Equipment health degree algorithm
By using equipment health algorithms, a multi-dimensional evaluation system is constructed to dynamically calculate equipment health scores and generate decision support reports. This solves the problem of equipment failure prediction relying on post-event maintenance and regular inspections, and realizes a comprehensive assessment of equipment health status and optimization of operation and maintenance costs.
Patent Information
- Application Number
- CN202511555566.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-29
- Publication Date
- 2026-02-27
AI Technical Summary
In existing technologies, equipment failure prediction relies on post-failure maintenance and periodic inspections, which leads to resource waste and high maintenance pressure. It cannot effectively predict equipment failures, resulting in high maintenance costs.
By employing equipment health algorithms, and through multi-source data acquisition, data preprocessing, multi-dimensional health assessment models, and intelligent early warning, a comprehensive assessment system is constructed to evaluate performance, alarms, lifespan, maintenance, spare parts, cost, and safety. The system dynamically calculates equipment health scores and generates decision support reports.
It enables a comprehensive and multi-dimensional assessment of equipment health status, improves the accuracy and reliability of fault prediction, reduces unplanned downtime, optimizes resource utilization, lowers operation and maintenance costs, and enhances management standardization and efficiency.
Smart Images
Figure CN121581835A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of industrial equipment operation and maintenance management, and in particular to a device health degree algorithm. BACKGROUND
[0002] With the continuous development of industrial technology, the complexity and integration of modern industrial production equipment have been significantly improved, and various devices have been gradually invented to greatly facilitate industrial development. At the same time, as time goes by, more and more devices enter the aging and maintenance stage, and from time to time, device failure or damage occurs, and the requirements for device maintenance are becoming higher and higher, and the maintenance tasks are becoming more and more heavy. At the same time, the labor cost is increasing year by year, and the operation and maintenance pressure is becoming greater and greater. Therefore, how to apply new technology to predict device failure in advance, respond in time, and reduce labor input has become an increasingly urgent problem.
[0003] At present, since it is impossible to know whether the device will fail, the mode adopted by the staff mainly depends on the following two modes: after-maintenance and regular maintenance. After-maintenance refers to maintenance after the device fails, and regular maintenance refers to device inspection and maintenance according to a fixed cycle. This mode has the dual problems of over-maintenance and insufficient maintenance, and about 40% of regular maintenance is a waste of resources. In view of the above situation, the present application provides a device health degree algorithm to solve the above problems. SUMMARY
[0004] In view of the deficiencies of the prior art, the present application provides a device health degree algorithm, which solves the problem that since it is impossible to know whether the device will fail, the mode adopted by the staff mainly depends on the following two modes: after-maintenance and regular maintenance. After-maintenance refers to maintenance after the device fails, and regular maintenance refers to device inspection and maintenance according to a fixed cycle. This mode has the dual problems of over-maintenance and insufficient maintenance, and about 40% of regular maintenance is a waste of resources.
[0005] To achieve the above purpose, the present application provides the following technical scheme: a device health degree algorithm, the algorithm steps comprising: S1. Multi-source data acquisition step: through a data acquisition transmission interface, using multiple industrial standard protocols, real-time acquisition of multi-source heterogeneous data of the device; S2. Data preprocessing and feature extraction step: cleaning, normalizing and fusing the collected multi-source heterogeneous data, and extracting key features for health assessment; S3. Constructing a multi-dimensional health assessment model step: establishing a comprehensive evaluation system including performance indicators, alarm events, device life, maintenance history, spare parts status, economy and safety dimensions, and configuring weight coefficients and maximum deduction values for each dimension; S4. Health score calculation step: according to the deduction score of each dimension and its weight, the comprehensive health score of the equipment is calculated by using a weighted fusion algorithm; S5. Result output step: the comprehensive health score is displayed through a visual interface; S6. Intelligent early warning step: when the comprehensive health score is lower than a preset threshold, the system automatically triggers an early warning and generates a decision support report containing maintenance recommendations.
