Rcm-based unit equipment anomaly detection and maintenance system and method

Abstract

The application discloses a unit equipment abnormality detection and maintenance system and method based on RCM, and relates to the technical field of detection and maintenance.The system comprises a forward risk judgment module, a reverse credibility verification module, an equipment state classification module and a hierarchical operation and maintenance disposal module.The method comprises the following steps: collecting real-time operation parameters, performing threshold comparison and abnormality identification, matching an RCM failure mode library to obtain a forward risk judgment result; determining an RCM risk benchmark and a health state base according to historical data, and performing reverse credibility verification on the forward judgment result; distinguishing pseudo-abnormality, early degradation risk and real risk according to parameter abnormality, verification results and risk trends; keeping normal operation for pseudo-abnormality, periodically monitoring early degradation risk, and performing hierarchical maintenance according to grades for real risk.The application effectively reduces the false positive rate, realizes early warning of the implicit degradation trend, and improves the precision and reliability of unit equipment operation and maintenance.

Description

Technical Field

[0001] This invention relates to the field of detection and maintenance technology, specifically to a system and method for detecting and maintaining abnormal equipment in generating units based on RCM. Background Technology

[0002] Reliability Management (RCM) equipment refers to industrial equipment whose operation and maintenance are managed around reliability as a core principle. It mainly includes power units, transmission units, and fluid transport units. This type of equipment consists of multiple components working collaboratively and requires long-term continuous operation. The equipment's operating status directly affects production line stability, production safety, and overall operational efficiency. For anomaly detection and maintenance of this type of equipment, the RCM analysis approach is mostly adopted. This involves collecting equipment operating parameters, identifying parameter deviations, matching them with corresponding failure modes, thereby assessing equipment risk and developing appropriate maintenance strategies.

[0003] Current conventional RCM analysis methods mostly adopt a one-way reasoning model, relying solely on parameter anomalies to directly determine equipment risk, lacking verification and review stages in the entire risk assessment process. In actual field operation, factors such as operating condition switching, instantaneous load changes, and slight sensor interference can all cause temporary fluctuations in equipment parameters. These fluctuations are normal operating condition changes and do not indicate a real equipment failure. However, traditional one-way analysis logic cannot verify the authenticity of anomalies, directly judging normal fluctuations as equipment risks and triggering alarms, interfering with maintenance judgments, and even causing unnecessary downtime for maintenance. Furthermore, this model can only respond to explicit anomalies and cannot combine historical risk baselines and health status for comprehensive judgment, making it difficult to identify hidden degradation trends where risk levels are continuously rising without obvious parameter manifestations. Summary of the Invention

[0004] The purpose of this invention is to provide a system and method for detecting and maintaining abnormal equipment in generating units based on RCM, so as to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: The RCM-based method for detecting and maintaining unit equipment anomalies includes the following steps: S1. Collect real-time operating parameters of the unit equipment, compare the real-time operating parameters with thresholds, and identify abnormal parameter characteristics; match the abnormal parameter characteristics with the preset RCM failure mode library to obtain the corresponding failure mode, and preliminarily determine the equipment risk level based on the failure mode to form a positive risk judgment result; S2. Based on the historical RCM analysis results, historical fault and maintenance records, and long-term operating status data of the unit equipment, determine the current RCM risk benchmark and health status base of the equipment; based on the RCM risk benchmark and health status base, perform reverse credibility verification on the positive risk judgment results to determine whether the positive risk judgment results meet the conditions for risk establishment. S3. When there are abnormal parameter characteristics and the positive risk assessment result does not meet the conditions for risk establishment, the current abnormality is determined to be a pseudo-abnormality; when no abnormal parameter characteristics are identified but the RCM risk benchmark shows a continuous upward trend and the increase exceeds the preset trend threshold, the equipment is determined to have early deterioration risk; when there are abnormal parameter characteristics and the positive risk assessment result meets the conditions for establishment, it is determined to be a real risk. S4. Maintain normal operation for pseudo-anomalies, perform periodic monitoring and status tracking for early deterioration risks, output early warning information for real risks according to their risk level and execute corresponding maintenance plans, and complete the anomaly detection and maintenance of unit equipment.

[0006] Furthermore, S1 includes the following: Real-time operating parameters of each monitoring point of the unit equipment are collected according to the preset collection frequency. The deviation of each operating parameter is calculated according to the preset normal threshold range corresponding to each operating parameter. The deviation is compared with the preset deviation judgment threshold. When the deviation is greater than the judgment threshold, the corresponding operating parameter is judged to be abnormal. All abnormal parameter types and deviation values ​​are summarized to form a parameter abnormality feature set. Retrieve the preset RCM failure mode library, which stores the feature matching conditions and risk weights corresponding to each failure mode. Match the abnormal feature set of parameters with the feature matching conditions of each failure mode one by one. Calculate the feature matching degree of each failure mode based on the deviation of the successfully matched abnormal features. Select failure modes with feature matching degrees greater than the preset matching degree judgment threshold as candidate failure modes. For each candidate failure mode, a single-mode risk score is calculated based on the feature matching degree and the corresponding risk weight. The maximum single-mode risk score among all candidate failure modes is selected as the overall equipment risk score. If there is no candidate failure mode, the overall equipment risk score is assigned to zero. Based on the preset risk threshold, the overall risk score of the equipment is divided into the corresponding risk level range to determine the risk level of the equipment. The candidate failure modes, the overall risk score of the equipment and the risk level of the equipment are integrated to generate a positive risk judgment result.

