Method and system for evaluating health state of key equipment of hydropower station

Abstract

The invention relates to the technical field of hydropower station equipment monitoring, in particular to a health state assessment method and system for key equipment of a hydropower station. The method comprises the following steps: establishing a plurality of monitoring units according to characteristics of hydropower station equipment, establishing a corresponding monitoring model for each monitoring unit, and obtaining multi-source heterogeneous monitoring data according to the monitoring models; preprocessing the multi-source heterogeneous monitoring data, extracting a feature index set reflecting the equipment operation state of the monitoring unit, and calculating according to the feature index set to obtain a real-time health index value; and comparing the health index value of each monitoring unit with a preset interval, and outputting an evaluation conclusion according to a comparison result. Accurate sensing and graded early warning of the health state of key equipment such as a water turbine and a generator are achieved, fault positioning and maintenance suggestion generation are supported, the operation safety and operation and maintenance efficiency of the equipment are improved, and effective support is provided for intelligent maintenance of a hydropower station.

Description

Technical Field

[0001] This invention relates to the field of hydropower station equipment monitoring technology, specifically to a method and system for assessing the health status of key equipment in a hydropower station. Background Technology

[0002] Traditional methods for assessing the health status of critical equipment in hydropower stations have limitations in data acquisition and utilization. Existing assessments often rely on single types of data, such as collecting only physical parameters like vibration and temperature, or focusing only on certain status signals of equipment operation while ignoring operational command data issued by the control system. They fail to effectively integrate multi-source heterogeneous data and thus cannot comprehensively reflect the equipment status.

[0003] Existing assessment methods have shortcomings in accuracy and timeliness of early warning. Traditional assessment methods rely heavily on periodic maintenance and manual judgment, resulting in low efficiency, slow response, and susceptibility to missed detections. With the development of sensor technology and data analysis methods, health status assessment methods based on multi-source data fusion have gradually become a research hotspot. However, existing methods still have shortcomings in feature extraction, weight allocation, and operating condition identification, necessitating a more comprehensive, accurate, and automated assessment solution. Summary of the Invention

[0004] The purpose of this application is to address the problems of insufficient data utilization, low assessment accuracy, and untimely early warning in existing health assessment methods for key equipment in hydropower stations, and to provide a health status assessment method and system based on multi-source heterogeneous data fusion, so as to achieve accurate perception and hierarchical early warning of equipment status and provide support for intelligent maintenance.

[0005] In some embodiments of this application, a method for assessing the health status of key equipment in a hydropower station includes the following steps:

[0006] Multiple monitoring units are established based on the characteristics of hydropower station equipment, and corresponding monitoring models are established for each monitoring unit. Multi-source heterogeneous monitoring data are obtained based on the monitoring models.

[0007] The multi-source heterogeneous monitoring data is preprocessed, and a set of feature indicators reflecting the operating status of the monitoring unit equipment is extracted. The real-time health index value is calculated based on the set of feature indicators.

[0008] The health index value of each monitoring unit is compared with a preset range, and an evaluation conclusion is output based on the comparison results.

[0009] In some embodiments of this application, multiple monitoring units are established based on the characteristics of the hydropower station equipment, including:

[0010] Construct a set of monitoring units U, U = {u1, u2, ..., u...} n}, where n is the number of monitoring units, ui This is the i-th monitoring unit;

[0011] Construct a set of health evaluation characteristic indicators F corresponding to the set of monitoring units U, where F = {f1, f2, ..., f m}, where m is the number of feature indicators, f j Let j be the j-th feature index;

[0012] Each feature index f is generated based on historical operational data. j Reference value r in a healthy state j And construct a health value vector set R for the monitoring unit set U, R = (r1, r2, ..., r m ).

[0013] In some embodiments of this application, a corresponding monitoring model is established for each monitoring unit, including:

[0014] For each monitoring unit and its corresponding key equipment, multiple sensor groups S1, S2, ..., S are deployed based on their fault mechanisms and operating characteristics. k , of which S k This is the Kth sensor group;

[0015] Each sensor group corresponds to a key part of the critical equipment and is used to collect multi-dimensional operational data of that part.

