A method for monitoring and judging real-time state data quality of hydroelectric equipment

Abstract

The application discloses a kind of real-time state data quality monitoring and discrimination method of hydroelectric equipment, belongs to the technical field of hydroelectric equipment monitoring, including acquisition measuring point data and working condition associated parameter is stored in classification;According to the classification processing of measuring point quantity, index category label is obtained and data quality discrimination model is established;Establish communication state discrimination model to carry out the detection processing of silence condition and obtain the communication state discrimination result of each measuring point;If communication is interrupted, push alarm information;If communication is normal, parallel trigger all data quality discrimination models to obtain data quality discrimination result;If abnormal, push alarm information and update the database containing working condition associated parameter label based on the standard value of the measuring point data of the current abnormal measuring point after revision;If normal, push state information, and update the database containing working condition associated parameter label.The application solves the problem of monitoring failure caused by unstable data quality and communication interruption in the operation of hydroelectric equipment in the prior art.

Description

TECHNICAL FIELD

[0001] The application relates to a water and electricity equipment real-time state data quality monitoring and distinguishing method and belongs to the technical field of water and electricity equipment monitoring. BACKGROUND

[0002] In the field of water and electricity equipment operation monitoring, the existing technology usually adopts a multi-level architecture to realize equipment state monitoring and data transmission. The core components include a sensor network deployed at key positions such as a generator set, a transformer and a water conveying pipeline, which is used for real-time collection of equipment operation parameters; a data collection module pre-processes the sensor signals and transmits the data to a monitoring center through a wired or wireless communication network; the monitoring center stores, analyzes and visualizes the data based on a SCADA system or an industrial Internet of Things platform, and realizes abnormal early warning in combination with preset thresholds or simple algorithm models. Some advanced systems also introduce edge computing nodes for localized data filtering and preliminary diagnosis to reduce the burden of the center server. The architecture realizes remote monitoring of the operation state of the water and electricity equipment through hierarchical cooperation and provides data support for operation and maintenance decisions.

[0003] However, the existing technology still has significant defects in the complex operation scene of water and electricity equipment: firstly, the sensors are easily affected by strong electromagnetic interference, humid environment and equipment vibration in the water and electricity station, which can cause signal drift or noise interference, resulting in distorted collected data; in the data transmission process, poor communication protocol compatibility or overly simplified compression algorithm further causes loss of key information such as vibration waveform distortion, making it difficult for the monitoring center to accurately restore the real state of the equipment. Secondly, wired networks are prone to line interruption due to geological disasters or equipment aging, and wireless networks are frequently disconnected due to the coverage blind area of remote water and electricity stations and bad weather; the existing technology lacks a breakpoint continuation and local cache mechanism, once the communication is interrupted, the historical data is lost and the real-time state cannot be updated, causing the monitoring system to fall into a "blind area", and the operation and maintenance personnel cannot respond to sudden failures in time, significantly increasing the risk of equipment damage and operation and maintenance cost. SUMMARY

[0004] The purpose of the application is to provide a water and electricity equipment real-time state data quality monitoring and distinguishing method, which collects the measurement point data and working condition related parameters of the water and electricity equipment in real time, constructs a hierarchical database and implements measurement point classification and multi-model dynamic discrimination, and combines communication state detection and data quality discrimination to solve the problem of unstable data quality in the existing technology in the operation of water and electricity equipment and the monitoring failure caused by communication interruption, and ensure the data accuracy and system operation reliability.

[0005] To solve the above technical problems, the application is implemented by using the following technical scheme:

[0006] The application provides a water and electricity equipment real-time state data quality monitoring and distinguishing method, which comprises:

[0007] Real-time acquisition of each measurement point data corresponding to each measurement point of the hydroelectric operation equipment and working condition associated parameters;

[0008] The measurement point data includes switch quantity, state quantity and analog quantity, and the working condition associated parameters include load rate, water head height and equipment operation mode;

[0009] The measurement point data is classified and stored according to the equipment-component-index-measurement point hierarchical structure, and a database containing working condition associated parameter labels is constructed;

[0010] According to the number of measurement points associated with each index, the index category label is obtained through classification processing, and a data quality discrimination model is established;