[0007] Furthermore, S2 includes the following: The system collects historical RCM risk data, historical fault record data, historical maintenance log data, and long-term steady-state operating parameter data for the unit equipment within a preset historical time period. The historical RCM risk data is obtained by calculating the comprehensive risk score and corresponding failure mode and risk level for each historical statistical period based on the historical operating data on a periodic basis. The historical fault record data includes at least the fault occurrence time, fault failure mode, faulty component, and fault handling measures. The historical maintenance log data includes at least the maintenance time, maintenance type, maintenance content, and replacement component information. The long-term steady-state operating parameter data are historical parameter sampling values ​​of the equipment under fault-free and stable operating conditions. The collected data is used to construct the equipment historical status dataset. The historical state dataset is updated at fixed time intervals using a sliding window method, and the sliding window contains multiple consecutive historical statistical periods. Based on the historical status dataset of the equipment, the arithmetic mean of all comprehensive risk scores within the historical statistical period is calculated and used as the RCM risk benchmark value. Based on the historical status dataset of the equipment, the normalized deviation mean of all operating parameters in each historical statistical period is extracted, and the health status baseline value is calculated based on the normalized deviation mean of all historical statistical periods. Based on historical failure boundary statistics, a risk fluctuation tolerance coefficient is set, and the risk benchmark value is combined with the health status baseline value to calculate the risk threshold. Extract the comprehensive risk score of the equipment from the positive risk assessment result, compare the comprehensive risk score of the equipment with the risk establishment threshold, and determine that the positive risk assessment result meets the risk establishment condition when the comprehensive risk score of the equipment is greater than the risk establishment threshold; otherwise, it is determined that the risk establishment condition is not met. Collect two sets of adjacent RCM risk benchmark values ​​before and after the sliding window update, and divide the difference between the latter benchmark value and the former benchmark value by a fixed time interval to obtain the slope of the risk benchmark change. The results of risk establishment criteria determination, risk benchmark change slope, and health status baseline value are summarized to form a reverse credibility verification result.

[0008] Furthermore, S3 includes the following: Extract the set of abnormal parameter features and the comprehensive risk score of the equipment from the positive risk assessment results, and extract the risk establishment condition assessment results and the risk benchmark change slope from the reverse credibility verification results; A preset risk trend threshold is defined based on statistical calibration of risk benchmark fluctuation data of long-term steady-state operation of unit equipment, and is used to quantitatively determine the effective deterioration range of continuous increase in equipment risk. When the set of abnormal parameter features is empty and the slope of the risk benchmark change is less than or equal to the risk trend threshold, the equipment is determined to be in normal operating condition. When the set of abnormal parameter features is not empty and the positive risk assessment result does not meet the conditions for risk establishment, the current parameter abnormality is determined to be a pseudo-anomaly; if the slope of the risk benchmark change is greater than the risk trend threshold in this case, an early deterioration risk warning will be output in addition to the pseudo-anomaly determination. When the set of abnormal parameters is empty and the slope of the risk benchmark change is greater than the risk trend threshold, it is determined that the equipment has an early deterioration risk. When the set of abnormal parameter features is not empty and the positive risk assessment result meets the conditions for risk establishment, the equipment is determined to have a real risk. The integrated judgment results are used to output the final abnormality judgment type of the device.

[0009] Furthermore, S4 includes the following: When the equipment is determined to be in normal operating condition, the current operating condition of the unit equipment remains unchanged, no abnormal warnings are triggered, no additional monitoring tasks are started, and no maintenance operations are performed; When the equipment is determined to be a false anomaly and there is no early deterioration risk warning, the unit equipment should be kept in normal continuous operation, the anomaly alarm should be masked, and no shutdown or maintenance intervention should be performed. When the equipment is identified as a pseudo-anomaly and an early deterioration risk warning is simultaneously superimposed, the current operating condition of the unit equipment is maintained, the current anomaly alarm is masked, and encrypted periodic status tracking and monitoring is started at the same time. The risk baseline update cycle and parameter sampling interval are shortened, and the trend of equipment risk baseline change is continuously recorded. When equipment is identified as having an early risk of deterioration, emergency alarms are not triggered, shutdown maintenance is not performed, a special deterioration tracking ledger is established, and the deviation of equipment parameters from the risk benchmark is continuously tracked according to the preset monitoring cycle. The health status and risk level of the equipment are dynamically updated, and the handling strategy is switched when the risk indicators reach the conditions for real risk judgment. When a device is identified as a real risk, the device risk level and candidate failure modes generated from the positive risk assessment results are retrieved, and a graded early warning and corresponding maintenance plan are executed according to the risk level.

[0010] The RCM-based unit equipment anomaly detection and maintenance system includes: a forward risk assessment module, a reverse reliability verification module, an equipment status classification module, and a hierarchical operation and maintenance handling module. The positive risk assessment module collects real-time operating parameters of the unit equipment, performs threshold comparison on the real-time operating parameters and identifies abnormal parameter characteristics, matches the abnormal parameter characteristics with the preset RCM failure mode library, determines the equipment risk level based on the matched failure modes, and generates a positive risk assessment result. The reverse reliability verification module determines the RCM risk benchmark and health status base based on the historical RCM analysis results, historical fault and maintenance records and long-term operating status data of the unit equipment. Based on the RCM risk benchmark and health status base, it performs reverse reliability verification on the positive risk judgment results to determine whether the positive risk judgment results meet the conditions for risk establishment. The equipment status classification module distinguishes between pseudo-anomalies, early degradation risks, and real risks based on the existence of abnormal parameter characteristics, the validity of positive risk judgment results, and the changing trend of RCM risk benchmarks, and outputs the final anomaly judgment type of the equipment. The hierarchical operation and maintenance module executes corresponding actions based on the final anomaly determination type of the equipment. It maintains normal equipment operation for false anomalies, performs periodic monitoring and status tracking for early deterioration risks, and outputs early warning information and executes corresponding maintenance plans for real risks based on risk level.