[0006] Preferably, in the step S1, the multi-source heterogeneous data includes but is not limited to performance parameter data of each component of the equipment, alarm information, life data, maintenance record, spare parts information, maintenance cost and safety log; and the industrial standard protocol includes TCP, SNMP and Redfish protocol.
[0007] Preferably, in the step S1, it further includes a vibration sensor, a temperature sensor and a noise sensor installed on the equipment, which are used to collect physical operation characteristic data of the equipment and input the physical operation characteristic data as a supplement of the performance indicator dimension.
[0008] Preferably, in the step S3, the weight coefficient is dynamically determined by the following way: a basic weight library based on the type of equipment, which is pre-set by expert experience; based on the correlation analysis of historical failure data and each dimension data, the basic weight is optimized by machine learning, and the optimization target is the prediction accuracy rate of health score on failure occurrence; wherein the higher the correlation of the dimension with failure occurrence, the higher the weight coefficient is dynamically adjusted.
[0009] Preferably, in the step S3, the score of the equipment life dimension adopts a technical depreciation method, and the deduction value calculation formula is: ; wherein, is the maximum deduction value of the dimension, t is the current service life, T is the preset total life, n is the acceleration attenuation coefficient, and the value range of n is 1.5 to 3, and the preferred value is 2; and the total life T is dynamically corrected according to the working load intensity of the equipment, and the correction formula is: ; wherein, is the corrected total life, is the reference total life under rated working condition, is the rated working load, is the actual average working load, and k is the load influence factor.
[0010] Preferably, in the step S3, the evaluation of the performance indicator dimension adopts a dynamic baseline method: The threshold in the preset rule is not a fixed value, but a range that is dynamically changed with operating conditions and time, generated through statistical learning based on the historical performance data of the equipment under normal conditions. When real-time performance data deviates from the dynamic baseline range, an early warning is triggered and points are deducted, with the deduction value being proportional to the degree of deviation. The evaluation of the alarm event dimensions includes: The system receives device alarm information uploaded via HTTP and SNMP protocols; it configures different base deduction values for different alarm levels based on their predefined levels; the system accumulates deductions according to preset rules based on the frequency of alarm occurrence within a preset time window, with high-frequency repetitive alarms triggering additional aggravated deductions.
[0011] Preferably, in step S3, the evaluation of the maintenance history dimension includes: Integrate historical maintenance records, including monthly, annual, and temporary maintenance; assess the lingering impact on equipment status based on the type of maintenance, quality evaluation, and time elapsed since the last maintenance; recent, high-quality maintenance will receive bonus points or deductions, while distant, poor-quality maintenance or missed maintenance not performed as planned will be subject to quantitative deductions. The evaluation of the spare parts status dimension includes: Assess the availability of critical spare parts, including checking whether the spare parts have been discontinued, whether there are fully replaceable spare parts, and whether the existing spare parts inventory meets the safety stock requirements. For each negative result, apply a corresponding risk deduction based on the magnitude of the risk of equipment downtime.
[0012] Preferably, in step S3, the evaluation of the economic dimension includes: Calculate the ratio of historical cumulative maintenance costs to the current residual value of the equipment; when the ratio exceeds a preset threshold, it is determined to be uneconomical and points are deducted; wherein, the threshold is set to 40%; The evaluation of the security dimensions includes: Identify whether the equipment has known but unresolved security risks, which are derived from security scan logs or manual reports; assess the potential impact level of the risks on operational safety and business continuity, and deduct points accordingly based on the impact level.
[0013] Preferably, in step S4, the specific calculation process of the weighted fusion algorithm includes: S41. Calculate the deduction value for each evaluation dimension using the following formula: ; in, Let be the deduction value for the i-th dimension. Let be the weight coefficient of the i-th dimension. This represents the cumulative deduction for all inspection items under the i-th dimension. The maximum deduction value for the i-th dimension; S42. Calculate the overall health score using the following formula: ; in, To calculate the overall health score, This is the preset total score.
