A hydropower station operation data intelligent management system and method based on a data mart

Abstract

This invention discloses an intelligent management system and method for hydropower station operation data based on a data mart, relating to the field of hydropower unit operation anomaly identification technology. The method includes: dividing the hydropower unit into operating condition intervals; obtaining a baseline feature set for all operating condition intervals; sampling actual power for a certain operating condition interval, fitting a curve based on the sampled data points, and determining a first fitted curve for that operating condition interval using a correlation calculation method; obtaining the first fitted curves for all operating condition intervals; storing the baseline feature set and the first fitted curve in a data mart; performing real-time monitoring of a certain hydropower unit and constructing a first monitoring sequence through random sampling; and based on the first monitoring sequence, retrieving the data mart to obtain the baseline feature set and the first fitted curve under the same operating condition, and issuing early warnings according to different situations. This invention has good scalability and adaptability, and is suitable for intelligent monitoring and health diagnosis in multi-unit collaborative operation scenarios.

Description

Technical Field

[0001] This invention relates to the field of hydropower unit operation anomaly identification technology, specifically a data mart-based intelligent management system and method for hydropower station operation data. Background Technology

[0002] As the core equipment of a hydropower station, the operating status of hydropower units directly affects power generation efficiency and equipment lifespan. With increasing operating time, factors such as guide vane scaling or jamming, impeller wear, increased bearing friction, aging of the excitation system, and poor cooling can all cause deviations between the actual output power and the theoretical power, thus affecting overall operating efficiency and energy conversion performance. Traditional hydropower unit operating status assessments mainly rely on periodic maintenance and manual experience, lacking real-time capabilities and data correlation analysis. This often results in long detection cycles, delayed anomaly detection, and an inability to quantify unit performance degradation.

[0003] In existing technologies, some hydropower stations attempt to calculate operating efficiency by real-time monitoring parameters such as head, flow rate, and power to reflect the operating status of the generating units. However, due to the significant phased characteristics of hydropower station operating loads, the operating status of the generating units varies considerably under different load conditions, making it difficult for a single efficiency calculation result to accurately reflect the overall performance changes of the unit. Furthermore, disturbances in the actual operating environment (such as water flow fluctuations and changes in regulation commands) can cause short-term fluctuations in power data, resulting in higher noise components in the efficiency curve and affecting the accuracy of anomaly identification.

[0004] Therefore, how to divide the operating conditions into reasonable ranges based on the operating characteristics of hydropower units under different load conditions, establish a benchmark feature group for each operating range based on historical operating efficiency, and then combine actual power sampling data to perform curve fitting and correlation analysis, so as to achieve real-time comparison and early warning of anomalies, has become a key problem that urgently needs to be solved in the field of hydropower station operation monitoring and intelligent diagnosis. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent management system and method for hydropower station operation data based on a data mart, so as to solve the problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent management of hydropower station operation data based on a data mart, specifically including the following steps:

[0007] Divide the operating condition ranges of the hydropower units; and obtain the baseline characteristic groups for all operating condition ranges based on operating efficiency;

[0008] For a certain operating condition range, several hydropower units are randomly selected as test samples. The actual power of the test samples is sampled within a predetermined time. Based on the sampled data points, a curve is fitted, and the first fitted curve for the operating condition range is determined by the correlation calculation method. The first fitted curves for all operating condition ranges are obtained.

[0009] Store the benchmark feature set and the first fitted curve in a data mart;

[0010] Real-time monitoring of a certain hydropower unit was conducted, and a first monitoring sequence was constructed by sampling inspections.

[0011] Based on the first monitoring sequence, the data mart is retrieved to obtain the benchmark feature group and the first fitting curve under the same working conditions, and early warning is issued according to different situations.

[0012] Furthermore, the process of dividing the hydropower unit into operating condition ranges and obtaining a baseline feature set for all operating condition ranges based on operating efficiency specifically involves:

[0013] Based on the load conditions during the operation of the hydropower station, the hydropower generating units are divided into several operating condition intervals, represented as: x1,...,x i ,...,x n ;x1,...,x i ,...,x n These represent the 1st, ..., i, ..., nth operating condition intervals, respectively.

[0014] Collect the effective head H, flow rate Q, and actual power P of the hydropower units during different operating conditions of the hydropower station. meas ;

[0015] Based on the effective head H, flow rate Q, and actual power P of the hydropower unit operating in the different operating conditions, meas The real-time efficiency η of the unit under different operating conditions is obtained, η=P meas / ρgHQ; where ρ represents the density of water; g represents the acceleration due to gravity; and the flow rate Q is in meters per second (m³). 3 / s; Effective head H is in meters; The average real-time efficiency μ over a given time period T is calculated. η and variance σ η As a reference characteristic set for the corresponding operating condition range of hydropower units;

[0016] The baseline feature sets for all operating condition intervals are obtained one by one, and the baseline feature sets for the operating condition intervals are stored in a data mart as a benchmark for analysis reference.