[0016] In some embodiments of this application, the real-time health index value is calculated based on the set of feature indicators, including:

[0017] The formula for calculating the real-time health index value is as follows:

[0018]

[0019] Where, d j w is the normalized deviation value of the j-th feature index. j is the weight coefficient corresponding to the j-th feature index.

[0020] In some embodiments of this application, each of the sensor groups S k It includes several different types of sensors that monitor the core physical quantity of the same key part at preset time intervals. The core physical quantity is determined according to the type of key equipment in the hydropower station.

[0021] Calculate the difference between the collected values ​​of each sensor and those of the same type of sensor in the same sensor group;

[0022] When the real-time acquired value of each sensor is not within the corresponding preset normal range, or the difference between the acquired values ​​is not within the preset normal acquisition value difference range, the sensor data is determined to be abnormal.

[0023] In some embodiments of this application, the output evaluation conclusions include:

[0024] Pre-set the first-level threshold range, the second-level range, and the third-level range;

[0025] When the real-time health index is within the preset first-level range, the assessment result is that the corresponding key part is in a healthy state;

[0026] When the real-time health index is in the preset secondary range, the assessment result is that the corresponding key part is in a state of attention. The index difference between the preset secondary range and the real-time health index is calculated, the preset time interval is corrected according to the index difference, and the corresponding key equipment is monitored according to the corrected preset time interval.

[0027] When the real-time health index is in the preset three-level range, the evaluation result is that the corresponding key part is in an abnormal state. The information of the top N feature indicators sorted by contribution and the corresponding maintenance suggestions or fault prediction information are extracted, where N is an integer greater than or equal to 1.

[0028] Based on historical fault correlation data of the aforementioned characteristic indicators, targeted maintenance suggestions or fault prediction information are provided.

[0029] The preset first-level interval corresponds to the healthy state, the preset second-level interval corresponds to the attention state, and the preset third-level interval corresponds to the abnormal state.

[0030] In some embodiments of this application, the preset time interval is corrected based on the exponential difference, including:

[0031] The first preset index difference range, the second preset index difference range, the third preset index difference range, and the fourth preset index difference range are preset.

[0032] When the exponential difference is within the first preset exponential difference range, the first preset correction coefficient a1 is selected to correct the preset time interval T, and the corrected preset time interval is T*a1.

[0033] When the exponential difference is within the second preset exponential difference range, the second preset correction coefficient a2 is selected to correct the preset time interval T. The corrected preset time interval is T*a2.

[0034] When the exponential difference is within the third preset exponential difference range, the third preset correction coefficient a3 is selected to correct the preset time interval T. The corrected preset time interval is T*a3.

[0035] When the exponential difference is within the fourth preset exponential difference range, the fourth preset correction coefficient a4 is selected to correct the preset time interval T, and the corrected preset time interval is T*a4.

[0036] Where 0 < a1 < a2 < a3 < a4 < 1.

[0037] In some embodiments of this application, the method for extracting the top N feature indicators with the greatest contribution includes:

[0038] Calculate each characteristic index f j Contribution to the Health Index (HI);

[0039] The formula for calculating the contribution is as follows:

[0040] C j =w j ·|d j |;

[0041] Among them, C j The contribution of the j-th feature indicator;

[0042] Sort the feature indicators from largest to smallest according to their contribution, and select the top N feature indicators;

[0043] Based on the data status of the sensor group to which the key equipment belongs, output the information of the first N characteristic indicators and their corresponding equipment location information.

[0044] Some embodiments of this application also include:

[0045] Based on the first N characteristic indicators and their corresponding sensor group data, a device health diagnosis report is generated;

[0046] The report includes the location of abnormal parts, health trend prediction, and maintenance recommendations based on historical maintenance records;

[0047] The health status trend prediction is based on the historical changes of the first N feature indicators and the fault development pattern for fitting and prediction.

[0048] The maintenance recommendations include the suggested maintenance content, the urgency of the maintenance, and the recommended maintenance time.

[0049] In some embodiments of this application, a health status assessment system for key equipment in a hydropower station includes:

[0050] The data acquisition module establishes multiple monitoring units based on the characteristics of hydropower station equipment, builds corresponding monitoring models for the multiple monitoring units, and acquires multi-source heterogeneous monitoring data based on the monitoring models.