[0014] Preferably, in step S6, the process of generating the decision support report includes: S61. Perform root cause analysis on the top N dimensions with the most deductions; S62. Match the maintenance strategy, required spare parts list and estimated repair time corresponding to the root cause from the maintenance knowledge base; S63. Integrate the analysis results, maintenance strategies, spare parts list, and estimated repair time to generate the decision support report.
[0015] The technical effects and advantages of this invention are as follows: This equipment health assessment algorithm breaks through the limitations of traditional operation and maintenance that rely on single parameters or human experience. It constructs a comprehensive evaluation system covering seven dimensions: performance, alarms, lifespan, maintenance, spare parts, cost, and safety, achieving a comprehensive and three-dimensional characterization of equipment health status. By introducing dynamic weights, a non-linear lifespan depreciation model, and a dynamic performance baseline based on machine learning, the evaluation model can better fit the objective laws of equipment performance degradation, significantly improving the accuracy and reliability of health status assessment and fault prediction, and providing a solid data foundation for subsequent decision-making. This equipment health algorithm, through a quantified health score and intelligent early warning mechanism, can identify potential risks in the early stages before equipment failure occurs, allowing sufficient preparation time for planned maintenance and effectively reducing unplanned downtime. At the same time, the decision support report generated by the system can transform operation and maintenance decisions from relying on personal experience to relying on data-driven approaches, making the formulation of maintenance strategies, the allocation of spare parts resources, and the planning of maintenance budgets more scientific and reasonable. This avoids the dual dilemma of "over-maintenance" and "under-maintenance," significantly optimizes resource utilization efficiency, and reduces overall operation and maintenance costs. This invention automates the entire process of equipment health assessment, from data collection, analysis, health scoring to early warning report generation. This significantly reduces reliance on the experience of senior maintenance personnel, lightens their workload, and makes maintenance management activities replicable and standardized. This highly intelligent management model not only improves the efficiency of daily maintenance but also enhances the standardization and traceability of the entire equipment asset management level through unified and objective evaluation standards, providing a core tool for enterprises' intelligent transformation and upgrading. Attached Figure Description
[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0017] Figure 1 This is a flowchart of the overall method of the present invention; Figure 2 This is a diagram showing the overall system framework of the present invention; Figure 3 This is a flowchart of the data preprocessing process of the present invention; Figure 4 This is a flowchart of the multidimensional model calculation process of the present invention; Figure 5 This is a flowchart of the health score calculation and early warning process of the present invention; Figure 6 This is a sample diagram of the algorithm model of the present invention. Detailed Implementation
[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0019] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.
[0020] This invention discloses a device health algorithm, based on the attached... Figures 1 to 2 As shown, the algorithm steps include: S1. Multi-source data acquisition steps: Through the data acquisition and transmission interface, and using various industrial standard protocols, multi-source heterogeneous data from the equipment are acquired in real time; S2. Data preprocessing and feature extraction steps: Clean, normalize and fuse the collected multi-source heterogeneous data, and extract key features for health assessment; S3. Steps for building a multidimensional health assessment model: Establish a comprehensive assessment system that includes performance indicators, alarm events, equipment lifespan, maintenance history, spare parts status, economic efficiency and safety dimensions, and configure weight coefficients and maximum deduction values for each dimension; S4. Health Score Calculation Steps: Based on the deductions and weights of each dimension, a weighted fusion algorithm is used to calculate the device's overall health score; S5. Results Output Steps: Display the overall health score through a visual interface; S6. Intelligent Early Warning Steps: When the overall health score falls below a preset threshold, the system automatically triggers an early warning and generates a decision support report containing maintenance recommendations.