[0017] The data mart is a small data storage and analysis system designed for specific business departments. In this application, it is used as a micro storage system with fast retrieval capabilities, similar to a high-speed cache. It has a small storage space but high-efficiency retrieval capabilities, which can greatly improve the speed of subsequent processing.

[0018] Furthermore, for a certain operating condition range, several hydropower units are randomly selected as test samples, and the actual electrical power of each test sample is sampled within a predetermined time. A curve is fitted based on the sampled data points, and the first fitted curve for that operating condition range is determined through a correlation calculation method. The first fitted curves for all operating condition ranges are obtained as follows:

[0019] Step S301: For a certain operating condition interval x i Several hydropower units were randomly selected as test samples, and their actual electrical power was sampled within a predetermined time T. The sampling results are characterized as: [(t 1 ,x i_P 1 sta ),...,(t J ,x i_P J sta ),...,(t L ,x i_P L sta )]; where t 1 ,…,t J ,…,t L These represent the 1st, ..., Jth, ..., Lth sampling times, respectively; x i_P 1 sta ,...,x i_P J sta ,...,x i_P L sta They represent t respectively 1 ,…,t J ,…,t L The corresponding actual electrical power; L represents the number of samples;

[0020] Step S302: Calculate the variance of any K consecutive actual power values, where K < L. Sort the K consecutive actual power values ​​with the smallest variance according to the sampling time to obtain the first sampling sequence; characterized as: [x i_P J sta ,…,x i_P J+K-1 sta ];x i_P J+K-1sta This represents the J+K-1th actual electrical power;

[0021] Step S303: Based on the first sampling sequence, with the sampling time of the actual power as the independent variable and the actual power corresponding to the sampling time as the dependent variable, the fitting curve of all test samples is obtained by fitting the data points.

[0022] Step S304: Calculate the similarity between the fitted curves of all test samples sequentially using the Pearson correlation method, and select the fitted curve with the highest similarity (greater than a preset threshold) as the operating condition interval x. i The first fitted curve;

[0023] Step S305: Calculate and obtain the first fitting curve for each working condition interval, and store the first fitting curve of the working condition interval into the data mart.

[0024] The independent variable takes the value of a finite number of sampling times within the given time T.

[0025] It should be noted that Pearson correlation calculation is often used for linear correlation calculation. In this application, due to multiple sampling, the actual power difference is small under normal operation of the hydropower unit. Moreover, faults such as scale or jamming of the guide vanes, wear of the impeller, increased bearing friction, aging of the excitation system, and poor cooling of the hydropower unit can all cause sudden changes in the actual power over the operating time, which will be reflected in the fitting curve as an upward and downward trend. Therefore, this application focuses on the similarity of the changing trends of the two curves through the Pearson correlation method.

[0026] Furthermore, real-time monitoring of a certain hydropower unit was conducted, and a first monitoring sequence was constructed through random sampling, specifically as follows:

[0027] Monitoring is conducted on a hydroelectric generator unit operating at a hydropower station. Any given moment is used as the starting point. Random checks are performed on the actual electrical power of the generator unit over a predetermined time interval T. Specifically, the operating condition interval x is monitored. i The hydropower units were monitored, and the sampling results were characterized as: [(t1,x i_P 1 meas ),...,(t j ,x i_P j meas ),...,(t l ,x i_P l meas )]; where t1,…,t j ,…,t l These represent the sampling times for the 1st, ..., jth, ..., lth sampling periods, respectively; x i_P1 meas ,…,x i_P j meas ,…,x i_P l meas Let t1, ..., t be respectively. j ,…,t l The corresponding actual power; l represents the number of samples, l≥L;

[0028] In the l actual power samples tested, the variance of any K consecutive actual power samples is calculated. The K consecutive actual power samples with the largest variances are sorted according to the sampling time to obtain the first monitoring sequence; characterized as: [x i_P j meas ,...,x i_P j+K-1 meas ];x i_P j+K-1 meas This represents the actual electrical power of the (j+K-1)th sampled item.