[0011] The communication state discrimination model is established to detect the silence condition of each measurement point to obtain the communication state discrimination result of each measurement point;

[0012] If the communication state discrimination result is communication interruption, an alarm information is pushed;

[0013] If the communication state discrimination result is normal communication, the data quality discrimination model corresponding to all index category labels is triggered in parallel to obtain the data quality discrimination result;

[0014] If the data quality discrimination result is abnormal, an alarm information is pushed, the standard value of the measurement point data of the current abnormal measurement point is corrected, and the database containing the working condition associated parameter labels is updated based on the corrected standard value of the measurement point data of the current abnormal measurement point;

[0015] If the data quality discrimination result is normal, a state information is pushed, and the database containing the working condition associated parameter labels is updated based on the measurement point data of all current measurement points.

[0016] Further, the communication state discrimination model includes a first communication state discrimination model and a second communication state discrimination model;

[0017] The first communication state discrimination model is used to judge whether the communication state discrimination result is communication interruption or normal communication, and includes:

[0018] If the current measurement point satisfies , it is judged that the communication state discrimination result of the current measurement point is communication interruption;

[0019] If the current measurement point does not satisfy , it is judged that the communication state discrimination result of the current measurement point is normal communication;

[0020] Wherein, represents the actual measurement value of the measurement point A at the current time T i , represents the actual measurement value of the measurement point A at the last time T , represents the actual measurement value of the measurement point A at the last time Ti-a the actual measurement value of the measurement point A, , is a positive integer and satisfies ;

[0021] The second communication state determination model is used for performing a silent condition detection process on each measurement point to obtain a communication state determination result of each measurement point, and includes the following steps:

[0022] Based on the measurement point data corresponding to each measurement point, the state transition frequency and the duration of the on-off quantity, the data refresh period of the state quantity, and the numerical fluctuation amplitude of the analog quantity are monitored within a preset time period;

[0023] According to the actual state transition frequency and the rated state transition frequency of the on-off quantity, an on-off quantity silence factor is calculated, wherein the on-off quantity silence factor = 1 - (actual state transition frequency / rated state transition frequency);

[0024] According to the actual data refresh period and the rated data refresh period of the state quantity, a state quantity silence factor is calculated, wherein the state quantity silence factor = 1 - (actual data refresh period / rated data refresh period);

[0025] According to the actual numerical fluctuation amplitude and the rated numerical fluctuation amplitude of the analog quantity, an analog quantity silence factor is calculated, wherein the analog quantity silence factor = 1 - (actual numerical fluctuation amplitude / rated numerical fluctuation amplitude);

[0026] According to the on-off quantity silence factor, the state quantity silence factor, and the analog quantity silence factor, a comprehensive silence index is calculated, wherein the comprehensive silence index = (on-off quantity silence factor + state quantity silence factor + analog quantity silence factor) / 3;

[0027] When the comprehensive silence index is greater than a preset first threshold value and the duration is greater than a preset second threshold value, it is determined that the communication is interrupted, otherwise it is determined that the communication is normal.

[0028] Further, the index category label includes L1 type index and L2 type index, wherein

[0029] If the number of measurement points associated with each index is greater than 2, it is determined to be an L1 type index;

[0030] If the number of measurement points associated with each index is less than or equal to 2, it is determined to be an L2 type index.

[0031] Further, a first data quality determination model and a second data quality determination model are established according to the L1 type index and the L2 type index, respectively;

[0032] The first data quality determination model includes a distance analysis module and a MAD optimization module;

[0033] The distance analysis module is configured to calculate Mahalanobis distance of data distribution between measuring points according to measuring point data corresponding to all measuring points under a certain type of index to screen out abnormal measuring points;

[0034] The MAD optimization module is configured to introduce a deterioration coefficient into median absolute deviation (MAD) to calculate a data quality discrimination threshold of measuring point data corresponding to all measuring points under a certain type of index, and to discriminate abnormal measuring points exceeding the index data quality discrimination threshold as data quality abnormality to generate a data quality discrimination result;

[0035] The second data quality discrimination model comprises a working condition parameter extraction and matching module and a weighted range calculation module.