[0011] Furthermore, the positive risk assessment module includes a parameter acquisition anomaly identification unit and a failure mode matching unit; The parameter acquisition anomaly identification unit collects real-time operating parameters of each monitoring point of the unit equipment according to the preset acquisition frequency, calculates the deviation degree according to the preset normal threshold range corresponding to the operating parameters, compares the deviation degree with the preset deviation degree judgment threshold, and summarizes the types of abnormal parameters and deviation degree values ​​to form a parameter anomaly feature set. The failure mode matching unit retrieves the preset RCM failure mode library, matches the set of abnormal parameter features with the feature matching conditions of each failure mode one by one, calculates the feature matching degree based on the deviation of the successfully matched abnormal features, and selects failure modes with feature matching degrees greater than the preset matching degree judgment threshold as candidate failure modes.

[0012] Furthermore, the positive risk assessment module also includes a risk score calculation unit and a risk level determination unit; The risk score calculation unit calculates the single-mode risk score based on the feature matching degree of the candidate failure modes and the corresponding risk weights, and selects the maximum value of the single-mode risk score as the comprehensive risk score of the equipment. When there are no candidate failure modes, the comprehensive risk score of the equipment is assigned to zero. The risk level determination unit divides the overall risk score of the equipment into the corresponding risk level range according to the preset risk boundary threshold, integrates the candidate failure modes, the overall risk score of the equipment and the risk level of the equipment, and generates a positive risk judgment result.

[0013] Furthermore, the reverse reliability verification module includes a historical data processing unit and a benchmark threshold calculation unit; The historical data processing unit collects historical RCM risk data, historical fault record data, historical maintenance log data, and long-term steady-state operation parameter data of the unit equipment within a preset historical period, constructs a historical status dataset of the equipment, and updates the historical status dataset using a sliding window method at fixed time intervals. The benchmark threshold calculation unit calculates the arithmetic mean of the comprehensive risk scores for the historical statistical period to obtain the RCM risk benchmark value. It calculates the health status baseline value based on the normalized deviation mean of the historical statistical period. The risk threshold is calculated by combining the risk fluctuation tolerance coefficient, the RCM risk benchmark value and the health status baseline value.

[0014] Furthermore, the reverse reliability verification module also includes a slope calculation unit and a verification result generation unit; The slope calculation unit collects two sets of adjacent RCM risk benchmark values ​​before and after the sliding window updates, and divides the difference between the latter benchmark value and the former benchmark value by a fixed time interval to obtain the slope of the risk benchmark change. The verification result generation unit integrates the risk establishment condition judgment results, the risk benchmark change slope, and the health status baseline value to form a reverse credibility verification result.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention adopts a two-way decision-making logic that combines positive risk assessment with reverse reliability verification, changing the traditional RCM's reliance on parameter anomalies to determine risk. By verifying abnormal information, it can effectively distinguish between normal operating condition changes such as operating condition switching, load fluctuations, and sensor interference, and actual equipment failures, significantly reducing the number of false alarms, avoiding invalid alarms from affecting maintenance judgments, and reducing unnecessary downtime for maintenance. This invention determines the RCM risk benchmark and health status baseline based on historical equipment operating data, and uses dynamic thresholds instead of fixed thresholds, making the risk assessment results closer to the actual operating conditions of the equipment and enhancing its applicability. Simultaneously, this invention utilizes the trend of risk benchmark changes to identify hidden degradation characteristics of the equipment, enabling the detection of early potential faults before obvious parameter anomalies appear, solving the problem that traditional methods struggle to monitor progressive hidden dangers, and improving the anticipation and completeness of anomaly detection. This invention implements graded and classified handling measures for different operating states, which shields pseudo-anomalies, continuously tracks early deterioration, and performs corresponding level early warning and maintenance for real risks. The operation and maintenance handling is more precise and reasonable. While ensuring the continuous and stable operation of equipment, it improves the efficiency of hidden danger handling and the reliability of equipment operation, and is more suitable for the complex and ever-changing actual operating environment of industrial sites. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1This is a flowchart of the unit equipment anomaly detection and maintenance system based on RCM of the present invention. Detailed Implementation

[0017] 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.

[0018] Please see Figure 1 The present invention provides the following technical solution: The RCM-based unit equipment anomaly detection and maintenance system includes: a forward risk assessment module, a reverse reliability verification module, an equipment status classification module, and a hierarchical operation and maintenance handling module. The positive risk assessment module collects real-time operating parameters of the unit equipment, performs threshold comparison on the real-time operating parameters and identifies abnormal parameter characteristics, matches the abnormal parameter characteristics with the preset RCM failure mode library, determines the equipment risk level based on the matched failure modes, and generates a positive risk assessment result. The reverse reliability verification module determines the RCM risk benchmark and health status base based on the historical RCM analysis results, historical fault and maintenance records and long-term operating status data of the unit equipment. Based on the RCM risk benchmark and health status base, it performs reverse reliability verification on the positive risk judgment results to determine whether the positive risk judgment results meet the conditions for risk establishment. The equipment status classification module distinguishes between pseudo-anomalies, early degradation risks, and real risks based on the existence of abnormal parameter characteristics, the validity of positive risk judgment results, and the changing trend of RCM risk benchmarks, and outputs the final anomaly judgment type of the equipment. The hierarchical operation and maintenance module executes corresponding actions based on the final anomaly determination type of the equipment. It maintains normal equipment operation for false anomalies, performs periodic monitoring and status tracking for early deterioration risks, and outputs early warning information and executes corresponding maintenance plans for real risks based on risk level.

