Multi-dimensional evaluation and intelligent judgment system for maintenance of industrial energy power system

Abstract

The invention discloses a multi-dimensional evaluation and intelligent judgment system for maintenance of an industrial energy power system. The system comprises an operation performance index acquisition module, a feature data acquisition module, a weight inference and correction reference generation module, a maintenance effect recovery rate calculation module, a hierarchical maintenance effect judgment module and a hierarchical level dynamic adjustment mechanism. The system dynamically deduces a physical characteristic weight through operation performance change before maintenance, generates a corrected maintenance target benchmark, calculates a comprehensive recovery rate based on an actual performance recovery degree after maintenance, hierarchically judges a maintenance effect grade according to the comprehensive recovery rate, introduces a dynamic monitoring window, finely adjusts a judgment threshold according to a performance fluctuation trend, and finally judges the maintenance effect grade. Therefore, accurate quantitative evaluation of the maintenance effect and intelligent output of a subsequent operation and maintenance strategy are realized. The method has the advantages of being high in adaptability, high in judgment accuracy and good in abnormal disturbance resistance.

Description

Technical Field

[0001] The present invention relates to the technical field of energy power system maintenance, and in particular to a multidimensional evaluation and intelligent determination system for industrial energy power system maintenance. Background Art

[0002] In modern industrial energy and power systems, system performance stability and efficiency are key factors in ensuring overall operational efficiency, reducing energy consumption, and reducing maintenance costs. Over long-term system operation, internal components such as heat exchangers, piping, sensors, and fluid channels often experience gradual performance degradation due to factors such as physical wear, chemical deposition, environmental contamination, and media aging. This degradation not only leads to decreased energy efficiency and increased fluid resistance, but can also cause slow response in key components, control instability, or increased failure rates, impacting the operational safety and economic efficiency of the entire system.

[0003] To maintain the long-term health of a system, engineering practices typically employ methods such as regular maintenance, component replacement, and system refreshes to restore the system's state. However, current practices often rely on static indicators, single parameter changes, or manual judgment to determine the effectiveness of system maintenance and repairs. Dynamic assessment methods closely tied to the system's overall performance recovery are lacking. This limitation makes it difficult to scientifically quantify and rank maintenance or tuning operations, often leading to inaccurate maintenance decisions, excessive maintenance, or delayed interventions, hindering the refined management and intelligent operation and maintenance of complex systems.

[0004] Therefore, a universal intelligent assessment mechanism for system performance recovery is urgently needed. This mechanism can combine key performance indicators and physical response characteristics before and after system operation to construct dynamic correction and multidimensional analysis models, accurately characterize performance trends, and implement a comprehensive assessment system with strong scenario adaptability, flexible response mechanisms, and highly operational assessment results. This system should not only serve the identification and assessment of operating status but also guide subsequent maintenance decisions and optimize strategy output, thereby enabling intelligent and precise system operation management in multiple industrial fields. Summary of the Invention

[0005] The purpose of the present invention is to provide a multidimensional evaluation and intelligent judgment system for industrial energy and power system maintenance, which has the advantages of high judgment accuracy, strong resistance to abnormal interference, good linkage of operation and maintenance strategies, and strong system adaptability.

[0006] The above technical objectives of the present invention are achieved through the following technical solutions: Multi-dimensional evaluation and intelligent judgment system for industrial energy and power system maintenance, including: An operating performance index collection module is used to collect operating performance indicators of the industrial energy power system before and after maintenance operations, wherein the operating performance indicators include heat exchange efficiency, heat exchange per unit flow rate, and total pressure drop level; Characteristic data acquisition module, used to collect physical characteristic data of industrial energy power systems before maintenance operations, including unit disturbance recovery time, pressure drop under standard flow, standing wave peak frequency offset and acoustic wave energy attenuation change rate; The weight inference and correction benchmark generation module is used to dynamically determine the final weight of each physical characteristic data based on the pre-maintenance operating performance indicators, calculate the health correction coefficient based on the final weight and physical characteristic data, and modify the original design benchmark parameters accordingly to generate the target benchmark parameters after maintenance; The maintenance effect recovery rate calculation module is used to compare the operating performance indicators collected after maintenance with the revised target benchmark parameters to calculate the comprehensive recovery rate; The hierarchical maintenance effect determination module is used to divide the maintenance effect into different levels according to the comprehensive recovery rate and output the corresponding subsequent operation and maintenance strategy.