[0051] The data processing module preprocesses the multi-source heterogeneous monitoring data and extracts a set of feature indicators that reflect the operating status of the monitoring unit equipment. Based on the set of feature indicators, the real-time health index value is calculated.

[0052] The health assessment module compares the health index value of each monitoring unit with a preset threshold range and outputs an assessment conclusion based on the comparison results.

[0053] This invention effectively overcomes the limitations of insufficient data utilization in traditional assessment methods by employing a multi-monitoring unit architecture and multi-source heterogeneous data fusion technology. Targeting the fault mechanisms and operating characteristics of different key equipment such as turbines and generators, it deploys multiple types of sensor groups, including temperature, vibration, and current sensors, while simultaneously collecting control system operation command data. After standardized preprocessing, it extracts multi-dimensional features such as time domain, frequency domain, and operating condition labels. Then, through a combined weighting algorithm that integrates subjective experience and objective data, it calculates a health index, comprehensively and accurately reflecting the true operating status of the equipment. This avoids the problems of missed detections and misjudgments caused by traditional single-parameter monitoring or manual judgment, significantly improving the accuracy and reliability of health assessment.

[0054] Meanwhile, this invention possesses highly efficient early warning and maintenance support capabilities. By comparing health indices with preset thresholds, it achieves a tiered early warning system (health-attention-early warning). When equipment is in an abnormal range, it can quickly identify the top N most contributing abnormal characteristic indicators and, combined with historical fault correlation data, provide targeted maintenance suggestions and fault development trend predictions. It can also predict the remaining lifespan of equipment based on historical health index data. This feature effectively addresses the pain points of low efficiency and slow response in traditional periodic maintenance, helping hydropower stations to develop intelligent maintenance plans in advance, reducing unplanned downtime, lowering maintenance costs, and providing strong technical support for the safe and stable operation of critical equipment. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating a method for assessing the health status of key equipment in a hydropower station, as described in a preferred embodiment of this application. Detailed Implementation

[0056] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate this application, but are not intended to limit the scope of this application.

[0057] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0058] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0059] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0060] like Figure 1 As shown in the figure, a method for assessing the health status of key equipment in a hydropower station according to an embodiment of this application includes the following steps:

[0061] Multiple monitoring units are established based on the characteristics of hydropower station equipment, and corresponding monitoring models are established for each monitoring unit. Multi-source heterogeneous monitoring data are obtained based on the monitoring models.

[0062] The multi-source heterogeneous monitoring data is preprocessed, and a set of feature indicators reflecting the operating status of the monitoring unit equipment is extracted. The real-time health index value is calculated based on the set of feature indicators.

[0063] The health index value of each monitoring unit is compared with a preset range, and an evaluation conclusion is output based on the comparison results.

[0064] In this embodiment, the key equipment of the hydropower station mainly includes turbines, generators, main transformers, governors, and excitation systems. The physical structures, operating mechanisms, and failure modes of different equipment vary significantly. For example, the core failures of turbines are bearing wear and blade cavitation, while the core failures of generators are stator winding overheating and rotor imbalance. Therefore, monitoring units need to be divided according to equipment type and functional modules.

[0065] In this embodiment, the multi-source heterogeneous data preprocessing and feature index set extraction include: data purification and extraction of core information. Preprocessing aims to eliminate noise and outliers in the original data, while feature index set extraction aims to transform massive amounts of data into quantifiable core indicators that reflect the equipment status, which is the core guarantee for evaluation accuracy. The data preprocessing steps include: noise reduction, missing value imputation, and normalization.

[0066] In this embodiment, the feature index set extraction includes three types of indicators: time-domain features, frequency-domain features, and operating condition-related features. For example, the feature index set of the turbine monitoring unit U1 includes: time-domain features: the average bearing vibration amplitude reflects the overall vibration level, the vibration peak value reflects the impact fault, and the temperature standard deviation reflects the temperature stability; frequency-domain features: the first harmonic amplitude of the vibration signal reflects the rotor imbalance fault, the second harmonic amplitude reflects the misalignment fault, and the 0.5 harmonic amplitude reflects the oil film eddy fault; operating condition-related features: the unit speed reflects the operating load, and the guide vane opening reflects the water flow impact intensity, ensuring that the indicators can be adapted to the state assessment under different operating conditions.