[0036] The working condition parameter extraction and matching module is configured to extract working condition associated parameters of a current measuring point from a database comprising working condition associated parameter tags and match a historical database comprising working condition associated parameter tags to obtain a normal value range and a standard deviation of the working condition associated parameters of the current measuring point.

[0037] The weighted range calculation module is configured to calculate a deviation allowable range of the working condition associated parameters of the current measuring point according to a seasonal correction coefficient and a standard deviation of the working condition associated parameters of the current measuring point, and to generate a data quality discrimination result according to the normal value range of the working condition associated parameters of the current measuring point and the deviation allowable range of the working condition associated parameters of the current measuring point.

[0038] Further, the abnormal measuring points are screened out by calculating Mahalanobis distance of data distribution between measuring points according to measuring point data corresponding to all measuring points under a certain type of index, comprising:

[0039] The actual measurement values of the measuring point data corresponding to all measuring points under a certain type of index are subjected to Z-score standardization to obtain standardized measuring point data corresponding to all measuring points;

[0040] Based on the standardized measuring point data corresponding to all measuring points, a measuring point inter-matrix is constructed to calculate Mahalanobis distance values of all measuring points;

[0041] If the Mahalanobis distance value of the current measuring point data is greater than a preset distance value alarm threshold, the current measuring point is marked as an abnormal measuring point.

[0042] Further, the deterioration coefficient is introduced into the median absolute deviation (MAD) to calculate a data quality discrimination threshold of measuring point data corresponding to all measuring points under a certain type of index, and the abnormal measuring points exceeding the index data quality discrimination threshold are discriminated as data quality abnormality to generate a data quality discrimination result, comprising:

[0043] The deterioration coefficient is calculated based on the equipment running time and the preset equipment life, and the calculation formula of the deterioration coefficient is represented as:

[0044]

[0045] wherein, represents the unit deterioration coefficient, represents the unit deterioration coefficient of the last overhaul period of a certain type of index the maximum value of the absolute value, represents the first occurrence corresponding to the time, represents the unit deterioration coefficient of the resumption time of the last overhaul period of a certain type of index the initial value of, represents the first occurrence corresponding to the time, T represents the overhaul period, , respectively, weight coefficients of,

[0046] extracting a sliding time window of the measuring point in the healthy state of the equipment, and calculating the median absolute deviation (MAD) of the measuring point data corresponding to all measuring points in a certain type of index in the sliding time window;

[0047] introducing the deterioration coefficient into the median absolute deviation (MAD) to calculate and generate the data quality discrimination threshold of the measuring point data corresponding to all measuring points in a certain type of index, wherein the calculation formula of the data quality discrimination threshold is represented as:

[0048] ;

[0049] wherein, represents the data quality discrimination threshold, represents the median absolute deviation of the measuring point data corresponding to a certain measuring point;

[0050] real-time quality discrimination is performed on the measuring point data corresponding to all measuring points in a certain type of index, including:

[0051] When the median absolute deviation of the measuring point data corresponding to a certain measuring point in a certain type of index is greater than the data quality discrimination threshold, the data quality discrimination result is judged to be abnormal;

[0052] When the median absolute deviation of the measuring point data corresponding to a certain measuring point in a certain type of index is less than or equal to the data quality discrimination threshold, the data quality discrimination result is judged to be normal.

[0053] Further, the working condition associated parameter of the current measuring point is extracted from the database containing the working condition associated parameter label, and the historical database containing the working condition associated parameter label is matched to obtain the normal value range and standard deviation of the working condition associated parameter of the current measuring point, including:

[0054] extracting a working condition correlation parameter of the current measuring point from a database containing working condition correlation parameter tags;

[0055] performing similarity matching in a historical database containing working condition correlation parameters based on a preset matching rule according to the working condition correlation parameter of the current measuring point, and screening a historical working condition correlation parameter of the current measuring point;

[0056] calculating a normal value range and a standard deviation of the working condition correlation parameter of the current measuring point according to the historical working condition correlation parameter of the current measuring point.