[0007] By adopting the above technical solution and building a closed-loop system from pre-maintenance collection, dynamic inference, post-maintenance recovery rate assessment to hierarchical judgment, the system can achieve accurate quantitative evaluation of maintenance effects, dynamically correct the design benchmark, make the basis for recovery rate calculation more realistic and reliable, avoid benchmark distortion caused by long-term aging, and introduce a hierarchical judgment and operation and maintenance strategy output mechanism, which can effectively guide subsequent operation and maintenance decisions and improve the overall operation efficiency and reliability of the system.

[0008] Further configuration: the feature data acquisition module specifically includes: The disturbance response acquisition submodule is used to apply a flow disturbance of a preset amplitude to the fluid path of the industrial energy power system before maintenance, and collect the response curve of the flow velocity or flow rate changing with time after the disturbance is applied in real time, and extract the unit disturbance recovery time constant based on curve fitting; The pressure drop acquisition submodule is used to collect the inlet and outlet pressure values of the industrial energy power system under standard flow conditions before maintenance, and calculate the pressure drop level under standard flow; The acoustic standing wave acquisition submodule is used to apply an acoustic wave signal within a preset frequency range inside the maintenance forward system, collect the reflected acoustic wave signal, and extract the strongest reflection peak frequency; The acoustic energy attenuation acquisition submodule is used to measure the ratio of the acoustic signal input power to the reflected signal received power, and calculate the acoustic energy attenuation change rate based on the change in the ratio.

[0009] By adopting the above technical solution, the feature data collection process before maintenance is modularized and subdivided, and various physical quantity detection methods such as disturbance response, standard voltage drop, acoustic standing wave, and acoustic energy attenuation are used to characterize the internal degradation state of the system from multiple angles. The collection method is standardized and has a physical basis, which improves the accuracy and adaptability of the detection data and facilitates flexible application in different systems.

[0010] Further configuration: the weight inference and correction benchmark generation module specifically includes: a macro-performance change correlation unit for inferring the contribution of each collected physical characteristic to the performance degradation of the industrial energy power system based on the decline of the operating performance indicators before maintenance, and assigning a preliminary weight factor to each physical characteristic accordingly, wherein the operating performance indicators include heat transfer efficiency, heat transfer per unit flow rate, and total pressure drop; A dynamic weight determination unit is used to determine the final weight based on the preliminary weight factors of each physical feature, select preset anomaly detection rules to analyze the physical feature data collected before maintenance, and increase the final weight of a physical feature accordingly when the physical feature data meets the anomaly condition; The benchmark correction generating unit is used to calculate the health correction coefficient based on the final weight of each physical feature and the physical feature data, and apply the health correction coefficient to the original design benchmark parameter to generate the target benchmark parameter after maintenance.

[0011] By adopting the above technical solution, the operating performance degradation characteristics are dynamically associated with physical characteristics to perform preliminary weight inference, avoiding the judgment deviation caused by fixed preset weights. Anomaly detection rules are introduced to dynamically adjust the weights, thereby improving the system's sensitivity to local abnormal deterioration trends. The benchmark correction generation process is based on physical quantitative data and a weighting mechanism to achieve adaptive correction of the maintenance target benchmark, thereby improving the scientificity and accuracy of the recovery rate judgment.

[0012] Further configuration: the aforementioned selecting preset anomaly detection rules to analyze the physical feature data collected before maintenance, and when a certain physical feature data meets an abnormal condition, correspondingly increasing the final weight of the physical feature specifically includes: Multi-level abnormality judgment thresholds are set for unit disturbance recovery time, pressure drop under standard flow, standing wave peak frequency offset and acoustic wave energy attenuation, including the first threshold, second threshold and third threshold; When the detection data of a certain physical feature is between the first threshold and the second threshold, the final weight of the physical feature is adjusted according to the first weighting factor k1>1; When the physical feature data is between the second threshold and the third threshold, the final weight is adjusted according to the second weighting factor k2>k1; When the physical feature data exceeds the third threshold, the final weight is adjusted according to the third weighting factor k3>k2.

[0013] By adopting the above technical solution, multi-level abnormality judgment and graded weighting methods are used to dynamically adjust weights according to different degrees of characteristic abnormalities, and the discriminative value of characteristic data is improved in a fine-grained manner, which strengthens the system's ability to identify degradation phenomena of different degrees of severity and improves the sensitivity and resolution of the overall system health assessment.