[0029] Furthermore, based on the first monitoring sequence, the data mart is retrieved to obtain the baseline feature group and the first fitted curve under the same operating conditions, and early warnings are issued according to different situations, specifically as follows:

[0030] Step 1: Based on the first monitoring sequence [x] i_P j meas ,...,x i_P j+K-1 meas ], retrieve the data mart, and obtain the first sampling sequence [x] corresponding to the same working condition interval. i_P J sta ,…,x i_P J+K-1 sta The sampling deviation W between the first monitoring sequence and the first sampling sequence is obtained by the deviation calculation method; the sampling deviation W is compared with the preset deviation anomaly threshold W', and a first type of power deviation warning is issued for the case where W > W';

[0031] Step 2: For the case where W≤W', perform trend deviation calculation and efficiency anomaly calculation. Specifically, the trend deviation calculation involves: fitting the data points of the first monitoring sequence according to the methods described in steps S303 and S304 to obtain a first monitoring curve; retrieving the data mart to obtain the first fitted curve corresponding to the same working condition interval; calculating the instantaneous change rate at K sampling points in the first monitoring curve and the first fitted curve respectively; taking the average value as the trend deviation of the first monitoring curve and the first fitted curve respectively, denoted as d1 and d2; when d1>d2+δ*σ, a second type of trend deviation warning is issued; where δ represents the adjustment coefficient; σ represents the standard deviation of the first sampling sequence.

[0032] The efficiency anomaly calculation specifically involves: obtaining the effective head, flow rate, and actual power of the hydropower units during the operation of the hydropower station; calculating the real-time average efficiency μ1 and variance σ1 of the hydropower units within a given time T; and retrieving the data mart to obtain the baseline characteristic set μ of the hydropower units under the same operating conditions. η and variance σ η ; Calculate the efficiency deviation Y η Characterized as: Y η =α*|μ1-μ η |+β*|σ1-σ η |;The efficiency deviation Y η A third type of efficiency anomaly warning is issued when the efficiency deviation exceeds the preset efficiency deviation threshold Y', where α and β represent the influence coefficients of the mean and variance, respectively, and α,β>0.

[0033] A data mart-based intelligent management system for hydropower station operation data includes: a working condition interval division module, a fitting curve construction module, a data mart management module, a real-time monitoring module, and an anomaly detection and early warning module.

[0034] The operating condition interval division module is used to divide the hydropower unit into several operating condition intervals according to the load conditions of the hydropower unit during operation.

[0035] The fitting curve construction module is used to randomly select several hydropower units as test samples for a certain operating condition range, sample the actual electric power of the test samples within a predetermined time, fit a curve based on the sampled data points, and determine the first fitting curve for the operating condition range through a correlation calculation method; and obtain the first fitting curve for all operating condition ranges.

[0036] The data mart management module is used to store the benchmark feature groups and the corresponding first fitting curves for each working condition interval;

[0037] The real-time monitoring module is used for real-time monitoring of the hydropower unit;

[0038] The anomaly detection and early warning module is used to retrieve the benchmark feature group and the first fitted curve of the same working condition range from the data mart based on the first monitoring sequence obtained by the real-time monitoring module, and to perform multi-dimensional deviation calculation and early warning judgment.

[0039] The data mart management module is built on a business-oriented lightweight data mart structure, enabling centralized management and fast access to operating data, and providing stable data support for the real-time monitoring module.

[0040] The anomaly detection and early warning module includes a first type of early warning unit, a second type of early warning unit, and a third type of early warning unit;

[0041] The first type of early warning unit is used for power deviation early warning;

[0042] The second type of early warning unit is used for power trend deviation early warning;

[0043] The third type of early warning unit is used for issuing early warnings of efficiency anomalies.

[0044] Furthermore, it also includes a visualization module; the visualization module is used to display the first fitting curve, real-time monitoring sequence and deviation results for different operating conditions, and generate early warning reports to enable maintenance personnel to quickly locate abnormal times.

[0045] Furthermore, it also includes a computing module; the computing module is used to achieve automated computing through a programmable device.

[0046] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention divides the operating status of hydropower units into multiple operating condition intervals according to load conditions, and constructs a benchmark feature group for each operating condition interval based on operating efficiency. This achieves hierarchical modeling and differentiated analysis of unit performance under different operating conditions, overcoming the shortcomings of ambiguous operating condition division and incomparable features in traditional methods. By sampling and fitting actual power data of several normal units within each operating condition interval, and using the Pearson correlation method to select the fitting curve with the highest similarity as the first fitting curve for that operating condition interval, a dynamic benchmark model of typical operating conditions is formed. This ensures the consistency of trend characteristics while suppressing random noise interference, improving the accuracy and stability of operating condition identification. Furthermore, by calculating the Euclidean distance deviation between the real-time monitoring sequence and the historical sampling sequence, and combining a multi-level judgment strategy of trend deviation and efficiency deviation, hierarchical early warning of different types of anomalies such as power mutation, trend drift, and efficiency degradation can be achieved, effectively improving the accuracy and sensitivity of unit operating status monitoring. Meanwhile, this invention employs a lightweight data mart structure tailored to specific business operations to store benchmark feature sets and fitted curves for each operating condition range, achieving efficient retrieval and rapid matching. This significantly improves real-time analysis and early warning response speed while reducing the system's computational burden. This method can achieve rapid identification and early warning of unit performance anomalies without the need for additional sensors, possessing good scalability and adaptability. It is suitable for intelligent monitoring and health diagnosis in multi-unit collaborative operation scenarios, significantly enhancing the safety and intelligence level of hydropower station operation. Attached Figure Description