[0067] In this embodiment, the preset interval setting includes: setting the preset scale to [0, 100], and dividing HI into 3 intervals, corresponding to different health states: Health interval HI∈[0, 30]: the deviation of each indicator of the equipment is small, there is no risk of failure, and the evaluation conclusion is "the equipment is healthy and does not require maintenance"; Attention interval HI∈(30, 60]: some indicators show slight deviations, such as the vibration 1st harmonic frequency being slightly higher than the normal range, there is a potential risk of failure, and the evaluation conclusion is "the equipment status needs attention, and it is recommended to increase the monitoring frequency, such as changing from once a day to once every 2 hours"; Warning interval HI∈(60, 100]: multiple indicators show significant deviations.

[0068] In some embodiments of this application, multiple monitoring units are established based on the characteristics of the hydropower station equipment, including:

[0069] Construct a set of monitoring units U, U = {u1, u2, ..., u...} n}, where n is the number of monitoring units, u i This is the i-th monitoring unit;

[0070] Construct a set of health evaluation characteristic indicators F corresponding to the set of monitoring units U, where F = {f1, f2, ..., f m}, where m is the number of feature indicators, f j Let j be the j-th feature index;

[0071] Each feature index f is generated based on historical operational data. j Reference value r in a healthy state j And construct a health value vector set R for the monitoring unit set U, R = (r1, r2, ..., r m ).

[0072] Establish corresponding monitoring models for each monitoring unit, including:

[0073] For each monitoring unit and its corresponding key equipment, multiple sensor groups S1, S2, ..., S are deployed based on their fault mechanisms and operating characteristics. k , of which S kThis is the Kth sensor group;

[0074] Each sensor group corresponds to a key part of the critical equipment and is used to collect multi-dimensional operational data of that part.

[0075] In some embodiments of this application, the real-time health index value is calculated based on the set of feature indicators, including:

[0076] The formula for calculating the real-time health index value is as follows:

[0077]

[0078] Where, d j w is the normalized deviation value of the j-th feature index. j The weight coefficient corresponding to the j-th feature index;

[0079] In this embodiment, the normalized deviation value of the j-th feature index is obtained by comparing the current measured value with its reference value r. j The difference is obtained by mapping the proportion of the difference to the preset deviation range onto the preset scale. The preset deviation range refers to the maximum allowable deviation range of this characteristic indicator during normal operation, which is determined based on equipment design standards or historical health data.

[0080] In this embodiment, the weight coefficient corresponding to the j-th feature index is predetermined based on the equipment's historical operating data and fault records, and satisfies the following conditions:

[0081] In this embodiment, determining the preset deviation range includes: the preset deviation range is d. j The core benchmark for calculation is defined as the indicator derived from the health reference value r. j To the fault threshold r 故障 The interval, i.e., the deviation range S j =r 故障j -r j Fault threshold value r 故障j Source: Prioritize equipment design standards, such as fault alarm values ​​provided by the manufacturer. If none are available, use historical fault data, specifically the measured value of the indicator when the equipment malfunctions, such as f for turbine U1. 12 The amplitude of the first harmonic of vibration, r j =0.15mm / s (health reference value), manufacturer-set fault alarm value r 故障j =0.45mm / s, then the deviation range S j =0.45-0.15=0.3mm / s.

[0082] In this embodiment, d j Specific calculation steps: 1. Calculate the actual deviation: Δx j =x实测j -r j x 实测j Let Δx be the measured value of this indicator at the current moment. j If the measured value is less than or equal to 0 and the reference value is less than or equal to 0, then d is considered to have no deviation. j =0; if Δx j >0, the measured value exceeds the reference value, indicating a deviation; proceed to the next step; 2. Calculate the deviation ratio: If α j If the measured value is greater than or equal to 1, and the deviation reaches the fault threshold, then d j =100; if 0 < α j If the measured value is less than 1 and falls between the reference value and the fault threshold value, then d j =α j *100. Using the f of turbine U1 12 (Vibration amplitude at 1st harmonic) as an example: Case 1: x 实测j =0.12mm / s (≤r j =0.15mm / s), then Δx j = -0.03mm / s, d j =0; Case 2: x 实测j =0.3mm / s (between r) j With r 故障j (between), Δx j =0.15mm / s, α j =0.15 / 0.3 = 0.5, d j =0.5 * 100 = 50; Case 3: x 实测j =0.5mm / s (≥r 故障j =0.45mm / s), α j =0.35 / 0.3≈1.17≥1,d j =100.