[0014] Further setting: the maintenance effect is divided into different levels according to the comprehensive recovery rate, and the corresponding subsequent operation and maintenance strategy is output, specifically including: Set a first maintenance threshold R1, a second maintenance threshold R2, and a third maintenance threshold R3, where R1>R2>R3; When the comprehensive recovery rate is greater than or equal to the first maintenance threshold, it is determined to be at the full recovery level, and an operation and maintenance strategy for maintaining normal operation is output; When the comprehensive recovery rate is less than the first maintenance threshold but greater than or equal to the second maintenance threshold, it is determined to be a good recovery level, and an operation and maintenance strategy of local periodic flushing and operation status monitoring is output; When the comprehensive recovery rate is less than the second maintenance threshold but greater than or equal to the third maintenance threshold, it is determined to be a local residual level, and an operation and maintenance strategy of strengthening local maintenance or shortening the maintenance cycle is output; When the comprehensive recovery rate is less than the third maintenance threshold, it is determined to be a maintenance failure level, and an operation and maintenance strategy of re-maintenance or replacement of relevant components is output.

[0015] By adopting the above technical solutions, comprehensive recovery rate determination and hierarchical classification mechanism, the maintenance effect evaluation is transformed from a single numerical value to a systematic classification. Each level corresponds to a clear operation and maintenance strategy recommendation, realizing the close linkage between maintenance effect evaluation and subsequent operation and maintenance actions, and improving the level of intelligent system management.

[0016] Further setting: The maintenance effect recovery rate calculation module further includes a dynamic monitoring unit, which is used to set a preset monitoring window period after the maintenance operation is completed, continuously collect the changing trend of the operating performance indicators of the industrial energy power system, and correct the comprehensive recovery rate based on the change amplitude of the operating performance indicators within the monitoring window, and correct the maintenance effect judgment result.

[0017] By adopting the above technical solution and introducing a dynamic post-maintenance monitoring window mechanism, the recovery rate is corrected in real time according to the performance change trend, avoiding the interference of short-term abnormal fluctuations in the judgment of maintenance effect, making the system evaluation results more stable and reliable, and enhancing the robustness and adaptability of the judgment mechanism.

[0018] Further configuration: the hierarchical maintenance effect determination module dynamically fine-tunes the first maintenance threshold, the second maintenance threshold, and the third maintenance threshold based on the change range of the operating performance indicator within the post-maintenance dynamic monitoring window after the maintenance operation, specifically including: When the fluctuation range of the operating performance indicator within the monitoring window is lower than the preset stability standard, the original set threshold is maintained; When the fluctuation range of the operating performance indicator within the monitoring window exceeds the preset standard, the set values of the first maintenance threshold and the second maintenance threshold are respectively lowered.

[0019] By adopting the above technical solution, the maintenance effect judgment threshold is dynamically fine-tuned, so that the stratification standard can be flexibly changed according to the actual status of the system, improving the system's ability to cope with complex operating environments, avoiding misjudgments caused by fixed threshold settings, and improving the accuracy and dynamic adaptability of maintenance effect stratification judgment.

[0020] Further configuration: the dynamic weight determination unit further includes a feature association weight modulation unit, which is used to detect the correlation between each physical feature data during the dynamic weight determination process, and when two or more physical feature data simultaneously meet the anomaly detection rules, jointly amplify the final weight of the corresponding physical feature.

[0021] By adopting the above technical solution and introducing a feature association weight modulation mechanism, when multiple physical features are abnormal at the same time, collaborative amplification processing is performed to strengthen the system's ability to identify systematic degradation trends, effectively improve the system's detection sensitivity to hidden and complex degradation phenomena, and enhance the depth and breadth of intelligent evaluation of maintenance effects.

[0022] In summary, the present invention has the following beneficial effects: by constructing an intelligent maintenance effect evaluation system based on dynamic benchmark correction and system performance orientation, introducing multi-angle collection of operating performance and physical characteristics, dynamic inference weights, anomaly detection and feature association weight modulation, as well as hierarchical maintenance effect judgment and threshold dynamic adjustment mechanism, accurate quantitative evaluation, intelligent grade classification and dynamic adaptive adjustment of the maintenance effect of industrial energy and power systems are achieved. The present invention not only overcomes the problems of traditional maintenance effect judgment being single, static and prone to misjudgment, but also can optimize the judgment criteria in real time according to the actual system operating status, significantly improving the scientific nature, intelligence and system reliability of maintenance effect management, and has broad application prospects in the fields of industrial energy-saving operation and maintenance, industrial energy and power system life management, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is an overall structural block diagram of the embodiment. DETAILED DESCRIPTION

[0024] The present invention will be further described in detail below with reference to the accompanying drawings.