[0019] The positive risk assessment module includes a parameter acquisition anomaly identification unit and a failure mode matching unit; The parameter acquisition anomaly identification unit collects real-time operating parameters of each monitoring point of the unit equipment according to the preset acquisition frequency, calculates the deviation degree according to the preset normal threshold range corresponding to the operating parameters, compares the deviation degree with the preset deviation degree judgment threshold, and summarizes the types of abnormal parameters and deviation degree values ​​to form a parameter anomaly feature set. The failure mode matching unit retrieves the preset RCM failure mode library, matches the set of abnormal parameter features with the feature matching conditions of each failure mode one by one, calculates the feature matching degree based on the deviation of the successfully matched abnormal features, and selects failure modes with feature matching degrees greater than the preset matching degree judgment threshold as candidate failure modes.

[0020] The positive risk assessment module also includes a risk score calculation unit and a risk level determination unit; The risk score calculation unit calculates the single-mode risk score based on the feature matching degree of the candidate failure modes and the corresponding risk weights, and selects the maximum value of the single-mode risk score as the comprehensive risk score of the equipment. When there are no candidate failure modes, the comprehensive risk score of the equipment is assigned to zero. The risk level determination unit divides the overall risk score of the equipment into the corresponding risk level range according to the preset risk boundary threshold, integrates the candidate failure modes, the overall risk score of the equipment and the risk level of the equipment, and generates a positive risk judgment result.

[0021] The reverse reliability verification module includes a historical data processing unit and a benchmark threshold calculation unit; The historical data processing unit collects historical RCM risk data, historical fault record data, historical maintenance log data, and long-term steady-state operation parameter data of the unit equipment within a preset historical period, constructs a historical status dataset of the equipment, and updates the historical status dataset using a sliding window method at fixed time intervals. The benchmark threshold calculation unit calculates the arithmetic mean of the comprehensive risk scores for the historical statistical period to obtain the RCM risk benchmark value. It calculates the health status baseline value based on the normalized deviation mean of the historical statistical period. The risk threshold is calculated by combining the risk fluctuation tolerance coefficient, the RCM risk benchmark value and the health status baseline value.

[0022] The reverse reliability verification module also includes a slope calculation unit and a verification result generation unit; The slope calculation unit collects two sets of adjacent RCM risk benchmark values ​​before and after the sliding window updates, and divides the difference between the latter benchmark value and the former benchmark value by a fixed time interval to obtain the slope of the risk benchmark change. The verification result generation unit integrates the risk establishment condition judgment results, the risk benchmark change slope, and the health status baseline value to form a reverse credibility verification result.

[0023] The RCM-based method for detecting and maintaining unit equipment anomalies includes the following steps: S1. Collect real-time operating parameters of the unit equipment, compare the real-time operating parameters with thresholds, and identify abnormal parameter characteristics; match the abnormal parameter characteristics with the preset RCM failure mode library to obtain the corresponding failure mode, and preliminarily determine the equipment risk level based on the failure mode to form a positive risk judgment result; S2. Based on the historical RCM analysis results, historical fault and maintenance records, and long-term operating status data of the unit equipment, determine the current RCM risk benchmark and health status base of the equipment; based on the RCM risk benchmark and health status base, perform reverse credibility verification on the positive risk judgment results to determine whether the positive risk judgment results meet the conditions for risk establishment. S3. When there are abnormal parameter characteristics and the positive risk assessment result does not meet the conditions for risk establishment, the current abnormality is determined to be a pseudo-abnormality; when no abnormal parameter characteristics are identified but the RCM risk benchmark shows a continuous upward trend and the increase exceeds the preset trend threshold, the equipment is determined to have early deterioration risk; when there are abnormal parameter characteristics and the positive risk assessment result meets the conditions for establishment, it is determined to be a real risk. S4. Maintain normal operation for pseudo-anomalies, perform periodic monitoring and status tracking for early deterioration risks, output early warning information for real risks according to their risk level and execute corresponding maintenance plans, and complete the anomaly detection and maintenance of unit equipment.