[0057] Further, a deviation allowable range of the working condition correlation parameter of the current measuring point is calculated according to the seasonal correction coefficient and the standard deviation of the working condition correlation parameter of the current measuring point, and a data quality discrimination result is generated according to the normal value range of the working condition correlation parameter of the current measuring point and the deviation allowable range of the working condition correlation parameter of the current measuring point, including:

[0058] calculating the seasonal correction coefficient according to a difference between the current environmental temperature and the standard temperature;

[0059] calculating the deviation allowable range of the working condition correlation parameter of the current measuring point according to the standard deviation of the working condition correlation parameter of the current measuring point and the seasonal correction coefficient;

[0060] if the actual measurement value of the working condition correlation parameter of the current measuring point is within the normal value range, judging that the data quality discrimination result is normal;

[0061] if the actual measurement value of the working condition correlation parameter of the current measuring point is within the deviation allowable range but exceeds the normal value range or the actual measurement value exceeds the deviation allowable range, judging that the data quality discrimination result is abnormal.

[0062] Further, the standard value of the measuring point data of the current abnormal measuring point is corrected, including:

[0063] correcting the standard value of the measuring point data of the current abnormal measuring point based on the actual measurement value of all measuring points at the current time point from the time interval from the last time of unit overhaul to the current time after the last time of unit overhaul, to obtain a corrected standard value of the measuring point data of the current abnormal measuring point.

[0064] Further, the corrected standard value of the measuring point data of the current abnormal measuring point is expressed as:

[0065] ;

[0066] wherein, the corrected standard value of the measuring point data of the current abnormal measuring point is represented by the standard value of the measuring point data of the current abnormal measuring point is represented by the last time of unit overhaul is represented by represents the time interval from the last maintenance of the unit to the current time, represents the actual measurement value of the measurement point data at the current time corresponding to the measurement point.

[0067] Compared with the prior art, the present application has the following beneficial effects:

[0068] 1. The present application synchronously collects switch value, state value, analog value and other measurement point data, and combines with load rate, water head height and other working condition related parameters to realize full-factor coverage of the equipment operation state, effectively solve the data island problem in traditional monitoring, and improve the accuracy of data correlation analysis; the present application stores data in a four-level hierarchical structure of equipment-component-index-measurement point, and attaches working condition parameter tags, so that the massive heterogeneous data has traceability; the present application innovatively classifies indexes according to the number of measurement points, constructs special quality discrimination models for different categories, and reduces the misjudgment rate; the present application identifies communication interruption through silent detection, and creatively adopts a parallel triggering multi-model discrimination strategy to improve the data repair efficiency in the interrupted state, and ensures the freshness of data in the continuous communication state through real-time database updating.

[0069] 2. The present application divides indexes into L1 type indexes and L2 type indexes according to the number of measurement points, and respectively constructs a first data quality discrimination model based on Mahalanobis distance and an optimized MAD model, and constructs a second data quality discrimination model based on working condition related parameter matching and weighted range calculation. The L1 type indexes are analyzed by multiple measurement points in cooperation, abnormal data are removed by combining with Mahalanobis distance, and a deterioration coefficient is introduced to calculate the MAD threshold, thereby reducing the misjudgment rate in the equipment aging stage; the L2 type indexes calculate the dynamic allowable range through historical working condition related parameter matching and seasonal correction, solve the false alarm problem in the working condition with few measurement points, and improve the reliability of single measurement point data. BRIEF DESCRIPTION OF DRAWINGS

[0070] Figure 1 is a flowchart of a real-time state data quality monitoring and discrimination method of a hydroelectric equipment provided by an embodiment of the present application. DETAILED DESCRIPTION

[0071] The technical scheme of the present application will be described in detail below with reference to the drawings and specific embodiments. It should be understood that the specific features in the embodiments and the specific embodiments of the present application are detailed descriptions of the technical scheme of the present application, and are not limitations of the technical scheme of the present application. In the case of no conflict, the technical features in the embodiments and the specific embodiments of the present application can be combined with each other.

[0072] The term "and / or", only describes the association relation of the associated objects, which means that there can be three relations, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " generally represents that the associated objects before and after are an "or" relationship.

[0073] Embodiment 1

[0074] As Figure 1 shown, the embodiment introduces a real-time state data quality monitoring and discrimination method of hydroelectric equipment, comprising:

[0075] Step one: real-time collection of each measuring point data corresponding to the measuring point of the hydroelectric operation equipment and the working condition associated parameters.