[0025] Example: like Figure 1 As shown in the figure, the multi-dimensional evaluation and intelligent judgment system for industrial energy and power system maintenance includes: An operating performance index collection module is used to collect operating performance indicators of the industrial energy power system before and after maintenance operations, wherein the operating performance indicators include heat exchange efficiency, heat exchange per unit flow rate, and total pressure drop level; Characteristic data acquisition module, used to collect physical characteristic data of industrial energy power systems before maintenance operations, including unit disturbance recovery time, pressure drop under standard flow, standing wave peak frequency offset and acoustic wave energy attenuation change rate; The weight inference and correction benchmark generation module is used to dynamically determine the final weight of each physical characteristic data based on the pre-maintenance operating performance indicators, calculate the health correction coefficient based on the final weight and physical characteristic data, and modify the original design benchmark parameters accordingly to generate the target benchmark parameters after maintenance; The maintenance effect recovery rate calculation module is used to compare the operating performance indicators collected after maintenance with the revised target benchmark parameters to calculate the comprehensive recovery rate; The hierarchical maintenance effect determination module is used to divide the maintenance effect into different levels according to the comprehensive recovery rate and output the corresponding subsequent operation and maintenance strategy.

[0026] The system's overall logic is as follows: Pre-maintenance data is collected to infer the system's current health status, baseline parameters are dynamically revised, and new data is collected and evaluated after maintenance. This allows for dynamic, adaptive, and highly accurate maintenance effectiveness assessments and operational decision-making. By integrating operational performance and physical characteristic information, maintenance effectiveness assessments are more comprehensive and accurate, avoiding misjudgments caused by a single data source and improving the operational stability and maintenance efficiency of industrial energy and power systems.

[0027] The operating performance index acquisition module is used to collect the heat exchange efficiency, heat exchange per unit flow rate, and total pressure drop level of the industrial energy power system before and after the maintenance operation. The heat exchange efficiency can be collected in real time by the temperature sensor to collect the inlet and outlet temperatures, and calculated in combination with the flow meter measurement results; the heat exchange per unit flow rate is derived in real time based on the relationship between heat exchange power and flow rate; the total pressure drop level can be directly measured by the pressure differential sensor under standard flow conditions. The heat exchange efficiency can reflect the overall thermal energy utilization capacity and is a key indicator for directly evaluating the maintenance effect. The heat exchange per unit flow rate can eliminate the influence of flow changes and ensure the consistency of comparison under different working conditions. The total pressure drop level is highly sensitive to internal structure blockage and flow resistance changes, and is an important reflection of the health status of the fluid channel before and after maintenance. The combination of the three can achieve a multi-dimensional comprehensive judgment of the maintenance effect and improve the accuracy and robustness of the system evaluation.

[0028] In a preferred embodiment, the feature data acquisition module includes the following submodules: The disturbance response acquisition submodule is used to apply a flow disturbance of a preset amplitude to the fluid pathways of the industrial energy and power system before maintenance. It then collects the response curve of the flow velocity or flow rate over time after the disturbance is applied in real time, and extracts the unit disturbance recovery time constant based on curve fitting. Preferably, the disturbance amplitude is within the range of 5%-15% of the standard flow rate to stimulate the system's dynamic characteristics without causing system instability. The reasons and advantages of using flow disturbance acquisition and recovery time extraction are: small disturbances will not disrupt the stable operation of the system, and the test is highly safe; changes in recovery time can early reflect the system's internal deposition, blockage, and dynamic response degradation processes. This indicator is characterized by high sensitivity and fast response speed. By analyzing the disturbance recovery speed, the degree of deviation of the current system operating state from the design state can be inferred.

[0029] The pressure drop acquisition submodule collects inlet and outlet pressures of industrial energy power systems at standard flow rates before maintenance and calculates the pressure drop at standard flow rates. The standard flow rate is set at 90%-110% of the design flow rate. The pressure drop directly reflects changes in internal flow resistance and can sensitively detect system blockages caused by scaling and clogging.

[0030] The acoustic standing wave acquisition submodule applies an acoustic signal within a preset frequency range to the maintenance front system, collects the reflected acoustic signal, and extracts the frequency of the strongest reflected peak. The acoustic excitation frequency is set between 1kHz and 20kHz. The standing wave frequency is highly sensitive to internal structural changes and can detect subtle local anomalies that are difficult to detect with conventional pressure drop testing. This non-contact acoustic detection enables real-time, online diagnosis without affecting system operation.

[0031] The acoustic energy attenuation acquisition submodule is used to measure the ratio of the input power of the acoustic signal to the received power of the reflected signal, and calculate the rate of change of the acoustic energy attenuation based on the change in the ratio. Acoustic energy attenuation is highly sensitive to microscopic changes within the system and is particularly suitable for discovering early local deposition, microcracks, corrosion and other phenomena. Compared with traditional single-frequency sound intensity monitoring, the energy ratio method has strong anti-interference ability and is suitable for application under complex background noise conditions in actual industrial environments. It can supplement the deficiencies of pressure drop and standing wave indicators and improve the stability and refinement of the overall system health assessment.