[0024] In this embodiment, S1 includes the following: Real-time operating parameters of each monitoring point of the unit equipment are collected according to a preset acquisition frequency. These real-time operating parameters include, but are not limited to, temperature parameters, vibration amplitude parameters, medium pressure parameters, shaft speed parameters, and medium flow parameters. Each parameter is acquired by a corresponding sensor at a preset acquisition frequency. A real-time operating parameter set Pr={p1,p2,...,pi,...,pm} is constructed, where i is the parameter type number, i=1,2,...,m, and m is the total number of monitored parameters. A preset parameter threshold library is retrieved to obtain the preset threshold interval [pi_min,pi_max] corresponding to the i-th type of operating parameter, where pi_min is the preset lower limit value of the i-th type of operating parameter, and pi_max is the preset upper limit value of the i-th type of operating parameter. The interval span Δpi=pi_max-pi_min is calculated. When Δpi≠0, the formula for calculating the parameter deviation Di is: Di=|pi-[(pi_min+pi_max) / 2]| / Δpi; When Δpi=0, the formula for calculating the parameter deviation Di is: Di=|pi-pi_min|; Where Di is the parameter deviation of the i-th type of operating parameter, and pi is the real-time collected value of the i-th type of operating parameter; Based on the historical normal operation data of the unit equipment, a preset deviation judgment threshold D0 is calculated. Di is compared with D0. When Di > D0, the i-th type of operating parameter is judged to be abnormal. The types and deviation values ​​of all abnormal parameters are summarized to form a parameter abnormality feature set C={c1,c2,...,cj,...,cn}; where j is the abnormal feature number, j=1,2,...,n, n is the total number of abnormal features, and cj is the j-th parameter abnormal feature. The system retrieves a pre-defined RCM failure mode library, which stores failure mode identifiers Mk, feature matching conditions Qk, and risk weights Wk, where k is the failure mode number and the risk weight Wk ranges from 0 to 1. The RCM failure mode library is established as follows: for each key component of the unit equipment, based on historical fault records and FMEA analysis results, various typical failure modes and their corresponding abnormal parameter feature combinations, i.e., feature matching conditions, are determined. Simultaneously, based on the impact of each failure mode on equipment safety, production, and maintenance costs, a corresponding risk weight Wk is assigned, with a value range of 0-1. The RCM failure mode library can be pre-established before equipment operation based on design data and experience with similar equipment, and dynamically updated during operation based on newly added failure cases. The set of abnormal parameter features C is matched one by one with the feature matching conditions Qk of each failure mode, and the feature matching degree Sk is calculated, where Sk = (∑ j∈[1,n] μjk·Dj) / ∑ j∈[1,n] Dj, where Sk is the feature matching degree of the k-th failure mode; μjk is the correlation coefficient, which is 1 when the abnormal feature of the j-th parameter satisfies the feature matching condition Qk, otherwise μjk=0; Dj is the deviation degree corresponding to the abnormal feature of the j-th parameter. Based on historical equipment failure sample data, a matching degree judgment threshold S0 is set, and failure modes with Sk > S0 are selected as candidate failure modes. All candidate failure modes are summarized to obtain a candidate failure mode set. If the candidate failure mode set is not empty, for each candidate failure mode, a single-mode risk score Sck is calculated, and Sck = Sk × Wk, where Sck is the single-mode risk score of the k-th type of candidate failure mode, and Wk is the risk weight of the k-th type of failure mode. The maximum value among all single-mode risk scores is selected as the comprehensive equipment risk score Sc. If the candidate failure mode set is empty, the comprehensive equipment risk score Sc is assigned the value 0. Based on historical risk data or expert experience, preset risk thresholds ScL, ScM, and SCH satisfy 0 < ScL < ScM < SCH ≤ 1; where the risk interval [0, ScL) is no risk, [ScL, ScM) is low risk, [ScM, ScH) is medium risk, and [ScH, 1] is high risk; the comprehensive risk score Sc of the equipment is compared with the risk intervals in turn to determine the unique risk level of the equipment; the candidate failure modes, the comprehensive risk score of the equipment, and the risk level of the equipment are integrated to generate a positive risk judgment result R.

[0025] In this embodiment, S2 includes the following: The system collects historical RCM risk data, historical fault record data, historical maintenance log data, and long-term steady-state operating parameter data for the unit equipment within a preset historical time period. Historical RCM risk data refers to the comprehensive risk score Sc_ht for each historical statistical period, calculated periodically from historical operating data using S1, along with its corresponding failure mode and risk level. Historical fault record data refers to actual fault records that occurred during the equipment's historical operation, including at least the fault occurrence time, failure mode, faulty component, and fault handling measures. Historical maintenance log data refers to records of historical equipment maintenance activities, including at least the maintenance time, maintenance type, maintenance content, and replacement component information. Long-term steady-state operating parameter data refers to historical parameter sampling values ​​of the equipment under fault-free and stable operating conditions, used to construct a baseline for normal equipment operation; its parameter type is consistent with real-time operating parameters. A historical status dataset Dh is constructed. The historical status dataset is updated using a sliding window method at fixed time intervals Δt. The sliding window contains T consecutive historical statistical periods, where t is the number of a single historical statistical period, t=1,2,...,T. Based on the historical equipment status dataset Dh, calculate the RCM risk baseline value Br, and Br=(∑ t∈[1,T] Sc_ht) / T, where Br is the RCM risk benchmark value, with a value range of [0,1]; Sc_ht is the historical comprehensive risk score of the t-th historical statistical period, and T is the total number of historical statistical periods included in the sliding window; Based on the historical equipment status dataset Dh, the normalized mean deviation of all operating parameters within the t-th historical statistical period is extracted and denoted as Dnt; the baseline health status value Bh is calculated based on the normalized mean deviation Dnt, and Bh = 1 - [(∑ t∈[1,T] Dnt) / T] 2 Where Bh is the baseline value of health status, with a value range of [0,1], and Dnt is the normalized average deviation of the operating parameters in the t-th historical statistical period. Based on historical fault boundary statistics, a risk fluctuation tolerance coefficient α is set, where 0 < α < 1; combining the RCM risk benchmark value Br and the health status baseline value Bh, the risk establishment threshold Thr is calculated, where Thr = Br + α × (1 - Br) × (1 - Bh); the equipment comprehensive risk score Sc generated by S1 is extracted, and the equipment comprehensive risk score Sc is compared with the risk establishment threshold Thr. When Sc > Thr, the positive risk judgment result is determined to meet the risk establishment condition; when Sc ≤ Thr, the positive risk judgment result is determined not to meet the risk establishment condition; two sets of adjacent RCM risk benchmark values ​​before and after the sliding window update are collected and recorded as the previous risk benchmark value Br1 and the subsequent risk benchmark value Br2, respectively. The risk benchmark change slope Kr is calculated, where Kr = (Br2 - Br1) / Δt, where Δt is the fixed time interval corresponding to the two sets of benchmark values; the risk establishment condition judgment result, the risk benchmark change slope, and the health status baseline value are summarized to form the reverse reliability verification result.