[0083] In some embodiments of this application, each of the sensor groups S k It includes several different types of sensors that monitor the core physical quantity of the same key part at preset time intervals. The core physical quantity is determined according to the type of key equipment in the hydropower station.

[0084] Calculate the difference between the collected values ​​of each sensor and those of the same type of sensor in the same sensor group;

[0085] When the real-time acquired value of each sensor is not within the corresponding preset normal range, or the difference between the acquired values ​​is not within the preset normal acquisition value difference range, the sensor data is determined to be abnormal.

[0086] In this embodiment, there are at least three sensors of several different types.

[0087] In this embodiment, the sensor arrangement includes: the physical structure of the equipment determines the position and function of the components. Sensors need to be placed on core functional components, avoiding placement in non-critical parts, such as the outer casing. For example, in a water turbine, the physical structure includes the runner, main shaft, thrust bearing, guide bearing, guide vanes, and tailrace. Sensors need to be placed on core components such as the runner (core physical quantity reflecting energy conversion), bearings (core physical quantity reflecting support status), and guide vanes (core physical quantity reflecting water flow control). Based on the fault-prone areas, key monitoring is conducted. Different equipment has different fault-prone areas, and the number of sensors needs to be increased in these areas. For example, in a water turbine, the fault-prone areas are: thrust bearing (wear failure accounts for 30%), runner blades (cavitation failure accounts for 25%), and guide vanes (jamming failure accounts for 20%). Therefore, two temperature sensors and one vibration sensor are placed on the thrust bearing, two vibration sensors are placed on the runner blades, and one displacement sensor and one position feedback sensor are placed on the guide vanes. Based on operating characteristics and to meet monitoring needs: The operating characteristics of the equipment determine the type and installation method of the sensor. For example, rotating parts, such as spindles and rotors, require non-contact sensors, such as eddy current vibration sensors, to avoid failures caused by wear of contact sensors; high-voltage parts, such as stator windings and main transformer windings, require insulated sensors, such as fiber optic temperature sensors, to avoid short circuits caused by sensor conductivity.

[0088] In this embodiment, the sensor group includes five core types of sensors: temperature, vibration, pressure, current, and voltage. The technical parameters, installation locations, and monitoring targets of each type are clearly defined.

[0089] In this embodiment, determining that the sensor data is abnormal includes:

[0090] If only one sensor data is abnormal, and the data of other sensors in the same group are all within the preset normal range, it is determined to be a single sensor failure. The abnormal data is discarded and replaced by the average value of other sensors in the same group.

[0091] If ≥2 sensor data in the same group are abnormal and the abnormality type is the same, it is determined that the equipment is in a critical part, triggering a secondary data verification of that part, and the verification result is included in the calculation of the feature index set.

[0092] In some embodiments of this application, the output evaluation conclusions include:

[0093] Pre-set the first-level threshold range, the second-level range, and the third-level range;

[0094] When the real-time health index is within the preset first-level range, the assessment result is that the corresponding key part is in a healthy state;

[0095] When the real-time health index is in the preset secondary range, the assessment result is that the corresponding key part is in a state of attention. The index difference between the preset secondary range and the real-time health index is calculated, the preset time interval is corrected according to the index difference, and the corresponding key equipment is monitored according to the corrected preset time interval.

[0096] When the real-time health index is in the preset three-level range, the evaluation result is that the corresponding key part is in an abnormal state. The information of the top N feature indicators sorted by contribution and the corresponding maintenance suggestions or fault prediction information are extracted, where N is an integer greater than or equal to 1.

[0097] Based on historical fault correlation data of the aforementioned characteristic indicators, targeted maintenance suggestions or fault prediction information are provided.

[0098] The preset first-level interval corresponds to the healthy state, the preset second-level interval corresponds to the attention state, and the preset third-level interval corresponds to the abnormal state.