[0032] The weight inference and correction benchmark generation module specifically includes: The macro performance change correlation unit is used to infer the contribution of each collected physical feature to the performance degradation of the industrial energy power system based on the decline of the operating performance indicators before maintenance, and to assign preliminary weight factors to each physical feature accordingly; A dynamic weight determination unit is used to determine the final weight based on the preliminary weight factors of each physical feature, select preset anomaly detection rules to analyze the physical feature data collected before maintenance, and increase the final weight of a physical feature accordingly when the physical feature data meets the anomaly condition; The benchmark correction generating unit is used to calculate the health correction coefficient based on the final weight of each physical feature and the physical feature data, and apply the health correction coefficient to the original design benchmark parameter to generate the target benchmark parameter after maintenance.

[0033] Before maintenance, the operating performance indicators of the industrial energy and power system are first collected to determine the overall performance degradation characteristics of the system; based on the correlation between each physical characteristic and different performance degradation types, the importance of each physical characteristic is dynamically inferred and a preliminary weight factor is assigned; then, the preset anomaly detection rules are applied to analyze the physical characteristic data collected before maintenance, and the preliminary weight is dynamically adjusted to obtain the final weight; finally, the health correction coefficient is calculated based on the final weight and physical characteristic data, the original design benchmark is corrected, and the target benchmark parameters after maintenance are generated.

[0034] Before maintenance, system performance indicators are collected, including heat transfer efficiency, heat transfer per unit flow rate, and total pressure drop. These indicators are compared with the original design baseline values, and the rate of decline of each indicator is calculated. Based on the pre-maintenance performance degradation pattern, the correlation between each physical characteristic and performance degradation is dynamically inferred, and a preliminary weight factor wi is assigned. Based on the influence of each physical characteristic on the change of system performance indicators, the present invention establishes the following dynamic correlation relationship: The rules for establishing correlation relationships are as follows: when heat transfer efficiency decreases primarily, the weight of the unit disturbance recovery time is increased; the weight of the standing wave peak shift is increased; and the weight of the acoustic energy attenuation change rate is moderately increased. When the total pressure drop level increases primarily, the weight of the standard flow pressure drop characteristic is increased; the weight of the standing wave peak shift is moderately increased; and the weight of the acoustic energy attenuation change rate is moderately increased. When the heat transfer rate per unit flow rate decreases significantly but the heat transfer efficiency changes little, the effectiveness of the overall heat transfer channel needs to be comprehensively considered. This is usually accompanied by a slight extension of the disturbance recovery time or a change in acoustic energy attenuation, and the weight of the acoustic energy attenuation change rate should be appropriately increased.

[0035] For each physical feature, the initial weight Calculated based on the decline rate of associated performance indicators:

[0036] in: : the j-th performance indicator decline rate; : The correlation weight coefficient of the i-th physical feature to the j-th performance index. If a feature has a high correlation with a performance index, then If the value is large and the correlation is low, then the value is small or even zero.

[0037] After assigning the preliminary weights, further introduce preset anomaly detection rules to analyze the physical feature data collected before maintenance: set the first, second, and third thresholds for each physical feature. The first threshold T1: the slight anomaly range; the second threshold T2: the medium anomaly range; the third threshold T3: the severe anomaly range. Among them, the threshold settings satisfy that the first threshold is less than the second threshold and the second threshold is less than the third threshold. Let the quantization value of the i-th physical feature detected before maintenance be v i , and the anomaly detection rules are as follows: If T1 ≤ v i < T2, it is considered to be in a slight anomaly state; If T2 ≤ v i < T3, it is considered to be in a medium anomaly state; If v i ≥ T3, it is considered to be in a severe anomaly state; If v i < T1, it is considered not to be abnormal.

[0038] According to the anomaly level, introduce multi-level weighting factors: the first weighting factor k1 > 1 (the slight anomaly amplification coefficient); the second weighting factor k2 > k1 (the medium anomaly amplification coefficient); the third weighting factor k3 > k2 (the severe anomaly amplification coefficient). The final weight after anomaly adjustment The calculation formula is:

[0039] Introducing the dynamic anomaly detection and multi-level weighting mechanism can achieve: distinguishing system degradation situations of different severities; intelligently adjusting the recognition priorities of each physical feature according to the degradation degree; improving the system's sensitivity to weak and hidden degradation trends; avoiding misjudgments or missed judgments caused by a single static threshold; seamlessly connecting the dynamic weight adjustment mechanism with the overall health correction model to form a continuous and consistent intelligent judgment system.