[0047] Fig. 1 This is a schematic diagram illustrating the process of constructing the data mart in this invention;

[0048] Fig. 2 This is a schematic diagram of the intelligent early warning process of the present invention. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] Example: Figs. 1-2 As shown, the present invention provides a technical solution, a method for intelligent management of hydropower station operation data based on a data mart, which specifically includes the following steps:

[0051] Divide the operating condition ranges of the hydropower units; and obtain the baseline characteristic groups for all operating condition ranges based on operating efficiency;

[0052] Furthermore, the process involves dividing the hydropower unit into operating condition ranges and obtaining a set of baseline features for all operating condition ranges based on operating efficiency; specifically:

[0053] Based on the load conditions during the operation of the hydropower station, the hydropower generating units are divided into several operating condition intervals, represented as: x1,...,x i ,...,x n ;x1,...,x i ,...,x n These represent the 1st, ..., i, ..., nth operating condition intervals, respectively.

[0054] Collect the effective head H, flow rate Q, and actual power P of the hydropower units during different operating conditions of the hydropower station. meas ;

[0055] Based on the effective head H, flow rate Q, and actual power P of the hydropower unit operating in the different operating conditions, meas The real-time efficiency η of the unit under different operating conditions is obtained, η=P meas / ρgHQ; where ρ represents the density of water; g represents the acceleration due to gravity; and the flow rate Q is in meters per second (m³). 3 / s; Effective head H is in meters; The average real-time efficiency μ over a given time period T is calculated. η and variance σ η As a reference characteristic set for the corresponding operating condition range of hydropower units;

[0056] The baseline feature sets for all operating condition intervals are obtained one by one, and the baseline feature sets for the operating condition intervals are stored in a data mart as a benchmark for analysis reference.

[0057] The data mart is a small data storage and analysis system designed for specific business departments. In this application, it is used as a micro storage system with fast retrieval capabilities, similar to a high-speed cache. It has a small storage space but high-efficiency retrieval capabilities, which can greatly improve the speed of subsequent processing.

[0058] It should be noted that the actual power output of a hydropower station is determined by the load conditions during operation. The full load condition is completely different from the partial load condition. Therefore, the hydropower units are divided into several operating condition ranges according to the load conditions during operation of the hydropower station, and are handled according to the actual situation.

[0059] For a certain operating condition range, several hydropower units are randomly selected as test samples. The actual power of the test samples is sampled within a predetermined time. Based on the sampled data points, a curve is fitted, and the first fitted curve for the operating condition range is determined by the correlation calculation method. The first fitted curves for all operating condition ranges are obtained.

[0060] Furthermore, for a certain operating condition range, several hydropower units are randomly selected as test samples, and the actual electrical power of each test sample is sampled within a predetermined time. A curve is fitted based on the sampled data points, and the first fitted curve for that operating condition range is determined through a correlation calculation method. The first fitted curves for all operating condition ranges are obtained as follows:

[0061] Step S301: For a certain operating condition interval x i Several hydropower units were randomly selected as test samples, and their actual electrical power was sampled within a predetermined time T. The sampling results are characterized as: [(t 1 ,x i_P 1 sta ),...,(t J ,x i_P J sta ),...,(t L ,x i_P L sta )]; where t 1 ,…,t J ,…,t L These represent the 1st, ..., Jth, ..., Lth sampling times, respectively; x i_P 1 sta ,...,x i_P J sta ,...,x i_P L sta They represent t respectively 1 ,…,t J ,…,t L The corresponding actual electrical power; L represents the number of samples;

[0062] Step S302: Calculate the variance of any K consecutive actual power values, where K < L. Sort the K consecutive actual power values ​​with the smallest variance according to the sampling time to obtain the first sampling sequence; characterized as: [x i_P J sta ,…,x i_P J+K-1 sta ];x i_P J+K-1 sta This represents the J+K-1th actual electrical power;

[0063] Step S303: Based on the first sampling sequence, with the sampling time of the actual power as the independent variable and the actual power corresponding to the sampling time as the dependent variable, the fitting curve of all test samples is obtained by fitting the data points.

[0064] Step S304: Calculate the similarity between the fitted curves of all test samples sequentially using the Pearson correlation method, and select the fitted curve with the highest similarity (greater than a preset threshold) as the operating condition interval x. i The first fitted curve;

[0065] Step S305: Calculate and obtain the first fitting curve for each working condition interval, and store the first fitting curve of the working condition interval into the data mart.

[0066] The independent variable takes the value of a finite number of sampling times within the given time T.