[0076] In the present application, the measuring point data includes switch quantity, state quantity and analog quantity, and the working condition associated parameters include load rate, water head height and equipment operation mode.

[0077] The present application realizes the full-dimensional perception of the hydroelectric equipment operation state and the working condition environment correlation analysis by real-time collection of multiple types of measuring point data such as switch quantity, state quantity, analog quantity and load rate, water head height and equipment operation mode, provides high-granularity, strong-association original data basis for subsequent data analysis, and significantly improves the integrity and accuracy of equipment state monitoring.

[0078] Step two: classifying and storing the measuring point data according to the equipment-component-index-measuring point hierarchical structure and constructing a database containing working condition associated parameter labels.

[0079] The present application realizes the standardized organization and multi-dimensional retrieval of data by constructing a four-level hierarchical database of equipment-component-index-measuring point and embedding working condition associated parameter labels, effectively solves the fragmentation problem of traditional data storage, at the same time supports quick data filtering according to equipment level or working condition, greatly improves the data reuse efficiency and cross-scene analysis capability.

[0080] Step three: classifying and processing according to the number of measuring points associated with each index to obtain index category labels, and establishing a data quality discrimination model.

[0081] The present application realizes the differentiated data quality verification strategy by dynamically classifying the index labels based on the number of measuring points and constructing the data quality discrimination model, reduces the single-measuring-point noise interference misjudgment rate, improves the logical rigor of multi-measuring-point collaborative analysis, and significantly enhances the accuracy and adaptability of data quality evaluation.

[0082] Step four: detecting and processing the silence condition of each measuring point by establishing a communication state discrimination model to obtain the communication state discrimination result of each measuring point.

[0083] If the communication state discrimination result is communication interruption, an alarm information is pushed.

[0084] If the communication state discrimination result is normal communication, data quality discrimination models corresponding to all index category labels are triggered in parallel to obtain a data quality discrimination result.

[0085] The application realizes the identification and immediate alarm of communication interruption failure by establishing a communication state discrimination model to detect the silence condition of the measuring point, and simultaneously triggers a multi-level data quality discrimination model in parallel for normal communication measuring points, thereby improving the data quality analysis efficiency through parallel computing and effectively shortening the abnormal detection response cycle on the basis of ensuring communication reliability.

[0086] Step five: judge the data quality discrimination result state:

[0087] If the data quality discrimination result is abnormal, an alarm information is pushed, the standard value of the measuring point data of the current abnormal measuring point is corrected, and the database containing the working condition associated parameter label is updated based on the corrected standard value of the measuring point data of the current abnormal measuring point.

[0088] If the data quality discrimination result is normal, a state information is pushed, and the database containing the working condition associated parameter label is updated based on the measuring point data of all current measuring points.

[0089] The application implements dynamic correction of the standard value of abnormal measuring point data and synchronous updating of the database through the construction of a data quality result driven closed loop correction mechanism, ensures the real-time accuracy of data storage, maintains the data timeliness through full database updating for normal data, forms an automatic operation and maintenance chain of "detection-correction-feedback", significantly reduces the manual intervention cost and improves the system self-healing ability.

[0090] Embodiment 2

[0091] Based on the same inventive concept as embodiment 1, this embodiment introduces an implementation step of a real-time state data quality monitoring and discrimination method of hydroelectric equipment, which includes:

[0092] Step 1: real-time acquisition of measuring point data corresponding to each measuring point of the hydroelectric operating equipment and working condition associated parameters.

[0093] In this embodiment, the measuring point data includes switch quantity, state quantity and analog quantity, and the working condition associated parameters include load rate, water head height and equipment operation mode.

[0094] Step 2: classifying and storing the measuring point data according to the equipment-component-index-measuring point hierarchical structure and constructing a database containing working condition associated parameter labels.

[0095] Step 3: Classify the indicators according to the number of associated measuring points to obtain indicator category labels, and establish a data quality discrimination model.

[0096] In this embodiment, the indicator category labels include L1 category indicators and L2 category indicators, wherein if the number of measuring points associated with each indicator is greater than 2, the indicator is determined as an L1 category indicator; and if the number of measuring points associated with each indicator is less than or equal to 2, the indicator is determined as an L2 category indicator.

