Equipment health assessment method under influence of multiple parameters

Abstract

The invention discloses an equipment health assessment method under multi-parameter influence, and relates to the technical field of equipment health management, and the method comprises the steps: obtaining multi-parameter original data of certain equipment, and extracting the feature vector of each parameter original data, the multi-parameter original data comprising current state data, historical operation data, and historical maintenance data; the parameter feature vectors, the configured vector weights and the vector reference conditions serve as input, the overall health value of the equipment in the normal operation process is calculated, and the vector reference conditions comprise multiple sections of threshold values. According to the method, multi-dimensional operation and maintenance data of the equipment are fully considered, the overall health state of the equipment is effectively represented, quantitative guidance is provided for maintenance and management of the equipment in the whole life cycle, and scientificity, comprehensiveness and accuracy of health state assessment results are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment health management, and specifically aims at the performance degradation analysis requirements during the equipment life cycle. In particular, it is an equipment health assessment method under the influence of multiple parameters. Background Art

[0002] The wide application of new generation information technology in various industries has continuously promoted the digital transformation and upgrading of modern equipment manufacturing. With the continuous deepening of multi-dimensional condition monitoring during the equipment's service process, the internal condition information of the equipment has gradually become more transparent, thus providing more accurate guidance for the equipment's operation and maintenance work. Against the background that the equipment intelligent operation and maintenance management system dominated by digital information gradually replaces the traditional automated monitoring system, various parametric data such as the operating state, usage environment, service time, usage frequency, and maintenance history during the equipment's entire life cycle have achieved deep online integration. These data provide sufficient big data support for the performance prediction and health management of the equipment.

[0003] However, the improvement of the equipment's digital level has also brought new challenges to data-driven operation and maintenance decision-making. Existing health assessment methods usually only rely on real-time condition monitoring data, fail to fully consider the long-term evolution process of the equipment during its entire life cycle, and also lack the comprehensive ability to fully characterize the overall performance of the equipment. This makes it difficult for users to form an efficient and unified inspection guidance line during the daily operation and maintenance process. In addition, variable factors such as the service environment, usage conditions, and human operation habits often cause the equipment to not operate in the best health state.

[0004] Therefore, designing a health assessment method that can fully consider the multi-dimensional operation and maintenance data of the equipment, effectively characterize the overall health state of the equipment, and provide quantitative guidance for the maintenance management during the equipment's entire life cycle will become the key to the future development of equipment prediction and health management technology. This method not only helps to improve the operation and maintenance efficiency of the equipment, but also can effectively extend its service life and optimize the overall performance. Summary of the Invention

[0005] In view of the above problems and technical requirements, the inventor of the present invention has proposed an equipment health assessment method under the influence of multiple parameters. By comprehensively considering multi-dimensional parameters inside and outside the equipment, covering various factors affecting the equipment's health state, and effectively incorporating the historical characteristics of the equipment into the health assessment, the scientificity, comprehensiveness, and accuracy of the health state assessment results are ensured. The technical solution of the present invention is as follows:

[0006] An equipment health assessment method under the influence of multiple parameters, comprising the following steps:

[0007] Obtain the multi-parameter original data of a certain equipment and extract the feature vectors of each parameter's original data;

[0008] Taking each parameter feature vector, the configured vector weights, and the vector reference conditions as inputs, calculate the overall health value during the normal operation of the equipment. The vector reference conditions include multiple segments of thresholds.

[0009] Its further technical solution is that the multi-parameter raw data is classified and input, including:

[0010] Such as Figure 1 shown, in order to fully consider the influence of the equipment's own factors and external human factors on the equipment's health status, the present invention classifies the input of multi-parameter raw data into three major categories: current status data, historical operation data, and historical maintenance data. Among them:

[0011] 1) Current status data: This type of data is obtained in real time through the equipment's sensor monitoring system, including parameters such as vibration, temperature, pressure, rotational speed, and power. Common monitoring means include vibration monitoring, thermal engineering monitoring, oil sample analysis, non-destructive flaw detection monitoring, power monitoring, etc. The current status data can directly reflect the operating performance of the equipment at a specific moment and is an important basis for evaluating the equipment's health status.