[0099] In this embodiment, each interval can be dynamically adjusted based on historical operating parameters.

[0100] In this embodiment, the information of the first N feature indicators refers to the full-dimensional correlation data supporting fault location and maintenance suggestion generation, including the following three core contents: 1. Basic information of feature indicators, including: indicator identifier: indicator name (e.g., "vibration amplitude of turbine thrust bearing at 1st harmonic frequency"), corresponding monitoring unit number (e.g., U1), and equipment location; numerical information: normalized deviation value d j Weighting coefficient w j Contribution C j 1. Status Information: Whether the indicator currently exceeds the preset normal range, and the maximum / minimum deviation value in the past 24 hours. 2. Sensor Group Association Information: Sensor group number and sensor type for this indicator; Sensor Status Information: Difference in collected values ​​among sensors of the same type in this sensor group, whether there are data anomalies (such as single sensor failure / equipment part anomaly), and sensor calibration time. 3. Historical Fault Association Information: Historical fault records: Fault types, number of faults, and average fault development cycle caused by this characteristic indicator in the past year; Fault Association Probability: The combined anomaly probability of this indicator with other characteristic indicators (e.g., when "vibration frequency exceeds the standard + temperature exceeds the standard", the probability of thrust bearing wear failure is 92%).

[0101] In some embodiments of this application, the preset time interval is corrected based on the exponential difference, including:

[0102] The first preset index difference range, the second preset index difference range, the third preset index difference range, and the fourth preset index difference range are preset.

[0103] When the exponential difference is within the first preset exponential difference range, the first preset correction coefficient a1 is selected to correct the preset time interval T, and the corrected preset time interval is T*a1.

[0104] When the exponential difference is within the second preset exponential difference range, the second preset correction coefficient a2 is selected to correct the preset time interval T. The corrected preset time interval is T*a2.

[0105] When the exponential difference is within the third preset exponential difference range, the third preset correction coefficient a3 is selected to correct the preset time interval T. The corrected preset time interval is T*a3.

[0106] When the exponential difference is within the fourth preset exponential difference range, the fourth preset correction coefficient a4 is selected to correct the preset time interval T, and the corrected preset time interval is T*a4.

[0107] Where 0 < a1 < a2 < a3 < a4 < 1.

[0108] In this embodiment, the first preset index difference interval < the second preset index difference interval < the third preset index difference interval < the fourth preset index difference interval.

[0109] In this embodiment, the index difference refers to the difference between the maximum index of the preset secondary interval and the real-time health index. The smaller the index difference, the more likely the key parts are to have abnormalities. Therefore, a smaller correction coefficient should be selected to correct the preset time interval, shorten the preset time interval, increase the monitoring frequency, and thus promptly detect the real-time health index and adjust the abnormal state of the key components, thereby reducing the operation and maintenance costs of the key equipment.

[0110] In some embodiments of this application, the method for extracting the top N feature indicators with the greatest contribution includes:

[0111] Calculate each characteristic index f j Contribution to the Health Index (HI);

[0112] The formula for calculating the contribution is as follows:

[0113] C j =w j ·|d j |;

[0114] Among them, C j The contribution of the j-th feature indicator;

[0115] Sort the feature indicators from largest to smallest according to their contribution, and select the top N feature indicators;

[0116] Based on the data status of the sensor group to which the key equipment belongs, output the information of the first N characteristic indicators and their corresponding equipment location information.

[0117] Based on the first N characteristic indicators and their corresponding sensor group data, a device health diagnosis report is generated;

[0118] The report includes the location of abnormal parts, health trend prediction, and maintenance recommendations based on historical maintenance records;

[0119] The health status trend prediction is based on the historical changes of the first N feature indicators and the fault development pattern for fitting and prediction.

[0120] The maintenance recommendations include the suggested maintenance content, the urgency of the maintenance, and the recommended maintenance time.

[0121] In this embodiment, the method for extracting the top N feature indicators with the largest contribution includes: locating the core anomaly source, the feature indicator with the largest weighted contribution value to HI has the largest weighted contribution value, where the weighted contribution value = weight coefficient w. j ×Normalized bias value d j By sorting by this value, the core indicators that cause abnormal equipment status can be quickly located.