[0097] In this embodiment, a first data quality discrimination model and a second data quality discrimination model are respectively established according to the L1 category indicators and the L2 category indicators.

[0098] In this embodiment, the first data quality discrimination model includes a distance analysis module and a MAD optimization module; the distance analysis module is configured to calculate the Mahalanobis distance of data distribution between measuring points according to the measuring point data corresponding to all measuring points under a certain category of indicators, and screen out abnormal measuring points; and the MAD optimization module is configured to introduce a degradation coefficient into the median absolute deviation (MAD) to calculate a data quality discrimination threshold of the measuring point data corresponding to all measuring points under a certain category of indicators, and determine the abnormal measuring points exceeding the indicator data quality discrimination threshold as data quality abnormality to generate a data quality discrimination result.

[0099] In this embodiment, the abnormal measuring points are screened out by calculating the Mahalanobis distance of data distribution between measuring points according to the measuring point data corresponding to all measuring points under a certain category of indicators, and the method includes:

[0100] The actual measurement values of the measuring point data corresponding to all measuring points under a certain category of indicators are subjected to Z-score standardization to obtain the standardized measuring point data corresponding to all measuring points;

[0101] Based on the standardized measuring point data corresponding to all measuring points, a measuring point matrix is constructed to calculate the Mahalanobis distance values of all measuring points;

[0102] If the Mahalanobis distance value of the current measuring point data is greater than a preset distance value alarm threshold, the current measuring point is marked as an abnormal measuring point.

[0103] In this embodiment, the degradation coefficient is introduced into the median absolute deviation (MAD) to calculate the data quality discrimination threshold of the measuring point data corresponding to all measuring points under a certain category of indicators, and the abnormal measuring points exceeding the indicator data quality discrimination threshold are determined as data quality abnormality to generate a data quality discrimination result, which includes:

[0104] The degradation coefficient is calculated based on the device running time and the preset device service life, and the calculation formula of the degradation coefficient is represented as:

[0105]

[0106] In the formula, This represents the unit's degradation coefficient. This represents the unit's degradation coefficient in the previous major overhaul cycle for a certain type of indicator. The maximum absolute value, Indicates the first appearance The corresponding time, This represents the unit deterioration coefficient at the time of resumption of production in the previous major overhaul cycle, representing a certain type of indicator. initial value, Indicates the first appearance The corresponding time, T represents the overhaul cycle. , They are respectively Weighting coefficients;

[0107] Extract the sliding time window of the measuring points under the health status of the equipment, and calculate the median absolute deviation (MAD) of the measuring point data corresponding to all measuring points under a certain type of index within the sliding time window.

[0108] The degradation coefficient is incorporated into the median absolute deviation (MAD) to calculate the data quality threshold for all measuring points under a certain type of index. The formula for calculating the data quality threshold is as follows:

[0109] ;

[0110] In the formula, This indicates the threshold for judging data quality. This represents the absolute deviation of the median of the measurement data corresponding to a given measurement point.

[0111] Real-time quality assessment of all measuring point data for a certain type of indicator, including:

[0112] When the median absolute deviation of the data at a certain measurement point under a certain type of indicator is greater than the data quality judgment threshold, the data quality judgment result is judged to be abnormal.

[0113] When the median absolute deviation of the data at a certain measurement point under a certain type of indicator is less than or equal to the data quality judgment threshold, the data quality judgment result is judged to be normal.

[0114] In the embodiment, the second data quality discrimination model comprises a working condition parameter extraction and matching module and a weighted range calculation module; the working condition parameter extraction and matching module is configured to extract working condition associated parameters of the current measuring point from a database comprising working condition associated parameter labels and match a historical database comprising working condition associated parameter labels to obtain a normal value range and a standard deviation of the working condition associated parameters of the current measuring point; and the weighted range calculation module is configured to calculate a deviation allowable range of the working condition associated parameters of the current measuring point according to a seasonal correction coefficient and the standard deviation of the working condition associated parameters of the current measuring point, and generate a data quality discrimination result according to the normal value range of the working condition associated parameters of the current measuring point and the deviation allowable range of the working condition associated parameters of the current measuring point.