[0021] In this embodiment, the core of this step lies in the fusion of "multi-source heterogeneous data," which is the foundation for achieving a comprehensive evaluation; specifically: Performance parameter data includes not only IT metrics such as CPU utilization and memory usage, but also core physical parameters specific to industrial scenarios, such as motor current, voltage, power factor, bearing temperature, equipment operating speed, and output accuracy. This data is acquired in real time through the equipment's PLC, SCADA system, or directly embedded sensors. Alarm information: The system can parse alarms from different sources and formats and standardize them; for example, it can normalize "Trap" alarms from the SNMP protocol, system logs from Syslog, and advanced alarms generated by device-specific management software to form a unified alarm level and event description. Lifetime data and maintenance records: This section emphasizes integration with the equipment management system; lifetime data includes not only the equipment's manufacturing and installation dates, but may also integrate detailed work order history, component replacement records, lubrication records, etc. from the CMMS, thereby building a "full lifecycle digital archive" for the equipment; Economic efficiency and spare parts data: This section innovatively incorporates operation and maintenance business data into the technical evaluation system; the system obtains the inventory status, procurement lead time, whether production is stopped, and historical maintenance cost details of key spare parts through interfaces with ERP or supply chain management systems, providing direct input for quantitative evaluation of economic efficiency.
[0022] According to the appendix Figures 1 to 2 As shown, further, in step S1, the multi-source heterogeneous data includes, but is not limited to, performance parameter data, alarm information, lifespan data, maintenance records, spare parts information, maintenance costs, and safety logs of various components of the equipment; industry standard protocols include TCP / IP, SNMP, and Redfish protocols.
[0023] According to the appendix Figures 1 to 2 As shown, further, step S1 also includes a vibration sensor, a temperature sensor and a noise sensor installed on the equipment, which are used to collect physical operating characteristic data of the equipment and use the physical operating characteristic data as a supplementary input for performance index dimensions.
[0024] According to the appendix Figures 3 to 6 As shown, specifically disclosed, in step S3, the weighting coefficients are dynamically determined in the following manner: The basic weight library based on device type is pre-set by expert experience; Based on the correlation analysis of historical fault data and data of various dimensions, machine learning is used to optimize the basic weights, with the optimization goal being the accuracy of the health score in predicting fault occurrence; among them, the weight coefficient of the dimension with higher correlation to fault occurrence is dynamically increased. In step S3, the equipment lifespan dimension is scored using the technical depreciation method, and the deduction value is calculated using the following formula: ; in, The maximum deduction value for this dimension is t, where t is the current service life, T is the preset total lifespan, and n is the accelerated decay coefficient, with n ranging from 1.5 to 3, and the preferred value being 2. Furthermore, the total lifespan T is dynamically adjusted based on the equipment's workload intensity, using the following formula: ; in, For the corrected total lifetime, The reference total life under rated operating conditions. For rated working load, Where k is the actual average workload, and k is the load influence factor. In step S3, the evaluation of performance metrics uses the dynamic baseline method: The thresholds in the preset rules are not fixed values, but rather ranges that are generated through statistical learning based on the historical performance data of the equipment under normal conditions and that change dynamically with operating conditions and time. When real-time performance data deviates from the dynamic baseline range, an alert is triggered and points are deducted, with the deduction value being proportional to the degree of deviation. The assessment of alarm events includes: Receive device alarm information uploaded via HTTP and SNMP protocols; configure different base deduction values for different alarm levels according to the predefined alarm levels; accumulate deductions according to preset rules based on the frequency of alarm occurrence within a preset time window, where high-frequency repeated alarms will trigger additional aggravated deductions. In step S3, the assessment of the maintenance history dimension includes: Integrate historical maintenance records, including monthly, annual, and temporary maintenance; assess the lingering impact on equipment status based on the type of maintenance, quality evaluation, and time elapsed since the last maintenance; recent, high-quality maintenance will receive bonus points or deductions, while distant, poor-quality maintenance or missed maintenance not performed as planned will be subject to quantitative deductions. The assessment of spare parts condition includes: Assess the availability of critical spare parts, including checking whether the spare parts have been discontinued, whether there are fully replaceable spare parts, and whether the existing spare parts inventory meets the safety stock requirements. For each negative result, apply a corresponding risk deduction based on the magnitude of the risk of equipment downtime. In step S3, the assessment of the economic dimension includes: Calculate the ratio of historical cumulative maintenance costs to the current residual value of the equipment; when the ratio exceeds a preset threshold, it is deemed to have poor economic efficiency and points are deducted; the threshold is set at 40%. The security dimension assessment includes: Identify whether the equipment has known but not eliminated security risks, which may originate from security scan logs or manual reports; assess the potential impact level of the risks on operational safety and business continuity, and deduct points accordingly based on the impact level.