[0040] Based on the final weight and the standardized quantization values of each physical feature collected before maintenance , calculate the health correction coefficient H:

[0041] Apply the health correction coefficient H to correct the original design benchmark parameters and generate the target benchmark parameters after maintenance.

[0042] In a preferred embodiment, based on the completion of anomaly detection and dynamic weighting of a single physical feature, further considering that different physical features in actual systems often have intrinsic correlations, the present invention introduces a feature association weight modulation mechanism to enhance the ability to identify the overall degradation feature group of the system.

[0043] Detect feature combination anomalies: Detect physical feature data collected before maintenance to determine whether two or more physical features simultaneously meet the anomaly detection conditions. The so-called "simultaneous anomaly" means that the detection value of each feature exceeds its corresponding first anomaly threshold T1. Let the feature set that meets the simultaneous anomaly condition be S, and the number of features contained in the set is |S|≥2. When a feature combination anomaly is detected, a unified feature association amplification factor is further introduced on the basis of the existing dynamic weighting factor of a single feature. (in >1), and the weights of all simultaneously abnormal features are jointly amplified. The final adjusted weight calculation formula is:

[0044] That is: first determine according to the abnormality of a single feature Then, for i that satisfies the abnormal feature combination, multiply it by the associated amplification factor get ; When calculating the health correction factor H, Calculated as the final weight.

[0045] The introduction of the feature correlation amplification mechanism can achieve the following: accurately identify systematic problems caused by common abnormalities of multiple physical features under complex degradation patterns; further improve the detection sensitivity of latent and systematic degradation phenomena compared with single feature weighting; avoid missed detection or misjudgment caused by relying solely on single-point features; and strengthen the overall intelligent maintenance effect judgment and health status correction system.

[0046] In a preferred embodiment, the maintenance effect recovery rate calculation module is used to: collect the current operating performance indicators of the industrial energy power system after completing the maintenance operation; and compare the performance indicators after maintenance with the aforementioned corrected target benchmark parameters; thereby quantifying the degree of system recovery achieved by this maintenance operation and forming an "overall recovery rate" indicator.

[0047] After the maintenance operation is completed, the operating performance indicators are re-collected: actual heat transfer efficiency, actual heat transfer per unit flow rate, and actual total pressure drop level. The above module has generated the revised target baseline parameters: revised target heat transfer efficiency, revised target heat transfer per unit flow rate, and revised target total pressure drop level. The recovery rate of each of the three operating performance indicators is calculated. In order to uniformly measure the overall maintenance effect, a comprehensive recovery rate is defined. , as a weighted combination of the three recovery rates above:

[0048] in: + + =1, which is the weighted ratio of the three performance indicators in the comprehensive evaluation. The preferred weight setting is as follows: =0.5, =0.3, =0.2, indicating that heat transfer efficiency is the main evaluation index, followed by unit heat transfer and then pressure drop.

[0049] The calculation of the comprehensive recovery rate is based on the modified target benchmark to avoid the distortion problem of "taking the design value as the benchmark"; the pressure drop index adopts inverse logic, which is more in line with the actual physical meaning; the weight factor can be personalized according to different application scenarios, with good adaptability and scalability; the recovery rate index will serve as the input of the subsequent "hierarchical judgment module" to realize intelligent grading and strategy linkage.

[0050] In a preferred embodiment, the hierarchical maintenance effectiveness assessment module is used to categorize maintenance effectiveness into different levels based on the calculated comprehensive recovery rate and output subsequent operation and maintenance strategy recommendations based on the different levels. The system has preset maintenance effectiveness assessment thresholds: the first maintenance threshold R1 (highest standard); the second maintenance threshold R2 (good standard); and the third maintenance threshold R3 (minimum acceptable standard), where the relationship is set as R1>R2>R3. Based on the comparison of the comprehensive recovery rate Rtotal with the thresholds, the maintenance effectiveness is categorized into the following four levels: The comprehensive recovery rate judgment condition can be formally expressed as:

[0051] Based on the maintenance effect judgment results, the system can automatically output the following operation and maintenance strategy suggestions: The maintenance effect is refined from a simple qualitative judgment to a continuous and quantifiable classification; different levels correspond to clear operation and maintenance decision-making suggestions, improving the intelligence level of the system; the hierarchical judgment is based on the revised dynamic benchmark, which can truly reflect the health status of the system; the preset threshold can be flexibly adjusted according to the actual working conditions of different systems to improve adaptability.

[0052] In a preferred embodiment, the hierarchical level dynamic adjustment mechanism is used to: after completing the maintenance operation and preliminarily determining the maintenance effect level, continue to collect system operation performance indicators in real time within the set dynamic monitoring window period, dynamically correct the maintenance effect judgment threshold based on the system performance fluctuations during the monitoring period, and further correct the final maintenance effect level to ensure that the judgment result can truly reflect the actual recovery level after the maintenance operation.