[0067] It should be noted that Pearson correlation calculation is often used for linear correlation calculation. In this application, due to multiple sampling, the actual power difference is small under normal operation of the hydropower unit. Moreover, faults such as scale or jamming of the guide vanes, wear of the impeller, increased bearing friction, aging of the excitation system, and poor cooling of the hydropower unit can all cause sudden changes in the actual power over the operating time, which will be reflected in the fitting curve as an upward and downward trend. Therefore, this application focuses on the similarity of the changing trends of the two curves through the Pearson correlation method.

[0068] In this embodiment, several hydropower units are selected as test samples, ensuring they are in normal operating condition. After obtaining the fitting curves of several test samples, the fitting curve of the first test sample is used as a benchmark. The similarity between the first and remaining test samples is calculated using the Pearson correlation method. Test samples with similarity greater than a preset threshold are counted. Then, the fitting curves of subsequent test samples are used as benchmarks, and the counts are obtained. The fitting curve with the highest number of similarities is taken as the first fitting curve for the corresponding operating condition interval. In this embodiment, based on the number of sampling points and the preset threshold of the residual power in history, the ratio of the number of sampling points within the residual range to the total number of sampling points is recorded as the preset threshold. The residual power can be taken as the average of the actual power and the predicted power within a predetermined time T. The predicted power P... for Characterized as: P for =ρgHQ; adjustments can also be made based on historical experience, and no restrictions are imposed here.

[0069] Store the benchmark feature set and the first fitted curve in a data mart;

[0070] Real-time monitoring of a certain hydropower unit was conducted, and a first monitoring sequence was constructed by sampling inspections.

[0071] Based on the first monitoring sequence, the data mart is retrieved to obtain the benchmark feature group and the first fitting curve under the same working conditions, and early warning is issued according to different situations.

[0072] Furthermore, real-time monitoring of a certain hydropower unit was conducted, and a first monitoring sequence was constructed through random sampling, specifically as follows:

[0073] Monitoring is conducted on a hydroelectric generator unit operating at a hydropower station. Any given moment is used as the starting point. Random checks are performed on the actual electrical power of the generator unit over a predetermined time interval T. Specifically, the operating condition interval x is monitored. i The hydropower units were monitored, and the sampling results were characterized as: [(t1,x i_P 1 meas ),...,(t j ,x i_P j meas ),...,(t l ,x i_P l meas )]; where t1,…,t j ,…,t l These represent the sampling times for the 1st, ..., jth, ..., lth sampling periods, respectively; x i_P 1 meas ,…,x i_P j meas ,…,x i_P l meas Let t1, ..., t be respectively. j ,…,t l The corresponding actual power; l represents the number of samples, l≥L;

[0074] In the l actual power samples tested, the variance of any K consecutive actual power samples is calculated. The K consecutive actual power samples with the largest variances are sorted according to the sampling time to obtain the first monitoring sequence; characterized as: [x i_P j meas ,...,x i_P j+K-1 meas ];x i_P j+K-1 meas This represents the actual electrical power of the (j+K-1)th sampled item.

[0075] Furthermore, based on the first monitoring sequence, the data mart is retrieved to obtain the benchmark feature group and the first fitted curve under the same operating conditions, such as... Fig. 2As shown, warnings are issued based on different situations, specifically:

[0076] Step 1: Based on the first monitoring sequence [x] i_P j meas ,...,x i_P j+K-1 meas ], retrieve the data mart, and obtain the first sampling sequence [x] corresponding to the same working condition interval. i_P J sta ,…,x i_P J+K-1 sta The sampling deviation W between the first monitoring sequence and the first sampling sequence is obtained by the deviation calculation method; the sampling deviation W is compared with the preset deviation anomaly threshold W', and a first type of power deviation warning is issued for the case where W > W';

[0077] Step 2: For the case where W≤W', perform trend deviation calculation and efficiency anomaly calculation. Specifically, the trend deviation calculation involves: fitting the data points of the first monitoring sequence according to the methods described in steps S303 and S304 to obtain a first monitoring curve; retrieving the data mart to obtain the first fitted curve corresponding to the same working condition interval; calculating the instantaneous change rate at K sampling points in the first monitoring curve and the first fitted curve respectively; taking the average value as the trend deviation of the first monitoring curve and the first fitted curve respectively, denoted as d1 and d2; when d1>d2+δ*σ, a second type of trend deviation warning is issued; where δ represents the adjustment coefficient; σ represents the standard deviation of the first sampling sequence.

[0078] The efficiency anomaly calculation specifically involves: obtaining the effective head, flow rate, and actual power of the hydropower units during the operation of the hydropower station; calculating the real-time average efficiency μ1 and variance σ1 of the hydropower units within a given time T; and retrieving the data mart to obtain the baseline characteristic set μ of the hydropower units under the same operating conditions. η and variance σ η ; Calculate the efficiency deviation Y η Characterized as: Y η =α*|μ1-μ η |+β*|σ1-σ η |;The efficiency deviation Y η A third type of efficiency anomaly warning is issued when the efficiency deviation exceeds the preset efficiency deviation threshold Y', where α and β represent the influence coefficients of the mean and variance, respectively, and α,β>0.