[0025] In this embodiment, the "dynamic weighting" mechanism is a key feature that enables accurate evaluation; its operating logic can be further described as a closed-loop optimization process: Basic weight library: For different types of equipment, domain experts pre-set a set of initial weights based on their failure modes and impact analysis; Machine learning optimization: The system continuously collects historical data, including evaluation results of each dimension and labels indicating whether a failure has occurred. Through correlation analysis and other means, the model automatically identifies the dimension that contributes the most to the failure prediction of a specific device. This optimization is a continuous process to ensure that the model can adapt to dynamic factors such as equipment aging and changes in operating conditions, and achieve the self-evolution capability of becoming more accurate with use.
[0026] According to the appendix Figures 3 to 6 As shown, specifically disclosed in step S4, the specific calculation process of the weighted fusion algorithm includes: S41. Calculate the deduction value for each evaluation dimension using the following formula: ; in, Let be the deduction value for the i-th dimension. Let be the weight coefficient of the i-th dimension. This represents the cumulative deduction for all inspection items under the i-th dimension. The maximum deduction value for the i-th dimension; S42. Calculate the overall health score using the following formula: ; in, To calculate the overall health score, This is the preset total score.
[0027] According to the appendix Figures 1 to 2 As shown, it is particularly important to emphasize that the decision support report generation process in step S6 includes: S61. Perform root cause analysis on the top N dimensions with the most deductions; S62. Match the maintenance strategy, required spare parts list, and estimated repair time corresponding to the root cause from the maintenance knowledge base; S63. Integrate the analysis results, maintenance strategies, spare parts list, and estimated repair time to generate a decision support report.
[0028] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A device health algorithm, characterized in that, The algorithm steps include: S1. Multi-source data acquisition steps: Through the data acquisition and transmission interface, and using various industrial standard protocols, multi-source heterogeneous data from the equipment are acquired in real time; S2. Data preprocessing and feature extraction steps: The collected multi-source heterogeneous data is cleaned, normalized and fused, and key features for health assessment are extracted; S3. Steps for building a multidimensional health assessment model: Establish a comprehensive assessment system that includes performance indicators, alarm events, equipment lifespan, maintenance history, spare parts status, economic efficiency and safety dimensions, and configure weight coefficients and maximum deduction values for each dimension; S4. Health Score Calculation Steps: Based on the deductions and weights of each dimension, a weighted fusion algorithm is used to calculate the device's overall health score; S5. Result Output Steps: Display the comprehensive health score through a visual interface; S6. Intelligent early warning step: When the overall health score is lower than the preset threshold, the system automatically triggers an early warning and generates a decision support report containing maintenance suggestions.
2. The device health algorithm according to claim 1, characterized in that, In step S1, the multi-source heterogeneous data includes, but is not limited to, performance parameter data, alarm information, lifespan data, maintenance records, spare parts information, maintenance costs, and safety logs of various components of the equipment; the industrial standard protocols include TCP, SNMP, and Redfish protocols.
3. The device health algorithm according to claim 2, characterized in that, Step S1 also includes vibration sensors, temperature sensors and noise sensors installed on the equipment, which are used to collect physical operating characteristic data of the equipment and use the physical operating characteristic data as supplementary input to the performance index dimension.