[0053] During the dynamic monitoring window, continuously collect system performance indicators: heat exchange efficiency, heat exchange per unit flow rate, and total pressure drop level, where t is the time within the monitoring period. For each performance indicator within the monitoring window, calculate its maximum fluctuation range: heat exchange efficiency fluctuation range, heat exchange per unit flow rate fluctuation range, and total pressure drop fluctuation range.

[0054] Define the comprehensive amplitude of performance fluctuations, weighted sum or maximum value: It is preferred to use the maximum value as the most unstable parameter indicator of the system.

[0055] Set a preset stability reference standard, usually setting a tolerance upper limit based on system design requirements. Dynamic adjustment rules are as follows: If the comprehensive amplitude of performance fluctuation is less than or equal to the stability reference standard, the system is judged to be stable and the original maintenance effect judgment level is maintained; If the comprehensive amplitude of performance fluctuation is greater than the stability reference standard, the system is considered to have performance fluctuation risks, and some maintenance effect judgment thresholds are dynamically lowered and the levels are recalibrated.

[0056] Dynamic downgrade method: lower the first maintenance threshold and the second maintenance threshold ; Keep the third maintenance threshold R3 unchanged (to ensure the minimum maintenance standard bottom line).

[0057] Threshold dynamic adjustment formula diagram:

[0058]

[0059] Where: γ is a dynamic adjustment coefficient, and the preferred range is 0.01≤γ≤0.05.

[0060] Combined with short-term operational performance monitoring, the maintenance effect determination results can be dynamically revised; it can identify performance fluctuations caused by residual contamination or local damage after maintenance; further improve the accuracy of the system's intelligent operation and maintenance decisions; and ensure that the maintenance effect determination can provide timely feedback and reflect the system's actual long-term recovery status.

[0061] In summary, the present invention has the following beneficial effects: By constructing an intelligent maintenance effectiveness assessment system based on dynamic benchmark correction and system performance guidance, introducing multi-angle acquisition of operating performance and physical characteristics, dynamic weight inference, anomaly detection and feature-related weight modulation, and a hierarchical maintenance effectiveness determination and dynamic threshold adjustment mechanism, this system achieves precise quantitative assessment, intelligent grading, and dynamic adaptive adjustment of the maintenance effectiveness of industrial energy and power systems. This invention not only overcomes the limitations of traditional maintenance effectiveness determination, which is single, static, and prone to misjudgment, but also optimizes the determination criteria in real time based on the actual system operating status, significantly improving the scientific nature, intelligence, and system reliability of maintenance effectiveness management. It has broad application prospects in areas such as industrial energy-saving operation and maintenance, and industrial energy and power system life management. The above-described embodiments do not constitute a limitation on the scope of protection of this technical solution. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the above-described embodiments shall be included in the scope of protection of this technical solution.

Claims

1. A multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance, characterized by: include: An operating performance index collection module is used to collect operating performance indicators of the industrial energy power system before and after maintenance operations, wherein the operating performance indicators include heat exchange efficiency, heat exchange per unit flow rate, and total pressure drop level; Characteristic data acquisition module, used to collect physical characteristic data of industrial energy power systems before maintenance operations, including unit disturbance recovery time, pressure drop under standard flow, standing wave peak frequency offset and acoustic wave energy attenuation change rate; The weight inference and correction benchmark generation module is used to dynamically determine the final weight of each physical characteristic data based on the pre-maintenance operating performance indicators, calculate the health correction coefficient based on the final weight and physical characteristic data, and modify the original design benchmark parameters accordingly to generate the target benchmark parameters after maintenance; The maintenance effect recovery rate calculation module is used to compare the operating performance indicators collected after maintenance with the revised target benchmark parameters to calculate the comprehensive recovery rate; The hierarchical maintenance effect determination module is used to divide the maintenance effect into different levels according to the comprehensive recovery rate and output the corresponding subsequent operation and maintenance strategy.

2. The multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance according to claim 1 is characterized in that: The feature data acquisition module specifically includes: The disturbance response acquisition submodule is used to apply a flow disturbance of a preset amplitude to the fluid path of the industrial energy power system before maintenance, and collect the response curve of the flow velocity or flow rate changing with time after the disturbance is applied in real time, and extract the unit disturbance recovery time constant based on curve fitting; The pressure drop acquisition submodule is used to collect the inlet and outlet pressure values of the industrial energy power system under standard flow conditions before maintenance, and calculate the pressure drop level under standard flow; The acoustic standing wave acquisition submodule is used to apply an acoustic wave signal within a preset frequency range inside the maintenance forward system, collect the reflected acoustic wave signal, and extract the strongest reflection peak frequency; The acoustic energy attenuation acquisition submodule is used to measure the ratio of the acoustic signal input power to the reflected signal received power, and calculate the acoustic energy attenuation change rate based on the change in the ratio.