[0122] In this embodiment, the maintenance recommendations include: Step 1: Matching historical faults: Matching the abnormal deviation range of the current top N indicators with the deviation range in the database to determine the most likely fault type; Step 2: Screening the optimal repair plan: Among the matched fault types, selecting the repair plan with the shortest repair time, the best post-repair indicator recovery, and the longest recurrence cycle; Step 3: Determining the maintenance timing: Based on the severity of the fault (setting the HI range 0-100) and the equipment operation plan, the following maintenance timing is recommended: Attention range (HI is 30-60): It is recommended to use the next planned shutdown for maintenance (e.g., within 1 month); Warning range (HI is 60-100): It is recommended to "emergency shutdown maintenance within 72 hours".

[0123] In this embodiment, the fault prediction information includes: predicting the fault development trend; the fault prediction information is based on the current indicator anomaly level and historical fault conditions, predicting the time it takes for the fault to develop from the current state to a severe fault, providing a time reference for the operation and maintenance plan; and using linear fitting and historical similarity comparison: Step 1: Extract the deviation value change curve of the current indicator over the past 7 days; Step 2: In the historical fault database, filter fault cases with the same indicator and similar change trends; Step 3: Predict d by linearly fitting the current curve. j The time it takes for the value to rise to 90, which is the critical fault threshold.

[0124] In some embodiments of this application, a health status assessment system for key equipment in a hydropower station is also included:

[0125] The data acquisition module establishes multiple monitoring units based on the characteristics of hydropower station equipment, builds corresponding monitoring models for the multiple monitoring units, and acquires multi-source heterogeneous monitoring data based on the monitoring models.

[0126] The data processing module preprocesses the multi-source heterogeneous monitoring data and extracts a set of feature indicators that reflect the operating status of the monitoring unit equipment. Based on the set of feature indicators, the real-time health index value is calculated.

[0127] The health assessment module compares the health index value of each monitoring unit with a preset threshold range and outputs an assessment conclusion based on the comparison results.

[0128] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of this application, and these improvements and substitutions should also be considered within the scope of protection of this application.

Claims

1. A method for assessing the health status of key equipment in a hydropower station, characterized in that, Includes the following steps: Multiple monitoring units are established based on the characteristics of hydropower station equipment, and corresponding monitoring models are established for each monitoring unit. Multi-source heterogeneous monitoring data are obtained based on the monitoring models. The multi-source heterogeneous monitoring data is preprocessed, and a set of feature indicators reflecting the operating status of the monitoring unit equipment is extracted. The real-time health index value is calculated based on the set of feature indicators. The health index value of each monitoring unit is compared with a preset range, and an evaluation conclusion is output based on the comparison results.

2. The health status assessment method for key equipment in a hydropower station as described in claim 1, characterized in that, Multiple monitoring units were established based on the characteristics of the hydropower station equipment, including: Construct a set of monitoring units U, U = {u1, u2, ..., u...} n }, where n is the number of monitoring units, u i This is the i-th monitoring unit; Construct a set of health evaluation characteristic indicators F corresponding to the set of monitoring units U, where F = {f1, f2, ..., f m }, where m is the number of feature indicators, f j Let j be the j-th feature index; Each feature index f is generated based on historical operational data. j Reference value r in a healthy state j And construct a health value vector set R for the monitoring unit set U, R = (r1, r2, ..., r m ).

3. The health status assessment method for key equipment in a hydropower station according to claim 2, characterized in that, Establish corresponding monitoring models for each monitoring unit, including: For each monitoring unit and its corresponding key equipment, multiple sensor groups S1, S2, ..., S are deployed based on their fault mechanisms and operating characteristics. k , among which, S k This is the Kth sensor group; Each sensor group corresponds to a key part of the critical equipment and is used to collect multi-dimensional operational data of that key part.

4. The health status assessment method for key equipment in a hydropower station as described in claim 3, characterized in that, The real-time health index value is calculated based on the aforementioned feature index set, including: The formula for calculating the real-time health index value is as follows: Where, d j w is the normalized deviation value of the j-th feature index. j is the weight coefficient corresponding to the j-th feature index.