[0026] In this embodiment, S3 includes the following: Extract the abnormal feature set C of parameters obtained from S1 and the comprehensive risk score of equipment Sc. Extract the risk establishment condition judgment result and risk benchmark change slope Kr obtained from S2. Preset risk trend threshold K0, and K0 is statistically calibrated based on the risk benchmark fluctuation data of the unit equipment in long-term steady-state operation. It is used to quantify the effective deterioration range of the continuous increase of equipment risk. K0 is a non-negative fixed threshold. Equipment status types are classified through a multi-dimensional combination of judgment logic based on the presence or absence of abnormal parameter characteristics, risk establishment conditions, and risk benchmark change trends; specifically: When the set of abnormal parameters C is empty, that is, no abnormal parameters of the equipment in real time are identified, and the slope of the risk benchmark change Kr ≤ K0, the equipment is determined to be in normal operation and no abnormal handling or maintenance action is required. When the parameter anomaly feature set C is a non-empty set, meaning that the equipment has real-time parameter anomalies and the positive risk assessment result does not meet the conditions for risk establishment, the parameter anomaly currently triggered by the equipment is determined to be a pseudo-anomaly. A pseudo-anomaly indicates that the deviation of the equipment's real-time parameters is only a temporary anomaly caused by operating condition switching, instantaneous load fluctuations, or slight sensor interference, and is not caused by equipment failure or deterioration. In particular, if Kr > K0 is satisfied in this case, in addition to determining it as a pseudo-anomaly, an early deterioration risk warning is output, indicating that although the current anomaly of the equipment is a temporary fluctuation, the long-term risk baseline has a continuous upward trend. When the set of abnormal parameters C is empty, meaning no abnormal parameters of the equipment in real time are identified, and the slope of the risk benchmark change Kr > K0, it is determined that the RCM risk benchmark shows a continuous upward trend and the increase exceeds the preset trend threshold, indicating that the equipment has an early deterioration risk. Among them, the slope of the risk benchmark change Kr > K0 indicates that the long-term overall risk baseline of the equipment continues to rise, and there is a gradual, implicit deterioration of the internal components of the equipment without explicit parameter representation, which belongs to the potential risk of early failure. When the set of abnormal parameters C is a non-empty set, that is, when the equipment has real-time abnormal parameters and the positive risk assessment result meets the conditions for risk establishment, it is determined that the equipment currently has a real risk; the real risk indicates that the real-time abnormal parameters of the equipment are caused by the actual deterioration of the equipment's main components due to failure or performance degradation, and do not belong to temporary operating condition fluctuation abnormalities. Integrate the judgment results, output the final abnormality judgment type of the equipment, and complete the multi-dimensional risk classification judgment process of the equipment.

[0027] In this embodiment, S4 includes the following: When the equipment is determined to be in normal operating condition, the current operating condition of the unit equipment shall remain unchanged, and no abnormal warning shall be triggered, no additional monitoring tasks shall be started, and no maintenance operations shall be performed. When the equipment is determined to be a pseudo-anomaly and there is no early deterioration risk warning, the unit equipment should be kept in normal and continuous operation. The abnormal alarm caused by the temporary parameter fluctuation should be ignored, and no shutdown operation or maintenance intervention should be performed to avoid ineffective operation and maintenance and unnecessary shutdown. When the equipment is identified as a false anomaly and an early deterioration risk warning is simultaneously superimposed, the current operating condition of the unit equipment is maintained, the false anomaly alarm is blocked, and encrypted periodic status tracking and monitoring is started at the same time. The risk baseline update cycle and parameter sampling interval are shortened, the equipment risk baseline change trend is continuously recorded, and the hidden deterioration is prevented from continuing to worsen in advance. When equipment is identified as having an early risk of degradation, emergency alarms are not triggered and shutdown maintenance is not performed. Instead, a special degradation tracking ledger is established, and the deviation of equipment parameters and the trend of changes in the RCM risk benchmark are continuously tracked according to a preset fixed monitoring cycle. The health status and risk level of the equipment are dynamically updated, and the risk handling strategy is switched in a timely manner when the risk indicators reach the conditions for real risk judgment, so as to achieve early warning of equipment degradation. When equipment is determined to pose a real risk, the risk level and candidate failure modes generated by S1 are retrieved, and graded early warning and matching maintenance plans are executed according to different risk levels. Specifically, low-risk levels output routine operation and maintenance early warnings and arrange planned inspections and component status checks; medium-risk levels output key monitoring early warnings, shorten equipment inspection cycles, increase parameter sampling frequency, and specifically investigate key components corresponding to candidate failure modes; high-risk levels output emergency risk early warnings, initiate equipment fault investigation processes, and, if necessary, perform load reduction operation or shutdown maintenance operations to eliminate real deterioration faults in the equipment itself. Complete differentiated monitoring, early warning, and maintenance actions corresponding to all states, and realize refined anomaly detection and hierarchical maintenance based on RCM bidirectional verification.

[0028] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0029] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are 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 method for detecting and maintaining unit equipment anomalies based on RCM, characterized in that: The method includes the following steps: S1. Collect real-time operating parameters of the unit equipment, compare the real-time operating parameters with thresholds, and identify abnormal parameter characteristics; match the abnormal parameter characteristics with the preset RCM failure mode library to obtain the corresponding failure mode, and preliminarily determine the equipment risk level based on the failure mode to form a positive risk judgment result; S2. Based on the historical RCM analysis results, historical fault and maintenance records, and long-term operating status data of the unit equipment, determine the current RCM risk benchmark and health status base of the equipment; based on the RCM risk benchmark and health status base, perform reverse credibility verification on the positive risk judgment results to determine whether the positive risk judgment results meet the conditions for risk establishment. S3. When there are abnormal parameter characteristics and the positive risk assessment result does not meet the conditions for risk establishment, the current abnormality is determined to be a pseudo-abnormality; when no abnormal parameter characteristics are identified but the RCM risk benchmark shows a continuous upward trend and the increase exceeds the preset trend threshold, the equipment is determined to have early deterioration risk; when there are abnormal parameter characteristics and the positive risk assessment result meets the conditions for establishment, it is determined to be a real risk. S4. Maintain normal operation for pseudo-anomalies, perform periodic monitoring and status tracking for early deterioration risks, output early warning information for real risks according to their risk level and execute corresponding maintenance plans, and complete the anomaly detection and maintenance of unit equipment.