3. The multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance according to claim 2 is characterized in that: The weight inference and correction benchmark generation module specifically includes: a macro-performance change correlation unit for inferring the contribution of each collected physical characteristic to the performance degradation of the industrial energy power system based on the decline of the operating performance indicators before maintenance, and assigning a preliminary weight factor to each physical characteristic accordingly, wherein the operating performance indicators include heat transfer efficiency, heat transfer per unit flow rate, and total pressure drop; A dynamic weight determination unit is used to determine the final weight based on the preliminary weight factors of each physical feature, select preset anomaly detection rules to analyze the physical feature data collected before maintenance, and increase the final weight of a physical feature accordingly when the physical feature data meets the anomaly condition; The benchmark correction generating unit is used to calculate the health correction coefficient based on the final weight of each physical feature and the physical feature data, and apply the health correction coefficient to the original design benchmark parameter to generate the target benchmark parameter after maintenance.

4. The multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance according to claim 3 is characterized in that: The method of selecting a preset anomaly detection rule to analyze the physical feature data collected before maintenance and correspondingly increasing the final weight of a physical feature when a physical feature data meets an anomaly condition specifically includes: Multi-level abnormality judgment thresholds are set for unit disturbance recovery time, pressure drop under standard flow, standing wave peak frequency offset and acoustic wave energy attenuation, including the first threshold, second threshold and third threshold; When the detection data of a certain physical feature is between the first threshold and the second threshold, the final weight of the physical feature is adjusted according to the first weighting factor k1>1; When the physical feature data is between the second threshold and the third threshold, the final weight is adjusted according to the second weighting factor k2>k1; When the physical feature data exceeds the third threshold, the final weight is adjusted according to the third weighting factor k3>k2.

5. The multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance according to claim 1 is characterized in that: The maintenance effect is divided into different levels according to the comprehensive recovery rate, and the corresponding subsequent operation and maintenance strategies are output, specifically including: Set a first maintenance threshold R1, a second maintenance threshold R2, and a third maintenance threshold R3, where R1>R2>R3; When the comprehensive recovery rate is greater than or equal to the first maintenance threshold, it is determined to be at the full recovery level, and an operation and maintenance strategy for maintaining normal operation is output; When the comprehensive recovery rate is less than the first maintenance threshold but greater than or equal to the second maintenance threshold, it is determined to be a good recovery level, and an operation and maintenance strategy of local periodic flushing and operation status monitoring is output; When the comprehensive recovery rate is less than the second maintenance threshold but greater than or equal to the third maintenance threshold, it is determined to be a local residual level, and an operation and maintenance strategy of strengthening local maintenance or shortening the maintenance cycle is output; When the comprehensive recovery rate is less than the third maintenance threshold, it is determined to be a maintenance failure level, and an operation and maintenance strategy of re-maintenance or replacement of relevant components is output.

6. The multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance according to claim 5 is characterized in that: The maintenance effect recovery rate calculation module further includes a dynamic monitoring unit, which is used to set a preset monitoring window period after the maintenance operation is completed, continuously collect the changing trend of the operating performance indicators of the industrial energy power system, and correct the comprehensive recovery rate based on the change amplitude of the operating performance indicators within the monitoring window to correct the maintenance effect judgment result.

7. The multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance according to claim 6 is characterized in that: The hierarchical maintenance effect determination module dynamically fine-tunes the first maintenance threshold, the second maintenance threshold, and the third maintenance threshold based on the change range of the operating performance indicator within the post-maintenance dynamic monitoring window after the maintenance operation, specifically including: When the fluctuation range of the operating performance indicator within the monitoring window is lower than the preset stability standard, the original set threshold is maintained; When the fluctuation range of the operating performance indicator within the monitoring window exceeds the preset standard, the set values of the first maintenance threshold and the second maintenance threshold are respectively lowered.

8. The multi-dimensional evaluation and intelligent determination system for industrial energy and power system maintenance according to claim 3 is characterized in that: The dynamic weight determination unit further includes a feature association weight modulation unit, which is used to detect the correlation between each physical feature data during the dynamic weight determination process, and when two or more physical feature data simultaneously meet the anomaly detection rules, jointly amplify the final weight of the corresponding physical features.

Images