Data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system
By using multi-dimensional sensors and intelligent modeling technology, the problems of lagging maintenance mode and reliance on human experience in intelligent manufacturing equipment have been solved, realizing intelligent fault prediction and maintenance, and improving equipment operation stability and maintenance efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- TANAC AUTOMATION
- Filing Date
- 2026-01-28
- Publication Date
- 2026-06-05
AI Technical Summary
The existing maintenance model for intelligent manufacturing equipment mainly relies on periodic maintenance and post-event repairs, which lacks specificity and leads to over-maintenance or under-maintenance. The reliance on human experience results in low accuracy of fault prediction and an inability to avoid potential faults in a timely manner, causing economic losses and product quality problems.
By employing multi-dimensional sensor modules, data acquisition and preprocessing, FMEA reliability assessment, fault prediction modeling and intelligent maintenance strategy generation units, combined with Weibull distribution and multi-algorithm collaborative modeling, early fault warning and personalized maintenance plans can be achieved, thereby reducing equipment failure rate.
It enables early warning of faults, reduces downtime losses, improves maintenance efficiency, reduces reliance on manual labor, optimizes maintenance strategies, ensures long-term equipment stability, and provides data traceability.
Smart Images
Figure CN122155673A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent manufacturing equipment reliability technology, and discloses a data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system. Background Technology
[0002] In the context of the large-scale advancement of intelligent manufacturing, equipment, as the core carrier of production activities, directly determines production efficiency, product quality, and production costs through its continuous and stable operation. Currently, equipment maintenance models in the field of intelligent manufacturing still face the following problems: Current maintenance methods mainly rely on periodic maintenance and reactive repair. Periodic maintenance adopts a one-size-fits-all fixed-cycle strategy, which lacks targeted consideration of the actual operating status of the equipment, and is prone to over-maintenance (resulting in a waste of human and material resources) or under-maintenance (failing to avoid potential failures in a timely manner). Reactive repair, on the other hand, depends on reactive handling after a failure occurs, and the failure warning is seriously delayed. Once the equipment suddenly fails, it often causes the production line to stop, which not only causes huge economic losses, but may also lead to a chain of problems such as deviations in product processing accuracy and batch quality defects.
[0003] Current fault diagnosis and maintenance strategies rely heavily on human experience. Maintenance personnel must use their personal experience to determine fault types and causes, making it difficult to comprehensively cover potential fault modes under complex equipment operating conditions, and hindering the scientific modeling and prediction of equipment and critical component lifespans. Furthermore, the lack of in-depth analysis of historical fault data and intelligent optimization mechanisms for maintenance strategies results in persistently high equipment failure rates, low maintenance efficiency, and severely restricts the high-quality development of the intelligent manufacturing industry. Summary of the Invention
[0004] The core objective of this invention is to provide a data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system. This system addresses the technical problems of outdated equipment maintenance models, low fault prediction accuracy, lack of scientific maintenance strategies, and over-reliance on human experience. Through multi-dimensional data collection, reliability assessment, multi-algorithm fusion modeling, and collaborative design of intelligent maintenance decision-making, the system enables early fault warning, personalized maintenance plan formulation, and precise maintenance guidance. Ultimately, this reduces equipment failure rates and downtime losses, and improves maintenance efficiency and equipment operational stability.
[0005] To achieve the above objectives, this invention provides a data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system, including a multi-dimensional sensor module, a data acquisition and preprocessing unit, an FMEA reliability assessment unit, a fault prediction modeling unit, an intelligent maintenance strategy generation unit, and an early warning execution unit. Each unit is sequentially connected to form a complete fault prediction and maintenance closed loop.
[0006] Preferably, the multi-dimensional sensor module includes a vibration sensor, a temperature sensor, and an operating parameter sensor, which are respectively deployed in key parts of the equipment to collect data on equipment vibration amplitude, temperature rise, core component rotation speed, and load status parameters in real time.
[0007] Preferably, the FMEA reliability assessment unit includes a fault mode identification subunit and a fault cause analysis subunit. The fault mode identification subunit identifies potential fault types such as mechanical wear and circuit overheating based on preprocessed equipment status data. The fault cause analysis subunit analyzes the root causes of faults such as component aging and insufficient lubrication in conjunction with the equipment structure and operating principle.