2. The method for unit equipment anomaly detection and maintenance based on RCM according to claim 1, characterized in that: S1 includes the following: Real-time operating parameters of each monitoring point of the unit equipment are collected according to the preset collection frequency. The deviation of each operating parameter is calculated according to the preset normal threshold range corresponding to each operating parameter. The deviation is compared with the preset deviation judgment threshold. When the deviation is greater than the judgment threshold, the corresponding operating parameter is judged to be abnormal. All abnormal parameter types and deviation values ​​are summarized to form a parameter abnormality feature set. Retrieve the preset RCM failure mode library, which stores the feature matching conditions and risk weights corresponding to each failure mode. Match the abnormal feature set of parameters with the feature matching conditions of each failure mode one by one. Calculate the feature matching degree of each failure mode based on the deviation of the successfully matched abnormal features. Select failure modes with feature matching degrees greater than the preset matching degree judgment threshold as candidate failure modes. For each candidate failure mode, a single-mode risk score is calculated based on the feature matching degree and the corresponding risk weight. The maximum single-mode risk score among all candidate failure modes is selected as the overall equipment risk score. If there is no candidate failure mode, the overall equipment risk score is assigned to zero. Based on the preset risk threshold, the overall risk score of the equipment is divided into the corresponding risk level range to determine the risk level of the equipment. The candidate failure modes, the overall risk score of the equipment and the risk level of the equipment are integrated to generate a positive risk judgment result.

3. The method for unit equipment anomaly detection and maintenance based on RCM according to claim 2, characterized in that: S2 includes the following: The system collects historical RCM risk data, historical fault record data, historical maintenance log data, and long-term steady-state operating parameter data of the unit equipment within a preset historical time period. The historical RCM risk data is obtained by calculating the comprehensive risk score and corresponding failure mode and risk level of each historical statistical period by calculating the historical operating data period by period. Historical fault record data includes at least the fault occurrence time, fault failure mode, faulty component and fault handling measures. Historical maintenance log data includes at least the maintenance time, maintenance type, maintenance content and replacement component information. Long-term steady-state operating parameter data are historical parameter sampling values ​​of the equipment under fault-free and stable operating conditions. The collected data is used to construct the equipment historical status dataset. The historical state dataset is updated at fixed time intervals using a sliding window method, and the sliding window contains multiple consecutive historical statistical periods. Based on the historical status dataset of the equipment, the arithmetic mean of all comprehensive risk scores within the historical statistical period is calculated and used as the RCM risk benchmark value. Based on the historical status dataset of the equipment, the normalized deviation mean of all operating parameters in each historical statistical period is extracted, and the health status baseline value is calculated based on the normalized deviation mean of all historical statistical periods. Based on historical failure boundary statistics, a risk fluctuation tolerance coefficient is set, and the risk benchmark value is combined with the health status baseline value to calculate the risk threshold. Extract the comprehensive risk score of the equipment from the positive risk assessment result, compare the comprehensive risk score of the equipment with the risk establishment threshold, and determine that the positive risk assessment result meets the risk establishment condition when the comprehensive risk score of the equipment is greater than the risk establishment threshold; otherwise, it is determined that the risk establishment condition is not met. Collect two sets of adjacent RCM risk benchmark values ​​before and after the sliding window update, and divide the difference between the latter benchmark value and the former benchmark value by a fixed time interval to obtain the slope of the risk benchmark change. The results of risk establishment criteria determination, risk benchmark change slope, and health status baseline value are summarized to form a reverse credibility verification result.

4. The method for unit equipment anomaly detection and maintenance based on RCM according to claim 3, characterized in that: S3 includes the following: Extract the set of abnormal parameter features and the comprehensive risk score of the equipment from the positive risk assessment results, and extract the risk establishment condition assessment results and the risk benchmark change slope from the reverse credibility verification results; A preset risk trend threshold is defined based on statistical calibration of risk benchmark fluctuation data of long-term steady-state operation of unit equipment, and is used to quantitatively determine the effective deterioration range of continuous increase in equipment risk. When the set of abnormal parameter features is empty and the slope of the risk benchmark change is less than or equal to the risk trend threshold, the equipment is determined to be in normal operating condition. When the set of abnormal parameter features is not empty and the positive risk assessment result does not meet the conditions for risk establishment, the current parameter abnormality is determined to be a pseudo-abnormality. If the slope of the risk benchmark change is greater than the risk trend threshold under this circumstance, an early deterioration risk warning will be output in addition to identifying false anomalies. When the set of abnormal parameters is empty and the slope of the risk benchmark change is greater than the risk trend threshold, it is determined that the equipment has an early deterioration risk. When the set of abnormal parameter features is not empty and the positive risk assessment result meets the conditions for risk establishment, the equipment is determined to have a real risk. The integrated judgment results are used to output the final abnormality judgment type of the device.