4. The device health algorithm according to claim 1, characterized in that, In step S3, the weighting coefficients are dynamically determined in the following manner: The basic weight library based on device type is pre-set by expert experience; Based on the correlation analysis between historical fault data and data of various dimensions, the basic weights are optimized by machine learning. The optimization objective is to improve the prediction accuracy of the health score for fault occurrence. Among them, the weight coefficient of the dimension with higher correlation to fault occurrence is dynamically increased.
5. The device health algorithm according to claim 4, characterized in that, In step S3, the equipment lifespan dimension is scored using the technical depreciation method, and the deduction value is calculated using the following formula: ； in, The maximum deduction value for this dimension is t, where t is the current service life, T is the preset total lifespan, and n is the accelerated decay coefficient, with n ranging from 1.5 to 3, and the preferred value being 2. Furthermore, the total lifespan T is dynamically adjusted based on the equipment's workload intensity, using the following formula: ; in, For the corrected total lifetime, The reference total life under rated operating conditions. For rated working load, denoted as the actual average workload, and k as the load influence factor.
6. The device health algorithm according to claim 5, characterized in that, In step S3, the evaluation of the performance metric dimensions adopts the dynamic baseline method: The threshold in the preset rule is not a fixed value, but a range that is dynamically changed with operating conditions and time, generated through statistical learning based on the historical performance data of the equipment under normal conditions. When real-time performance data deviates from the dynamic baseline range, an early warning is triggered and points are deducted, with the deduction value being proportional to the degree of deviation. The evaluation of the alarm event dimensions includes: The system receives device alarm information uploaded via HTTP and SNMP protocols; it configures different base deduction values for different alarm levels based on their predefined levels; the system accumulates deductions according to preset rules based on the frequency of alarm occurrence within a preset time window, with high-frequency repetitive alarms triggering additional aggravated deductions.
7. The device health algorithm according to claim 6, characterized in that, In step S3, the evaluation of the maintenance history dimension includes: Integrate historical maintenance records, including monthly, annual, and temporary maintenance; assess the lingering impact on equipment status based on the type of maintenance, quality evaluation, and time elapsed since the last maintenance; recent, high-quality maintenance will receive bonus points or deductions, while distant, poor-quality maintenance or missed maintenance not performed as planned will be subject to quantitative deductions. The evaluation of the spare parts status dimension includes: Assess the availability of critical spare parts, including checking whether the spare parts have been discontinued, whether there are fully replaceable spare parts, and whether the existing spare parts inventory meets the safety stock requirements. For each negative result, apply a corresponding risk deduction based on the magnitude of the risk of equipment downtime.
8. The device health algorithm according to claim 7, characterized in that, In step S3, the evaluation of the economic dimension includes: Calculate the ratio of historical cumulative maintenance costs to the current residual value of the equipment; when the ratio exceeds a preset threshold, it is determined to be uneconomical and points are deducted; wherein, the threshold is set to 40%; The evaluation of the security dimensions includes: Identify whether the equipment has known but unresolved security risks, which are derived from security scan logs or manual reports; assess the potential impact level of the risks on operational safety and business continuity, and deduct points accordingly based on the impact level.
9. The device health algorithm according to claim 1, characterized in that, In step S4, the specific calculation process of the weighted fusion algorithm includes: S41. Calculate the deduction value for each evaluation dimension using the following formula: ； in, Let be the deduction value for the i-th dimension. Let be the weight coefficient of the i-th dimension. This represents the cumulative deduction for all inspection items under the i-th dimension. The maximum deduction value for the i-th dimension; S42. Calculate the overall health score using the following formula: ； in, To calculate the overall health score, This is the preset total score.
10. The device health algorithm according to claim 1, characterized in that, In step S6, the process of generating the decision support report includes: S61. Perform root cause analysis on the top N dimensions with the most deductions; S62. Match the maintenance strategy, required spare parts list and estimated repair time corresponding to the root cause from the maintenance knowledge base; S63. Integrate the analysis results, maintenance strategies, spare parts list, and estimated repair time to generate the decision support report.