[0079] It should be noted that the first monitoring sequence and the first sampling sequence need to be converted into numerical values ​​for calculation, keeping the units consistent while removing the units.

[0080] First, the sampling sequence is the one with the smallest variance, meaning that the actual power output of the hydropower unit remains the most stable during operation within this operating range. Using this as a benchmark ensures the accuracy of the deviation calculation. Second, when selecting the monitoring sequence, the one with the largest variance is chosen to ensure that the data in this set best reflects the instability of the hydropower unit. Calculating the deviation between the two sequences can determine to the greatest extent whether the hydropower unit has experienced power anomalies.

[0081] Scaling or jamming of the guide vanes of hydroelectric generators, wear of the impeller, increased bearing friction, aging of the excitation system, and poor cooling can all affect power. Such faults do not immediately produce huge power changes, but they can cause greater hidden dangers in the long run. Therefore, it is necessary to monitor such faults regularly to ensure the operation of the equipment and the prevention of subsequent risks.

[0082] In this embodiment, the deviation calculation method uses Euclidean distance, which is the sum of the absolute differences between each data point in the first monitoring sequence and the first sampling sequence. Alternatively, methods such as dynamic time warping can be used, and there is no limitation here. However, it should be noted that the setting of the deviation anomaly threshold needs to correspond to the deviation calculation method. In this embodiment, the deviation calculation method uses Euclidean distance, so the sum of the standard deviations of each actual power in the first sampling sequence is used as the deviation anomaly threshold. Alternatively, the deviation anomaly threshold can be set by normal simulation or by setting the sum of standard deviations that are integer multiples of each other (quasi-normal) according to the actual situation.

[0083] It should be noted that the efficiency deviation Y η Based on the mean and variance over a given time period T, the mean represents the baseline of efficiency over the given time period T, while the variance represents the fluctuation of efficiency over the given time period T. The combination of the two includes both the baseline difference and reflects the abnormal fluctuations.

[0084] A data mart-based intelligent management system for hydropower station operation data includes: a working condition interval division module, a fitting curve construction module, a data mart management module, a real-time monitoring module, and an anomaly detection and early warning module.

[0085] The operating condition interval division module is used to divide the hydropower unit into several operating condition intervals based on the load conditions of the hydropower unit during operation.

[0086] The curve fitting module is used to randomly select several hydropower units as test samples for a certain operating condition range, sample the actual electric power of the test samples within a predetermined time, fit a curve based on the sampled data points, and determine the first fitted curve for the operating condition range through a correlation calculation method; and obtain the first fitted curve for all operating condition ranges.

[0087] The data mart management module is used to store the baseline feature sets and the corresponding first fitted curves for each operating condition range;

[0088] The real-time monitoring module is used for real-time monitoring of the hydropower unit;

[0089] The anomaly detection and early warning module is used to retrieve the benchmark feature group and the first fitted curve of the same working condition range from the data mart based on the first monitoring sequence obtained by the real-time monitoring module, and to perform multi-dimensional deviation calculation and early warning judgment.

[0090] The data mart management module is built on a business-oriented lightweight data mart structure, enabling centralized management and fast access to operational data, and providing stable data support for the real-time monitoring module.

[0091] The anomaly detection and early warning module includes a first-type early warning unit, a second-type early warning unit, and a third-type early warning unit;

[0092] The first type of early warning unit is used for power deviation early warning;

[0093] The second type of early warning unit is used for power trend deviation early warning;

[0094] The third type of early warning unit is used for issuing early warnings of efficiency anomalies.

[0095] Furthermore, it also includes a visualization module; the visualization module is used to display the first fitted curve, real-time monitoring sequence and deviation results for different operating conditions, and generate early warning reports to help maintenance personnel quickly locate abnormal times.

[0096] Furthermore, it also includes a computing module; the computing module is used to achieve automated computing through a programmable device.