5. The method for unit equipment anomaly detection and maintenance based on RCM according to claim 4, characterized in that: S4 includes the following: When the equipment is determined to be in normal operating condition, the current operating condition of the unit equipment remains unchanged, no abnormal warnings are triggered, no additional monitoring tasks are started, and no maintenance operations are performed; When the equipment is determined to be a false anomaly and there is no early deterioration risk warning, the unit equipment should be kept in normal continuous operation, the anomaly alarm should be masked, and no shutdown or maintenance intervention should be performed. When the equipment is identified as a pseudo-anomaly and an early deterioration risk warning is simultaneously superimposed, the current operating condition of the unit equipment is maintained, the current anomaly alarm is masked, and encrypted periodic status tracking and monitoring is started at the same time. The risk baseline update cycle and parameter sampling interval are shortened, and the trend of equipment risk baseline change is continuously recorded. When equipment is identified as having an early risk of deterioration, emergency alarms are not triggered, shutdown maintenance is not performed, a special deterioration tracking ledger is established, and the deviation of equipment parameters from the risk benchmark is continuously tracked according to the preset monitoring cycle. The health status and risk level of the equipment are dynamically updated, and the handling strategy is switched when the risk indicators reach the conditions for real risk judgment. When a device is identified as a real risk, the device risk level and candidate failure modes generated from the positive risk assessment results are retrieved, and a graded early warning and corresponding maintenance plan are executed according to the risk level.

6. An RCM-based unit equipment anomaly detection and maintenance system, applied to the RCM-based unit equipment anomaly detection and maintenance method according to any one of claims 1-5, characterized in that: The system includes: a positive risk assessment module, a reverse reliability verification module, an equipment status classification module, and a hierarchical operation and maintenance handling module; The positive risk assessment module collects real-time operating parameters of the unit equipment, performs threshold comparison on the real-time operating parameters and identifies abnormal parameter characteristics, matches the abnormal parameter characteristics with the preset RCM failure mode library, determines the equipment risk level based on the matched failure modes, and generates a positive risk assessment result. The reverse reliability verification module determines the RCM risk benchmark and health status base based on the historical RCM analysis results, historical fault and maintenance records and long-term operating status data of the unit equipment. Based on the RCM risk benchmark and health status base, it performs reverse reliability verification on the positive risk judgment results to determine whether the positive risk judgment results meet the conditions for risk establishment. The equipment status classification module distinguishes between pseudo-anomalies, early degradation risks, and real risks based on the existence status of abnormal parameter characteristics, the validity status of positive risk judgment results, and the changing trend of RCM risk benchmark, and outputs the final anomaly judgment type of the equipment. The hierarchical operation and maintenance module executes corresponding actions based on the final anomaly determination type of the equipment. It maintains normal equipment operation for false anomalies, performs periodic monitoring and status tracking for early deterioration risks, and outputs early warning information and executes corresponding maintenance plans for real risks based on risk levels.

7. The unit equipment anomaly detection and maintenance system based on RCM according to claim 6, characterized in that: The positive risk assessment module includes a parameter acquisition anomaly identification unit and a failure mode matching unit; The parameter acquisition anomaly identification unit acquires real-time operating parameters of each monitoring point of the unit equipment according to a preset acquisition frequency, calculates the deviation degree according to the preset normal threshold range corresponding to the operating parameters, compares the deviation degree with the preset deviation degree judgment threshold, and summarizes the types of abnormal parameters and the deviation degree values ​​to form a parameter anomaly feature set. The failure mode matching unit retrieves a preset RCM failure mode library, matches the parameter abnormal feature set with the feature matching conditions of each failure mode one by one, calculates the feature matching degree based on the deviation of the successfully matched abnormal features, and selects failure modes with feature matching degrees greater than the preset matching degree judgment threshold as candidate failure modes.

8. The unit equipment anomaly detection and maintenance system based on RCM according to claim 6, characterized in that: The positive risk assessment module also includes a risk score calculation unit and a risk level determination unit; The risk score calculation unit calculates the single-mode risk score based on the feature matching degree of the candidate failure modes and the corresponding risk weights, selects the maximum value of the single-mode risk score as the comprehensive risk score of the equipment, and assigns the comprehensive risk score of the equipment to zero when there are no candidate failure modes. The risk level determination unit divides the overall risk score of the equipment into corresponding risk level ranges according to a preset risk boundary threshold, integrates candidate failure modes, overall risk score of the equipment and risk level of the equipment, and generates a positive risk judgment result.

9. The unit equipment anomaly detection and maintenance system based on RCM according to claim 6, characterized in that: The reverse credibility verification module includes a historical data processing unit and a benchmark threshold calculation unit; The historical data processing unit collects historical RCM risk data, historical fault record data, historical maintenance ledger data, and long-term steady-state operation parameter data of the unit equipment within a preset historical period, constructs a historical status dataset of the equipment, and updates the historical status dataset using a sliding window method at fixed time intervals. The benchmark threshold calculation unit calculates the arithmetic mean of the comprehensive risk scores for the historical statistical period to obtain the RCM risk benchmark value, calculates the health status baseline value based on the normalized deviation mean of the historical statistical period, and calculates the risk establishment threshold by combining the risk fluctuation tolerance coefficient, the RCM risk benchmark value and the health status baseline value.

10. The unit equipment anomaly detection and maintenance system based on RCM according to claim 6, characterized in that: The reverse credibility verification module also includes a slope calculation unit and a verification result generation unit; The slope calculation unit collects two sets of adjacent RCM risk benchmark values ​​before and after the sliding window updates, and divides the difference between the latter benchmark value and the former benchmark value by a fixed time interval to obtain the slope of the risk benchmark change. The verification result generation unit integrates the risk establishment condition judgment result, the risk benchmark change slope, and the health status baseline value to form a reverse credibility verification result.

Images