[0008] Preferably, the data acquisition and preprocessing unit uses a wavelet denoising algorithm to denoise the original acquired data, and simultaneously performs outlier removal and data standardization operations to remove environmental interference and data noise, providing high-quality data input for subsequent units.
[0009] Preferably, the fault prediction modeling unit has a self-learning function. After each equipment maintenance operation is completed, it automatically optimizes the model parameters based on the fault handling data and maintenance records of this operation, thereby continuously improving the accuracy of fault diagnosis.
[0010] Preferably, the intelligent maintenance strategy generation unit is based on the CBM strategy, the maintenance plan formulation subunit formulates personalized maintenance plans based on real-time equipment status data, fault prediction results and life modeling conclusions, and the maintenance cycle optimization subunit optimizes the maintenance cycle by statistically analyzing historical fault records of the equipment to mine the fault patterns of key components.
[0011] Preferably, the alarm subunit of the early warning execution unit immediately issues an audible and visual alarm when the fault prediction modeling unit identifies a fault risk, and the maintenance guidance push subunit pushes a targeted maintenance operation guide containing maintenance plans, operating procedures and safety precautions to maintenance personnel.
[0012] Preferably, the life modeling subunit of the fault prediction modeling unit uses Weibull distribution to perform life modeling on the overall equipment and key components respectively, so as to accurately predict the remaining service life of the equipment and each key component.
[0013] Preferably, the fault identification algorithm module of the fault prediction modeling unit uses the collaborative operation of decision tree algorithm and neural network algorithm to mine the correlation between equipment status data and fault type and fault degree, so as to realize the early identification and accurate prediction of potential faults.
[0014] Preferably, the present invention further includes a data storage unit for storing the raw data collected by the multi-dimensional sensor module, the processed data of the data acquisition and preprocessing unit, the FMEA reliability assessment results, the fault prediction data and maintenance records, supporting data traceability and subsequent analysis and retrieval.
[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Early warning of faults to reduce downtime losses: By collecting data in real time through multi-dimensional sensors, combined with FMEA reliability assessment and Weibull distribution + dual algorithm collaborative modeling, faults can be predicted 3-7 days in advance (the specific time is adjusted according to the equipment type), effectively avoiding production line downtime caused by sudden faults and reducing economic losses and product quality defects caused by downtime. 2. Personalized maintenance strategies reduce maintenance costs. Based on the CBM condition-based maintenance strategy, combined with real-time equipment status, fault prediction results and historical fault patterns, targeted maintenance plans are formulated and maintenance cycles are dynamically optimized to avoid the problems of over-maintenance and under-maintenance in traditional periodic maintenance, thereby reducing maintenance costs. 3. Reduce reliance on manual labor and improve maintenance efficiency. The early warning execution unit pushes precise maintenance guidance, providing maintenance personnel with clear operating guidelines. Efficient repairs can be completed without relying on extensive manual experience, improving fault resolution efficiency by more than 40% and reducing the error rate of maintenance operations. 4. Data-driven continuous optimization ensures long-term stability. The system has self-learning capabilities and data storage units, which can continuously iterate fault prediction models and maintenance strategies based on actual operating data. As the operating time increases, the accuracy of fault prediction improves, ensuring the long-term stability of equipment operation. 5. Data traceability facilitates management decisions. Through the data storage unit, all types of data can be traced throughout the entire process. Managers can keep track of equipment operating status, maintenance history, and failure patterns in real time, providing a scientific basis for decision-making regarding equipment upgrades and production plan adjustments. Attached Figure Description
[0016] Figure 1 This is a block diagram of the overall system architecture of the present invention; Figure 2 This is a block diagram of the internal structure of the multi-dimensional sensor module of the present invention; Figure 3 This is a block diagram of the internal structure of the FMEA reliability assessment unit of the present invention; Figure 4 This is a block diagram of the internal structure of the fault prediction modeling unit of the present invention; Figure 5 This is a block diagram of the internal structure of the early warning execution unit of the present invention. Detailed Implementation
[0017] Please refer to Figures 1 to 5 This invention provides a data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system, including a multi-dimensional sensor module, a data acquisition and preprocessing unit, an FMEA reliability assessment unit, a fault prediction modeling unit, an intelligent maintenance strategy generation unit, an early warning execution unit, and a data storage unit. Each unit is connected in sequence to form a complete closed loop of data acquisition, processing, evaluation, prediction, decision-making, execution, storage, and iteration.