[0012] 2) Historical operation data: mainly includes the start-stop records and operating time of the equipment. These data are closely related to the fatigue accumulation and service life of the equipment during its life cycle, can reveal the working load and operating conditions of the equipment under different operating conditions, and provide key support for evaluating the health level of the equipment after long-term use.

[0013] 3) Historical maintenance data: mainly includes the alarm records, repair records, maintenance records, and replacement records of the equipment, etc. Such data reflects the equipment's fault history, maintenance frequency, and repair situation, and is closely related to the reliability, maintainability, and operation standardization of the equipment during its life cycle, providing important historical background information for health assessment.

[0014] Its further technical solution is that the parameter feature vector generation includes:

[0015] During the equipment health assessment process, the original multi-parameter data usually has challenges such as large quantity, low value density, and difficulty in directly characterizing the equipment's operating characteristics. Therefore, before conducting a health assessment, it is necessary to effectively extract the features of various types of parameter raw data and generate feature vectors. The specific steps are as Figure 2 shown:

[0016] 1) Feature extraction of current state data: For the multi-parameter raw data obtained by means such as thermal monitoring, vibration monitoring, oil monitoring, and non-destructive testing monitoring, the characteristic values that are of great significance to the operating state of the equipment need to be extracted. For example: Extract the vibration RMS value (root mean square value) and the first-order main frequency spectrum value from the vibration raw data. The vibration RMS value reflects the vibration amplitude and the smoothness of the equipment operation, and the first-order main frequency spectrum value is used to analyze the mechanical performance and fault diagnosis of the equipment; Extract the temperature data of the key parts of concern from the temperature raw data to detect whether the equipment is within the normal working temperature range and avoid failures caused by overheating; Extract the peak pressure from the pressure raw data to reveal the possible overload or pressure fluctuation conditions during the equipment operation; Extract the rotational speed volatility from the rotational speed raw data to reflect the stability and operating characteristics of the equipment rotational speed; Extract the relevant data characterizing the energy efficiency and operating load of the equipment from the power raw data.

[0017] 2) Feature extraction of historical operation data: Extract the start-stop times or frequencies from the start-stop records of the equipment to reflect the frequency of equipment operation and the impact of the use environment on the equipment health; Extract the operating hours from the operating time to evaluate the cumulative working time of the equipment, thereby inferring the fatigue degree and service life.

[0018] 3) Feature extraction of historical maintenance data: Extract the annual average alarm rate (AFR) from the alarm records, which represents the frequency of faults or abnormalities of the equipment within a specific time period; Extract the mean time to repair (MTTR) from the maintenance records, that is, divide the total maintenance duration by the number of maintenance times to reflect the repair efficiency after the equipment failure occurs; Extract the maintenance time difference from the maintenance records, which is the difference between the planned maintenance cycle time and the actual maintenance time, revealing the timeliness of the equipment regular maintenance and affecting its long-term health; Extract the latest equipment replacement time from the replacement records as the starting time for extracting historical operation and maintenance data, providing a time benchmark for subsequent health assessment.

[0019] Its further technical solution is the weight configuration of the parameter feature vector, including:

[0020] As Figure 3 shown, each feature in the parameter feature vector has different importance for the description of the equipment health state. Therefore, when performing the health assessment calculation, corresponding weights must be assigned to each parameter feature according to the actual situation of the equipment. This weight configuration is determined according to the influence degree of each feature on the overall health state of the equipment to ensure that the assessment result can accurately reflect the actual operating condition of the equipment. Specifically, the weight assignment needs to follow the following principles:

[0021] 1) Feature importance assessment: Based on the contribution of different parameter feature vectors to the health status of the equipment, evaluate the importance of each feature vector in the overall health assessment and assign corresponding weights. For example, some key parameters (such as temperature, vibration RMS, etc.) may have a greater impact on the health status, while other auxiliary features (such as operating hours, maintenance time difference, etc.) are relatively less important.

[0022] 2) Dynamic adjustment: The weight configuration should be flexible. According to the working environment, usage conditions, and historical performance of the equipment, dynamically adjust the weights of each parameter feature vector. This adjustment can be achieved through historical data analysis, expert experience, or machine learning methods.

[0023] 3) Weight sum: The sum of the weights of all parameter feature vectors is 100%, ensuring the consistency and comparability of the evaluation model, and avoiding bias in the evaluation results caused by too high or too low weight of a certain feature.

