health status, and wear and tear, combined with actual production needs. The further functional description of each of the above modules and units is the same as that of the corresponding embodiment above and will not be repeated here.

[0164] The tool control device based on the Internet of Things and the enhanced Apriori algorithm in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0165] The embodiment of the present invention also provides a computer device having the above Figure 5 The tool control device based on the Internet of Things and enhanced Apriori algorithm is shown.

[0166] See also Figure 6 , Figure 6 is a structural diagram of a computer device provided by an optional embodiment of the present invention, such as Figure 6 As shown, the computer device includes: one or more processors 10, memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 6 A processor 10 is taken as an example.

[0167] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0168] The memory 20 stores instructions that can be executed by at least one processor 10, so as to enable at least one processor 10 to execute the method shown in the above embodiment.

[0169] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0170] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0171] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 can be connected via a bus or other means. Figure 6 The bus connection is taken as an example.

[0172] The input device 30 can receive input digital or character information and generate key signal input related to user settings and function control of the computer device, such as a touch screen, etc. The output device 40 can include a display device, etc.

[0173] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0174] A portion of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0175] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are intended to fall within the scope of this application.

Claims

1. A tool management and control method based on the Internet of Things and enhanced Apriori algorithm, characterized in that: The method comprises: Combine the Internet of Things (IoT) RFID radio frequency technology and wireless sensors to collect tool data, including tool usage data and tool environment data; The improved DBSCAN algorithm is used to perform dynamic cluster analysis on the tool data, and the spatiotemporal semantic enhancement Apriori algorithm is used to filter frequent item sets from the dynamic cluster analysis results, and association rules are generated based on the frequent item sets. The spatiotemporal semantic enhancement Apriori algorithm embeds three-dimensional semantic labels of time, space, and working conditions; Using the frequent item sets and the association rules as data sets, a cost prediction model is established using a multivariate adaptive regression spline algorithm, a health status prediction model is constructed based on a neural network model with an attention mechanism, and a remaining life prediction model is constructed using a support vector machine regression algorithm. The cost prediction model is used to predict the procurement and management costs of tools, the health status prediction model is used to predict the health status of tools, and the remaining life prediction model is used to predict tool wear and tear; Make tool management decisions based on the predicted procurement and management costs, health status, and losses of tools and equipment, combined with actual production needs.

2. The method according to claim 1, characterized in that The improved DBSCAN algorithm model is used to perform dynamic cluster analysis on the tool data, and the spatiotemporal semantic enhanced Apriori algorithm is used to filter out frequent item sets from the dynamic cluster analysis results, and association rules are generated based on the frequent item sets, including: A dynamic working condition adaptation mechanism is introduced, and the improved DBSCAN algorithm model is used to perform dynamic cluster analysis on the tool data, where the cluster labels include power tools, hand tools, arc welding tools, and high-frequency electrical tools; The importance factor of spatiotemporal working conditions is introduced, a weighted support function is constructed, and frequent itemsets are obtained from the dynamic clustering analysis results. Integrate spatiotemporal working condition semantics to mine association rules that meet confidence requirements.

3. The method according to claim 2, characterized in that Improve the DBSCAN algorithm by following these steps: Introducing working condition perception, analyzing working condition data collected by wireless sensors, and calculating characteristic fluctuation rate indicators. The characteristic fluctuation rate indicators are used to characterize the severity of the tool operation. The working condition data include current fluctuations, temperature gradients, and vibration frequencies. Based on the characteristic volatility index, nonlinear mapping is performed to construct a density threshold adjustment function, and a reference neighborhood radius and minimum point number parameter combination adapted to the working condition data is output; The DBSCAN algorithm is improved based on the output parameter combination of benchmark neighborhood radius and minimum number of points.

4. The method according to claim 2, characterized in that The frequent itemsets obtained by screening from the dynamic clustering analysis results include: Embed three-dimensional semantic labels into the dynamic clustering analysis results to obtain candidate item sets; Calculate the support of candidate item sets and filter the item sets with support greater than the preset support threshold as frequent item sets.

5. The method according to claim 1, characterized in that The neural network model based on the attention mechanism constructs a health status prediction model, including: Construct a neural network model, which includes an input layer, a hidden layer, and an output layer. The input layer receives the feature vector after four-dimensional preprocessing. There are two hidden layers, each with 32 neurons, using a ReLU activation function, and the output layer using a Sigmoid activation function. The feature vector processed by the attention mechanism is input into the neural network model for model training. Based on L1 regularization pruning, the weight of the neural network model is constrained to obtain a health status prediction model.

6. The method according to claim 1, characterized in that The method of establishing a cost prediction model using a multivariate adaptive regression spline algorithm includes: Using tool purchase price, maintenance cost, energy cost, and storage cost as input features, and purchase cost and management cost as output labels, a multivariate adaptive regression spline algorithm model is used to learn the relationship between input features and output labels. Combined with time series analysis, a cost prediction model is established. The cost prediction model is trained, and the regression coefficient is solved using the least squares method. The objective function is to minimize the sum of squares of the errors between the predicted value output by the cost prediction model and the actual cost value until the cost prediction model converges.

7. The method according to claim 1, characterized in that The support vector machine regression algorithm is used to construct a remaining life prediction model, including: Divide the historical data set into training set and test set according to the preset ratio; Use the SVR model for training, map the four-dimensional features through the RBF kernel function, and filter the support vectors through the Lagrange multiplier method; The model parameters are iteratively learned and optimized through the training set, and the penalty parameters and RBF kernel function parameters are adjusted; Calculate the loss function value of the model on the training set; The trained model is evaluated using the test set to obtain the remaining life prediction model.

8. The method according to claim 1, characterized in that Based on the predicted tool procurement and management costs, health status, and tool wear and tear, combined with actual production needs, tool management decisions are made, including: Develop a tool procurement management strategy based on the predicted tool procurement and management costs, combined with tool failure frequency and inventory information; Based on the predicted health status of tools and equipment and combined with tool usage and maintenance data, we formulate optimization suggestions for tool usage strategies; Based on the predicted tool loss and tool maintenance cycle, a tool maintenance plan is formulated.

9. The method according to claim 1, characterized in that After collecting tool data by combining the Internet of Things RFID radio frequency technology and wireless sensors, the method further includes: According to the application scenario and data transmission requirements, the corresponding transmission method is selected, and a data transmission gateway is set to transmit the tool data.

10. The method according to claim 1, characterized in that After collecting tool data by combining the Internet of Things RFID radio frequency technology and wireless sensors, the method further includes: Deleting outliers and duplicate values ​​in the tool data and filling missing data using linear interpolation; Extracting features from the tool data, the extracted features including tool usage frequency, usage duration, and fault interval duration; The tool data is normalized and standardized.

Images