[0018] Multi-dimensional sensor module: As the core of the system's data acquisition, the multi-dimensional sensor module includes vibration sensors, temperature sensors, and operating parameter sensors, all deployed in key parts of the equipment (such as the spindle, motor, transmission mechanism, circuit control box, and other core components). These sensors are used to collect multi-dimensional status data during equipment operation in real time: vibration sensors collect vibration characteristic parameters such as vibration amplitude and frequency; temperature sensors collect temperature rise data of core components and ambient temperature data; and operating parameter sensors collect operating status parameters of core components such as rotational speed, load, voltage, and current, providing comprehensive and real-time data support for subsequent fault prediction and reliability assessment.
[0019] Data acquisition and preprocessing unit: This unit connects to the multi-dimensional sensor module to receive raw data collected by the sensors. It employs a wavelet denoising algorithm to denoise the raw data, effectively filtering out irrelevant noise such as environmental vibrations and electromagnetic interference. Simultaneously, it uses the 3σ criterion to remove outliers caused by sensor malfunctions or transmission interference during data acquisition. Finally, it performs data standardization (converting the data to the [0,1] range) to ensure comparability of data of different types and magnitudes, providing high-quality, highly reliable data input for subsequent FMEA reliability assessment and fault prediction modeling.
[0020] FMEA Reliability Assessment Unit: Based on the preprocessed high-quality data, a comprehensive reliability assessment of the equipment is conducted using FMEA (Failure Mode and Effects Analysis) technology. This unit includes a failure mode identification subunit and a failure cause analysis subunit. Fault Mode Recognition Subunit: By extracting feature vectors from preprocessed data (such as abnormal peak vibration amplitude, sudden temperature rise trend, and fluctuation range of operating parameters), and combining them with a preset fault feature library, it automatically identifies potential fault types of equipment, including mechanical wear, transmission mechanism jamming, circuit overheating, and component aging. Fault Cause Analysis Subunit: Combining equipment structural design parameters, operating principles, and historical fault data, the root cause of identified fault types is traced, and the core factors causing the faults are analyzed, such as insufficient lubrication, overload, voltage instability, and component fatigue wear, providing a basis for the formulation of subsequent maintenance strategies.
[0021] Fault prediction modeling unit: As the core computing unit of the system, it has self-learning capabilities and includes a life modeling subunit and a fault identification algorithm module: Life modeling sub-unit: The Weibull distribution is used to model the life of the entire equipment and key components (such as spindle, bearing, motor winding, etc.). Data such as equipment running time, maintenance records, and fault history are input. The remaining service life of the equipment and each key component is calculated through parameter estimation (shape parameters, dimensional parameters), providing a quantitative basis for determining the maintenance timing. Fault identification algorithm module: Integrates decision tree algorithm and neural network algorithm. The decision tree algorithm can quickly screen key influencing factors of faults and reduce data dimensionality. The neural network algorithm (such as BP neural network) can be used to explore the non-linear correlation between equipment status data and fault type and fault degree to achieve early identification and accurate classification and prediction of potential faults (such as minor faults, moderate faults, and severe faults). Self-learning function: After each equipment maintenance operation is completed, the system automatically feeds back the fault handling data and maintenance records (such as maintenance location, replaced parts, and maintenance effect) to the fault prediction modeling unit, iteratively optimizes the model parameters (such as the weight coefficients of the neural network and the parameters of the Weibull distribution), and continuously improves the accuracy of fault diagnosis and prediction.
[0022] Intelligent maintenance strategy generation unit: Based on the CBM (State-Based Maintenance) strategy, a scientific and efficient maintenance plan is developed, including a maintenance planning sub-unit and a maintenance cycle optimization sub-unit. Maintenance plan formulation subunit: Based on real-time equipment status data, fault prediction results (fault type, fault level, probability of occurrence) and life modeling conclusions (remaining service life), it automatically formulates personalized maintenance plans, specifying core information such as maintenance time, maintenance location, required tools, and replacement parts models, to avoid unnecessary maintenance operations; Maintenance cycle optimization subunit: By statistically analyzing the equipment's historical fault records, we can uncover the fault patterns of key components (such as the duration of high-incidence operation and environmental factors), and combine this with the remaining service life prediction results to dynamically optimize the maintenance cycle, ensuring that the maintenance timing is neither too early (avoiding over-maintenance) nor too late (avoiding under-maintenance).