[0024] Its further technical solution is the configuration of the reference conditions for the parameter feature vector, including:

[0025] Such as Figure 4 As shown, each feature in the parameter feature vector needs to be configured with corresponding reference conditions, and the reference conditions at least include one level of alarm value and reference value. These reference conditions are used as the basis for health assessment to judge the normal range of each feature value and whether there are abnormalities. Moreover, the more levels there are, the closer the corresponding relationship between the evaluated health value and the actual performance of the equipment will be. The specific configuration requirements are as follows:

[0026] 1) The reference conditions for the current state data feature vector include the warning value, level II alarm value, level I alarm value, and reference value. Among them: The warning value indicates that when a certain feature value reaches this threshold, attention should be paid and preventive measures should be taken; when the feature value of this feature vector reaches the level II alarm value, it means that the equipment state has a relatively serious abnormality and emergency measures should be taken immediately; when the feature value of this feature vector exceeds the level I alarm value, it means that the equipment state is already very critical and fault troubleshooting and repair measures must be taken quickly; the reference value of this feature vector is the standard value within the normal range of equipment operation, usually determined based on the design parameters or historical data of the equipment, and used as the reference for the equipment health status.

[0027] 2) The reference conditions for the historical operation data feature vector: The reference value of the start-stop times or frequency is usually determined according to the calibrated life time or the predetermined usage cycle of the equipment, which can reflect the normal operation range of the equipment within its expected service life; the reference value of the operating hours is determined based on the calibrated life time of the equipment, combined with the usage intensity and working conditions of the equipment, to evaluate the working load and fatigue degree of the equipment.

[0028] 3) Benchmark conditions for the historical maintenance data feature vector: The benchmark value of AFR is usually determined based on the equipment's maintenance history, design standards, and expected reliability, reflecting the expected frequency of alarms under normal operation; the benchmark value of MTTR is usually determined based on the equipment's maintenance manual and fault repair history; the benchmark value of the maintenance time difference is usually determined based on the equipment's maintenance plan or the maintenance cycle recommended by the manufacturer, and is used to judge whether the equipment is maintained at the specified maintenance interval.

[0029] Its further technical solution is that the real-time health assessment model calculation includes:

[0030] As Figure 5 shown, based on the input of multi-parameter feature vectors, vector weights, and benchmark conditions, the real-time health assessment of the equipment is carried out. When conducting the health assessment, it is assumed that the equipment is in a steady-state operation state and the sensors are not disconnected or faulty. The health assessment process is divided into the following two main steps:

[0031] 1) Single-parameter feature health assessment: First, according to each single-parameter feature vector and the corresponding benchmark conditions, the first calculation model is used to calculate the current health value of the single-parameter feature vector. Among them, the first calculation model gives the corresponding health value according to the current single-parameter feature vector located in a certain stage of multiple thresholds. The calculation method of the first calculation model can be selected according to actual application requirements, such as a linear model, a quadratic model, etc. The specific model form depends on the physical meaning of the feature and the reflection of the equipment's health status. Repeat this step to generate the health values of all parameter feature vectors, providing a basis for subsequent comprehensive evaluation.

[0032] 2) Composite-parameter feature health assessment: On the basis of the single-parameter health assessment, further according to the health values of each parameter feature vector and the corresponding weights, the second calculation model is used to calculate the overall health value during the normal operation of the equipment. Among them, the calculation method of the second calculation model can be selected in various forms, such as a linear weighted sum calculation model or a radar model, etc. The specific method can be determined according to the characteristics of different equipment and actual requirements. This calculation process will consider the comprehensive impact of all features on the equipment's health, and conduct comprehensive calculations according to the configured weights, and finally obtain the evaluation result of the overall health status of the equipment.

[0033] When the equipment is in a shutdown state, the real-time health assessment model calculation ends, and the assessment result can be used for decision support, providing a scientific basis for subsequent maintenance decisions. Through this real-time assessment model, the health status of the equipment can be dynamically monitored, potential faults or performance degradation problems can be detected in a timely manner, so as to optimize the operation and maintenance management of the equipment and improve its reliability and availability.