5. The health status assessment method for key equipment in a hydropower station as described in claim 4, characterized in that, Each of the sensor groups S k It contains several different types of sensors that monitor the core physical quantities of the same key part at preset time intervals to obtain real-time acquisition values; Calculate the difference between the collected values ​​of each sensor and those of the same type of sensor in the same sensor group; When the real-time acquired value of each sensor is not within the corresponding preset normal range, or the difference between the acquired values ​​is not within the preset normal acquisition value difference range, the sensor data is determined to be abnormal.

6. The health status assessment method for key equipment in a hydropower station as described in claim 5, characterized in that, Output evaluation conclusions, including: Pre-set the first-level interval, the second-level interval, and the third-level interval; When the real-time health index is within the preset first-level range, the assessment result indicates that the corresponding key parts are in a healthy state; When the real-time health index is in the preset secondary range, the assessment result is that the corresponding key part is in a state of attention. The index difference between the preset secondary range and the real-time health index is calculated, the preset time interval is corrected according to the index difference, and the corresponding key equipment is monitored according to the corrected preset time interval. When the real-time health index is in the preset three-level range, the evaluation result is that the corresponding key part is in an abnormal state. The information of the top N feature indicators sorted by contribution and the corresponding maintenance suggestions or fault prediction information are extracted, where N is an integer greater than or equal to 1. Based on historical fault correlation data of the aforementioned characteristic indicators, targeted maintenance suggestions or fault prediction information are provided. The preset first-level interval corresponds to the healthy state, the preset second-level interval corresponds to the attention state, and the preset third-level interval corresponds to the abnormal state.

7. The health status assessment method for key equipment in a hydropower station as described in claim 6, characterized in that, The preset time interval is corrected based on the exponential difference, including: The first preset index difference range, the second preset index difference range, the third preset index difference range, and the fourth preset index difference range are preset. When the exponential difference is within the first preset exponential difference range, the first preset correction coefficient a1 is selected to correct the preset time interval T, and the corrected preset time interval is T*a1. When the exponential difference is within the second preset exponential difference range, the second preset correction coefficient a2 is selected to correct the preset time interval T. The corrected preset time interval is T*a2. When the exponential difference is within the third preset exponential difference range, the third preset correction coefficient a3 is selected to correct the preset time interval T. The corrected preset time interval is T*a3. When the exponential difference is within the fourth preset exponential difference range, the fourth preset correction coefficient a4 is selected to correct the preset time interval T. The corrected preset time interval is T*a4. Where 0 < a1 < a2 < a3 < a4 < 1.

8. The health status assessment method for key equipment in a hydropower station as described in claim 7, characterized in that, The extraction of information from the top N feature indicators ranked by contribution includes: Calculate each characteristic index f j Contribution to the Health Index (HI); The formula for calculating the contribution is as follows: C j =w j ·|d j |; Among them, C j The contribution of the j-th feature indicator; Sort the feature indicators from largest to smallest according to their contribution, and select the top N feature indicators; Based on the data status of the sensor group to which the key equipment belongs, output the information of the first N characteristic indicators and their corresponding equipment location information.

9. The health status assessment method for key equipment in a hydropower station as described in claim 8, characterized in that, Also includes: Based on the first N characteristic indicators and their corresponding sensor group data, a device health diagnosis report is generated; The report includes the location of abnormal parts, health trend prediction, and maintenance recommendations based on historical maintenance records; The health status trend prediction is based on the historical changes of the first N feature indicators and the fault development pattern for fitting and prediction. The maintenance recommendations include the suggested maintenance content, the urgency of the maintenance, and the recommended maintenance time.

10. A health status assessment system for key equipment in a hydropower station, employing the health status assessment method for key equipment in a hydropower station as described in any one of claims 1-9, characterized in that, include: The data acquisition module establishes multiple monitoring units based on the characteristics of hydropower station equipment, builds corresponding monitoring models for the multiple monitoring units, and acquires multi-source heterogeneous monitoring data based on the monitoring models. The data processing module preprocesses the multi-source heterogeneous monitoring data and extracts a set of feature indicators that reflect the operating status of the monitoring unit equipment. Based on the set of feature indicators, the real-time health index value is calculated. The health assessment module compares the health index value of each monitoring unit with a preset threshold range and outputs an assessment conclusion based on the comparison results.

Images