[0097] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for intelligent management of hydropower station operation data based on a data mart, characterized in that: Specifically, the steps include the following: The operating condition ranges of the hydropower units are divided; and a baseline feature set for all operating condition ranges is obtained based on operating efficiency, specifically: The hydropower units are divided into several operating condition ranges based on the load conditions during the operation of the hydropower station. Collect the effective head H, flow rate Q, and actual power P of the hydropower units during different operating conditions of the hydropower station. meas ; Based on the effective head H, flow rate Q, and actual power Pmeas of the hydropower unit operating in different operating conditions, the real-time efficiency η of the unit operating in different operating conditions is obtained, η = Pmeas / ρgHQ; where ρ represents water density; g represents gravitational acceleration; the average value μη and variance ση of the real-time efficiency within a given time T are used as the benchmark characteristic set of the corresponding operating conditions of the hydropower unit. For a certain operating condition range, several hydropower units are randomly selected as test samples. Actual power is sampled from each test sample within a predetermined time period. A curve is fitted based on the sampled data points, and the first fitted curve for that operating condition range is determined using a correlation calculation method. The first fitted curves for all operating condition ranges are obtained as follows: For a certain operating condition interval x i Several hydropower units were randomly selected as test samples, and their actual electrical power was sampled within a predetermined time T. The sampling results are characterized as: [(t 1 ,x i_P 1 sta ),...,(t J ,x i_P J sta ),...,(t L ,x i_P L sta )]; where t 1 ,…,t J ,…,t L These represent the 1st, ..., Jth, ..., Lth sampling times, respectively; x i_P 1 sta ,...,x i_P J sta ,...,x i_P L sta They represent t respectively 1 ,…,t J ,…,t L The corresponding actual electrical power; L represents the number of samples; Step S302: Calculate the variance of any K consecutive actual power values, where K < L. Sort the K consecutive actual power values ​​with the smallest variance according to the sampling time to obtain the first sampling sequence; characterized as: [x i_P J sta ,…,x i_P J+K-1 sta ];x i_P J+K-1 sta This represents the J+K-1th actual electrical power; Step S303: Based on the first sampling sequence, with the sampling time of the actual power as the independent variable and the actual power corresponding to the sampling time as the dependent variable, the fitting curve of all test samples is obtained by fitting the data points. Step S304: Calculate the similarity between the fitted curves of all test samples sequentially using the Pearson correlation method, and select the fitted curve with the highest similarity (greater than a preset threshold) as the operating condition interval x. i The first fitted curve; Step S305: Calculate and obtain the first fitted curve for each working condition interval; Store the benchmark feature set and the first fitted curve in a data mart; Real-time monitoring of a certain hydropower unit was conducted, and a first monitoring sequence was constructed by sampling inspections. Based on the first monitoring sequence, the data mart is retrieved to obtain the benchmark feature group and the first fitted curve under the same working conditions, deviation analysis is performed, and early warnings are issued according to different situations.

2. The intelligent management method for hydropower station operation data based on a data mart according to claim 1, characterized in that: Also includes: The baseline feature sets for all operating condition intervals are obtained one by one, and the baseline feature sets for the operating condition intervals are stored in a data mart as a benchmark for analysis reference.

3. The intelligent management method for hydropower station operation data based on a data mart according to claim 2, characterized in that: Also includes: The independent variable takes the value of a finite number of sampling times within the given time T.

4. The intelligent management method for hydropower station operation data based on a data mart according to claim 3, characterized in that: The real-time monitoring of a certain hydropower unit, and the sampling to construct the first monitoring sequence, specifically involves: Monitoring is conducted on a hydroelectric generator unit operating at a hydropower station. Any given moment is used as the starting point. Random checks are performed on the actual electrical power of the generator unit over a predetermined time interval T. Specifically, the operating condition interval x is monitored. i The hydropower units were monitored, and the sampling results were characterized as: [(t1,x i_P 1 meas ),...,(t j ,x i_P j meas ),...,(t l ,x i_P l meas )]; where t1,…,t j ,…,t l These represent the sampling times for the 1st, ..., jth, ..., lth sampling periods, respectively; x i_P 1 meas ,…,x i_P j meas ,…,x i_P l meas Let t1, ..., t be respectively. j ,…,t l The corresponding actual power; l represents the number of samples, l≥L; In the l actual power samples tested, the variance of any K consecutive actual power samples is calculated. The K consecutive actual power samples with the largest variances are sorted according to the sampling time to obtain the first monitoring sequence; characterized as: [x i_P j meas ,...,x i_P j+K-1 meas ];x i_P j+K-1 meas This represents the actual electrical power of the (j+K-1)th sampled item.