[0034] The beneficial technical effects of the present invention are:

[0035] The equipment health assessment method under the influence of multiple parameters proposed by the present invention can effectively consider the historical maintenance records and operation data of the equipment on the basis of considering the real-time status monitoring data of the equipment. This innovation makes the evaluated health index have higher accuracy in characterizing the health status of the equipment, and can provide more accurate and efficient quantitative guidance for equipment operation and maintenance personnel, thus significantly improving the service quality and operation and maintenance efficiency of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of multi-parameter data classification provided by this application.

[0037] Figure 2 It is a schematic diagram of parameter feature vector generation provided by this application.

[0038] Figure 3 It is a schematic diagram of weight configuration of parameter feature vectors provided by this application.

[0039] Figure 4 It is a schematic diagram of benchmark condition configuration of parameter feature vectors provided by this application.

[0040] Figure 5 It is a flowchart of real-time health assessment calculation of equipment provided by this application.

[0041] Figure 6 It is a schematic diagram of the first calculation model corresponding to each parameter feature vector provided by this application.

[0042] Figure 7 It is a schematic diagram of the radar model corresponding to the composite parameter feature vector provided by this application.

[0043] Figure 8 It is a historical trend chart of health values calculated within the historical period of the ship provided by this application.

[0044] Figure 9 It is a radar situation map of equipment health at the latest moment provided by this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0045] The following further describes the specific embodiments of the present invention with reference to the accompanying drawings.

[0046] An embodiment of the present application provides a method for evaluating the health of equipment under the influence of multiple parameters. Taking the health management of a certain ship power equipment as an example, the current state data of the equipment is the digital state monitoring signal parameters of the key component cylinder of the engine, mainly including the original data of the exhaust temperature, pressure, vibration, and cylinder jacket water temperature of the cylinder. The corresponding characteristic values are the exhaust temperature, explosion pressure, vibration RMS, and cylinder jacket water temperature. Combining historical operation data and historical maintenance data, the characteristic vector of the cylinder includes: {exhaust temperature, explosion pressure, vibration RMS, cylinder jacket water temperature, start-stop frequency, operating hours, AFR, maintenance time difference}.

[0047] Then, according to the influence degree of the 8 elements in the characteristic vector on the cylinder performance, the weight vector is given as follows: {20, 20, 10, 10, 10, 5, 10, 15}.

[0048] According to the monitoring alarm conditions of the cylinder and the equipment maintenance manual, the reference conditions of each parameter characteristic vector are shown in Table 1 below:

[0049] Table 1. Reference conditions of each parameter characteristic vector

[0050]

[0051] For the first calculation model used for the single-parameter characteristic health assessment in the characteristic vector, in this embodiment, linear functions and quadratic functions are selected, and the corresponding functions are used to fit the corresponding relationship between each characteristic vector and the health value in different stages divided according to the reference conditions, as Figure 6 shown.

[0052] For the second calculation model used for the composite-parameter characteristic health assessment in the characteristic vector, in this embodiment, a radar model is selected for comprehensive calculation. Then, the radar charts of the key component cylinder of the engine at different time periods are as Figure 7 shown. It should be noted that Figure 6 and Figure 7 the green, yellow, and red colors in represent healthy, sub-healthy, and unhealthy respectively. Exemplarily, an overall health assessment calculation formula of a radar model is given in this embodiment as:

[0053] H = S 0 / S

[0054] h i = hh i * Q i / 100, s.t.: ΣQ i = 100

[0055]

[0056] Among them, H represents the current overall health value of the equipment; S 0represents the polygonal area of ​​the current composite parameter feature vector on the radar chart, S represents the base area of ​​the radar chart; n represents the dimension of the feature vector considered in the equipment health assessment, for example, n=8 in this embodiment; hh i 、h i represents the current health value and normalized weighted health value of the i-th parameter feature vector, Q i represents the weight of the i-th parameter eigenvector.

[0057] The multi-parameter feature vector data for a period of time in the history of the ship are selected and shown in Table 2 below. It is concluded from the replacement records that the equipment replacement time during this period is April 2023.

[0058] Table 2. Historical data of multi-parameter feature vectors of ships

[0059]

[0060] Based on the historical data of the multi-parameter feature vector of the ship, the historical trend of the health value of the ship engine cylinder during the period is calculated as follows: Figure 8 As shown, the health radar situation map at the latest moment (i.e. 202412) is as follows Figure 9 As shown, the overall health value of the equipment is 75 points, which is in a sub-healthy state.