[0115] In the embodiment, the working condition parameter extraction and matching module is configured to extract working condition associated parameters of the current measuring point from a database comprising working condition associated parameter labels and match a historical database comprising working condition associated parameter labels to obtain a normal value range and a standard deviation of the working condition associated parameters of the current measuring point.

[0116] The working condition associated parameters of the current measuring point are extracted from the database comprising working condition associated parameter labels.

[0117] The historical working condition associated parameters of the current measuring point are filtered according to the working condition associated parameters of the current measuring point and similarity matching based on a preset matching rule in the historical database comprising working condition associated parameters.

[0118] The normal value range and the standard deviation of the working condition associated parameters of the current measuring point are calculated according to the historical working condition associated parameters of the current measuring point.

[0119] In the embodiment, the deviation allowable range of the working condition associated parameters of the current measuring point is calculated according to a seasonal correction coefficient and the standard deviation of the working condition associated parameters of the current measuring point, and the data quality discrimination result is generated according to the normal value range of the working condition associated parameters of the current measuring point and the deviation allowable range of the working condition associated parameters of the current measuring point.

[0120] The seasonal correction coefficient is calculated according to a difference between the current environmental temperature and the standard temperature.

[0121] The deviation allowable range of the working condition associated parameters of the current measuring point is calculated according to the standard deviation of the working condition associated parameters of the current measuring point and the seasonal correction coefficient.

[0122] If the actual measurement value of the working condition associated parameters of the current measuring point is within the normal value range, the data quality discrimination result is determined to be normal.

[0123] If the actual measurement value of the working condition associated parameters of the current measuring point is within the deviation allowable range but exceeds the normal value range or the actual measurement value exceeds the deviation allowable range, the data quality discrimination result is determined to be abnormal.

[0124] Step 4: The communication state discrimination model is established to detect the silence condition of each measuring point to obtain the communication state discrimination result of each measuring point.

[0125] In the embodiment, the communication state discrimination model includes a first communication state discrimination model and a second communication state discrimination model.

[0126] In the embodiment, the first communication state discrimination model is used to determine whether the communication state discrimination result is communication interruption or communication normality, and the measuring point A is taken as an example, which includes the following steps.

[0127] If the current measuring point satisfies the following condition: the communication state discrimination result of the current measuring point is determined as communication interruption.

[0128] If the current measuring point does not satisfy the following condition: the communication state discrimination result of the current measuring point is determined as communication normality.

[0129] Wherein, represents the actual measurement value of the measuring point A at the current time T i , represents the actual measurement value of the measuring point A at the previous time T i-a , , is a positive integer and satisfies .

[0130] In the embodiment, the second communication state discrimination model is used to detect the silence condition of each measuring point to obtain the communication state discrimination result of each measuring point, which includes the following steps.

[0131] Based on the measuring point data corresponding to each measuring point, the state jump frequency and the duration, the data refresh period of the state quantity and the numerical fluctuation amplitude of the analog quantity are monitored in a preset time period.

[0132] The switch quantity silence factor is calculated according to the actual state jump frequency and the rated state jump frequency, wherein the switch quantity silence factor = 1-(actual state jump frequency / rated state jump frequency).

[0133] The state quantity silence factor is calculated according to the actual data refresh period and the rated data refresh period, wherein the state quantity silence factor = 1-(actual data refresh period / rated data refresh period).

[0134] The analog quantity silence factor is calculated according to the actual numerical fluctuation amplitude and the rated numerical fluctuation amplitude, wherein the analog quantity silence factor = 1-(actual numerical fluctuation amplitude / rated numerical fluctuation amplitude).

[0135] The comprehensive silence index is calculated based on the silence factors of switching quantities, state quantities, and analog quantities, wherein the comprehensive silence index = (silence factor of switching quantities + silence factor of state quantities + silence factor of analog quantities) / 3.

[0136] When the overall silence index is greater than the preset first threshold and the duration is greater than the preset second threshold, it is determined that the communication is interrupted; otherwise, it is determined that the communication is normal.

[0137] If the communication status is determined to be interrupted, an alarm message will be sent.