5. The intelligent management method for hydropower station operation data based on a data mart according to claim 4, characterized in that: Based on the first monitoring sequence, the data mart is retrieved to obtain the benchmark feature group and the first fitted curve under the same working conditions, deviation analysis is performed, and early warnings are issued according to different situations, specifically as follows: Step 1: Based on the first monitoring sequence [x] i_P j meas ,...,x i_P j+K-1 meas ], retrieve the data mart, and obtain the first sampling sequence [x] corresponding to the same working condition interval. i_P J sta ,…,x i_P J+K-1 sta The sampling deviation W between the first monitoring sequence and the first sampling sequence is obtained by the deviation calculation method; the sampling deviation W is compared with the preset deviation anomaly threshold W', and a first type of power deviation warning is issued for the case where W > W'; Step 2: For the case where W≤W', perform trend deviation calculation and efficiency anomaly calculation. Specifically, the trend deviation calculation involves: fitting the data points of the first monitoring sequence according to the methods described in steps S303 and S304 to obtain a first monitoring curve; retrieving the data mart to obtain the first fitted curve corresponding to the same working condition interval; calculating the instantaneous change rate at K sampling points in the first monitoring curve and the first fitted curve respectively; taking the average value as the trend deviation of the first monitoring curve and the first fitted curve respectively, denoted as d1 and d2; when d1>d2+δ*σ, a second type of trend deviation warning is issued; where δ represents the adjustment coefficient; σ represents the standard deviation of the first sampling sequence. The efficiency anomaly calculation specifically involves: obtaining the effective head, flow rate, and actual power of the hydropower units during the operation of the hydropower station; calculating the real-time average efficiency μ1 and variance σ1 of the hydropower units within a given time T; and retrieving the data mart to obtain the baseline characteristic set μ of the hydropower units under the same operating conditions. η and variance σ η ; Calculate the efficiency deviation Y η Characterized as: Y η =α*|μ1-μ η |+β*|σ1-σ η |;The efficiency deviation Y η A third type of efficiency anomaly warning is issued when the efficiency deviation exceeds the preset efficiency deviation threshold Y', where α and β represent the influence coefficients of the mean and variance, respectively, and α,β>0.

6. A smart management system for hydropower station operation data based on a data mart, characterized in that: include: The module includes a working condition interval division module, a fitting curve construction module, a data mart management module, a real-time monitoring module, and an anomaly detection and early warning module. The operating condition interval division module is used to divide the hydropower unit into several operating condition intervals based on the load conditions during operation, and to collect the effective head H, flow rate Q, and actual power P of the hydropower unit in different operating condition intervals during the operation of the hydropower station. meas Based on the effective head H, flow rate Q, and actual power Pmeas of the hydropower unit operating in different operating conditions, the real-time efficiency η of the unit operating in different operating conditions is obtained, η = Pmeas / ρgHQ; where ρ represents water density; g represents gravitational acceleration; the average value μη and variance ση of the real-time efficiency within a given time T are used as the benchmark characteristic set of the corresponding operating conditions of the hydropower unit. The curve fitting module is used to randomly select several hydropower units as test samples for a certain operating condition range, sample the actual power of the test samples within a predetermined time, fit a curve based on the sampled data points, and determine the first fitted curve for that operating condition range through a correlation calculation method; specifically, the first fitted curves for all operating condition ranges are obtained as follows: Step S301: For a certain operating condition interval x i Several hydropower units were randomly selected as test samples, and the actual power of each test sample was sampled within a predetermined time T. Step S302: Calculate the variance of any K consecutive actual power values, where K < L. Sort the K consecutive actual power values ​​with the smallest variance according to the sampling time to obtain the first sampling sequence. Step S303: Based on the first sampling sequence, with the sampling time of the actual power as the independent variable and the actual power corresponding to the sampling time as the dependent variable, the fitting curve of all test samples is obtained by fitting the data points. Step S304: Calculate the similarity between the fitted curves of all test samples sequentially using the Pearson correlation method, and select the fitted curve with the highest similarity (greater than a preset threshold) as the operating condition interval x. i The first fitted curve; Step S305: Calculate and obtain the first fitted curve for each working condition interval; The data mart management module is used to store the benchmark feature groups and the corresponding first fitting curves for each working condition interval; The real-time monitoring module is used for real-time monitoring of the hydropower unit; The anomaly detection and early warning module is used to retrieve the benchmark feature group and the first fitted curve of the same working condition range from the data mart based on the first monitoring sequence obtained by the real-time monitoring module, and to perform multi-dimensional deviation calculation and early warning judgment.

7. The intelligent management system for hydropower station operation data based on a data mart according to claim 6, characterized in that: The data mart management module is built on a business-oriented lightweight data mart structure, enabling centralized management and fast access to operating data, and providing stable data support for the real-time monitoring module.

8. The intelligent management system for hydropower station operation data based on a data mart according to claim 6, characterized in that: The anomaly detection and early warning module includes a first type of early warning unit, a second type of early warning unit, and a third type of early warning unit; The first type of early warning unit is used for power deviation early warning; The second type of early warning unit is used for power trend deviation early warning; The third type of early warning unit is used for issuing early warnings of efficiency anomalies.

9. The intelligent management system for hydropower station operation data based on a data mart according to claim 8, characterized in that: It also includes a visualization module; the visualization module is used to display the first fitting curve, real-time monitoring sequence and deviation results for different operating conditions, and generate early warning reports to enable maintenance personnel to quickly locate abnormal times.

10. The intelligent management system for hydropower station operation data based on a data mart according to claim 9, characterized in that: It also includes a computing module; the computing module is used to perform automated calculations through a programmable device.

Images