[0061] The above is only a preferred embodiment of the present application, and the present invention is not limited to the above embodiments. It is understood that other improvements and changes directly derived or associated by those skilled in the art without departing from the spirit and concept of the present invention should be considered to be included in the protection scope of the present invention.

Claims

1. A method for equipment health assessment under the influence of multiple parameters, characterized in that: The method comprises: Obtain the multi-parameter raw data of a certain equipment and extract the feature vector of each parameter raw data; Taking the characteristic vectors of each parameter, the configured vector weights and the vector reference conditions as input, the overall health value of the equipment during normal operation is calculated; Wherein, the multi-parameter original data includes current status data, historical operation data, and historical maintenance data; The vector reference condition includes multiple thresholds.

2. The equipment health assessment method under the influence of multiple parameters according to claim 1 is characterized in that: The overall health value of the equipment during normal operation is calculated by taking the characteristic vectors of each parameter, the configured vector weights and the vector reference conditions as inputs, including: According to each single parameter feature vector and the corresponding benchmark condition, the current health value of the single parameter feature vector is calculated using the first calculation model, and the step is repeated to generate the health values ​​of all parameter feature vectors; According to the health value of each parameter feature vector and the corresponding weight, the second calculation model is used to calculate the overall health value of the equipment during normal operation.

3. The equipment health assessment method under the influence of multiple parameters according to claim 2 is characterized in that: The first calculation model gives a corresponding health value according to a stage where the current single parameter feature vector is located in the multi-segment threshold value; The second calculation model adopts a linear weighted sum calculation model or a radar model to perform comprehensive calculation according to the configured weights.

4. The equipment health assessment method under the influence of multiple parameters according to claim 1 is characterized in that: The obtaining of multi-parameter raw data of a certain equipment includes: The current state data includes vibration, temperature, pressure, speed, and power parameters obtained through real-time monitoring; The historical operation data includes the start and stop records and operation time of the equipment; The historical maintenance data includes equipment alarm records, repair records, maintenance records and replacement records.

5. The equipment health assessment method under the influence of multiple parameters according to claim 4 is characterized in that: The step of extracting the characteristic vector of the original data of each parameter includes: For the current state data, extract the vibration RMS value and the first-order main spectrum value from the vibration raw data, extract the temperature data of the key focus parts from the temperature raw data, extract the peak pressure from the pressure raw data, extract the speed fluctuation rate from the speed raw data, and extract the relevant data characterizing the energy efficiency and operating load of the equipment from the power raw data; For the historical operation data, extract the start and stop times or frequencies from the start and stop records, and extract the operating hours from the operating time; For the historical maintenance data, the annual average alarm rate is extracted from the alarm record, the average fault repair time is extracted from the maintenance record, the maintenance time difference is extracted from the maintenance record, and the maintenance time difference is the difference between the planned maintenance cycle time and the actual maintenance time. The latest equipment replacement time is extracted from the replacement record as the starting time for extracting historical operation and maintenance data.

6. The equipment health assessment method under the influence of multiple parameters according to claim 5 is characterized in that: When configuring the reference conditions of each parameter feature vector, the reference conditions at least include an alarm value and a reference value, where: The reference value of the current state data feature vector is determined based on the design parameters or historical data of the equipment; The reference value of the start-stop number or frequency is determined according to the calibrated life time or the predetermined use cycle of the equipment; The reference value of the operating hours is determined based on the rated life of the equipment, combined with the intensity of use and working conditions of the equipment; The benchmark value of the annual average alarm rate is determined based on the maintenance history, design standards and expected reliability of the equipment; The reference value of the mean fault repair time is determined based on the maintenance manual and fault repair history of the equipment; The baseline value of the maintenance time difference is determined based on the equipment's maintenance plan or the maintenance cycle recommended by the manufacturer.

7. The equipment health assessment method under the influence of multiple parameters according to claim 1 is characterized in that: Configure the weights of each parameter feature vector, including: According to the contribution of different parameter feature vectors to the health status of the equipment, the importance of each feature vector in the overall health assessment is evaluated and a corresponding weight is assigned, and the total weight of all parameter feature vectors is 100%.

8. The equipment health assessment method under the influence of multiple parameters according to claim 7 is characterized in that: Configure the weights of each parameter feature vector, including: The weight of each parameter feature vector is dynamically adjusted according to the equipment's working environment, usage and historical performance.

Images