[0023] Early warning execution unit: It works in conjunction with the fault prediction modeling unit and the intelligent maintenance strategy generation unit, including an alarm subunit and a maintenance guidance push subunit: Alarm Subunit: When the fault prediction modeling unit identifies a fault risk (especially a moderate or severe fault risk), it immediately issues an audible and visual alarm (such as a flashing red warning light and a buzzer alarm), and simultaneously displays the warning information on the equipment control system and maintenance personnel's terminal to remind staff to respond in a timely manner; Maintenance guidance push sub-unit: The personalized maintenance plan, step-by-step operation steps, safety precautions (such as power-off operation requirements, component installation accuracy standards) and solutions to common problems generated by the intelligent maintenance strategy generation unit are pushed to maintenance personnel in the form of text, pictures and videos, providing accurate guidance for maintenance work.
[0024] Data storage unit: This system is used to classify and store various types of data during system operation, including: raw data collected by multi-dimensional sensor modules, processed data from data acquisition and preprocessing units, FMEA reliability assessment results (failure modes, failure causes), failure prediction data (failure level, remaining service life), and maintenance records (maintenance plan, execution status, maintenance effect). The data adopts a distributed storage architecture, supports data traceability and historical data analysis, and provides data support for failure prediction model optimization and maintenance strategy iteration.
Claims
1. A data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system, characterized in that: It includes a multi-dimensional sensor module, a data acquisition and preprocessing unit, an FMEA reliability assessment unit, a fault prediction modeling unit, an intelligent maintenance strategy generation unit, and an early warning execution unit. Each unit is connected in sequence to form a complete fault prediction and maintenance closed loop.
2. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The multi-dimensional sensor module includes vibration sensors, temperature sensors, and operating parameter sensors, which are deployed in key parts of the equipment to collect data on equipment vibration amplitude, temperature rise, core component rotation speed, and load status parameters in real time.
3. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The FMEA reliability assessment unit includes a fault mode identification subunit and a fault cause analysis subunit. The fault mode identification subunit identifies potential fault types such as mechanical wear and circuit overheating based on preprocessed equipment status data. The fault cause analysis subunit analyzes the root causes of faults such as component aging and insufficient lubrication in combination with equipment structure and operating principle.
4. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The data acquisition and preprocessing unit uses wavelet denoising algorithm to denoise the original acquired data, and simultaneously performs outlier removal and data standardization operations to remove environmental interference and data noise, providing high-quality data input for subsequent units.
5. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The fault prediction modeling unit has a self-learning function. After each equipment maintenance operation is completed, it automatically optimizes the model parameters based on the fault handling data and maintenance records to continuously improve the accuracy of fault diagnosis.
6. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The intelligent maintenance strategy generation unit is based on the CBM strategy. The maintenance plan formulation subunit formulates personalized maintenance plans based on real-time equipment status data, fault prediction results, and life modeling conclusions. The maintenance cycle optimization subunit optimizes the maintenance cycle by statistically analyzing historical fault records of the equipment to uncover fault patterns of key components.
7. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The alarm subunit of the early warning execution unit immediately issues an audible and visual alarm when the fault prediction modeling unit identifies a fault risk, and the maintenance guidance push subunit pushes a targeted maintenance operation guide containing maintenance plans, operating procedures and safety precautions to maintenance personnel.
8. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The life modeling subunit of the fault prediction modeling unit uses Weibull distribution to perform life modeling on the overall equipment and key components, accurately predicting the remaining service life of the equipment and each key component.
9. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: The fault identification algorithm module of the fault prediction modeling unit uses the collaborative operation of decision tree algorithm and neural network algorithm to mine the correlation between equipment status data and fault type and fault degree, so as to realize the early identification and accurate prediction of potential faults.
10. The data-driven intelligent manufacturing equipment fault prediction and intelligent maintenance system according to claim 1, characterized in that: It also includes a data storage unit for storing raw data collected by the multi-dimensional sensor module, processed data from the data acquisition and preprocessing unit, FMEA reliability assessment results, fault prediction data, and maintenance records, supporting data traceability and subsequent analysis.