[0138] If the communication status determination result is that the communication is normal, then the data quality discrimination model corresponding to all indicator category labels is triggered in parallel to obtain the data quality discrimination result.

[0139] Step 5: Determine the status of the data quality assessment results:

[0140] If the data quality assessment result is abnormal, an alarm message will be pushed, and the standard value of the measurement point data of the current abnormal measurement point will be corrected. The database containing the working condition related parameter labels will be updated based on the corrected standard value of the measurement point data of the current abnormal measurement point.

[0141] In this embodiment, the standard value of the measurement point data of the current abnormal measurement point is corrected, including:

[0142] Based on the time cycle of the unit's most recent major overhaul and the time interval from the most recent maintenance and repair to the current moment, the standard values ​​of the current abnormal measuring points are corrected to obtain the corrected standard values ​​of the current abnormal measuring points.

[0143] In this embodiment, the standard value of the corrected measurement point data of the current abnormal measurement point is expressed as:

[0144] ;

[0145] In the formula, This represents the standard value of the measurement point data for the current abnormal measurement point after correction. This represents the standard value of the measurement data at the current abnormal measurement point. This indicates the time interval between the unit's most recent major overhaul. This indicates the time interval between the unit's most recent maintenance and repair and the current moment. Indicates the first The actual measured value of the measuring point at the current moment for each measuring point.

[0146] If the data quality assessment result is normal, a status message is pushed, and the database containing the labels of working condition related parameters is updated based on the current measurement point data of all measurement points.

[0147] Embodiment 3

[0148] Based on the same inventive concept as other embodiments, this embodiment introduces a computer readable storage medium, which stores computer instructions, and the computer instructions are executed by a processor to realize the steps of the method of Embodiment 1 or 2.

[0149] Embodiment 4

[0150] Based on the same inventive concept as other embodiments, this embodiment introduces a computer program product, which includes computer instructions, and the computer instructions are executed by a processor to realize the steps of the method of Embodiment 1 or 2.

[0151] In summary of the above embodiments, the present application realizes full-factor coverage of the equipment operation state by synchronously collecting switch value, state value, analog value and other measured point data, and combining with working condition related parameters such as load rate and water head height, effectively solves the data island problem in traditional monitoring, and improves the accuracy of data correlation analysis; the present application also stores data in a four-level hierarchical structure of equipment-component-index-measured point, and attaches working condition parameter tags, so that the massive heterogeneous data has traceability; the present application innovatively classifies and models the indexes according to the number of measured points, constructs special quality discrimination models for different categories, and reduces the misjudgment rate; the present application identifies communication interruption through silent detection, and innovatively adopts a parallel triggering multi-model discrimination strategy, improves the data repair efficiency in the interrupted state, and at the same time guarantees the freshness of data in the continuous communication state through real-time database updating.

[0152] The present application divides the indexes into L1 type indexes and L2 type indexes according to the number of measured points, and respectively constructs a first data quality discrimination model based on Mahalanobis distance and an optimized MAD model, and constructs a second data quality discrimination model based on working condition related parameter matching and weighted range calculation. The L1 type indexes are analyzed by multiple measured points in cooperation, abnormal data are removed by combining with Mahalanobis distance, and a deterioration coefficient is introduced to dynamically adjust the MAD threshold, thereby reducing the misjudgment rate in the equipment aging stage; the L2 type indexes dynamically allow the range through historical working condition related parameter matching and seasonal correction calculation, thereby solving the false alarm problem in the few-measured-point working condition and improving the reliability of single-measured-point data. Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer usable program code.

[0153] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0154] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0155] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0156] The embodiments of the present application described above are merely intended to illustrate the present application, but not to limit the present application. The skilled in the art can make many modifications and improvements without departing from the spirit and scope of the present application, which should be protected as long as they fall within the scope of the present application and the claims.

Claims

1. A method for real-time condition data quality monitoring and discrimination of hydroelectric equipment, characterized by, The method comprises the following steps: Real-time acquisition of point data corresponding to each measuring point of a hydroelectric power operation device and working condition associated parameters; The point data includes switch quantity, state quantity and analog quantity, and the working condition associated parameters include load rate, water head height and device operation mode; The point data is classified and stored according to the device-component-index-measuring point hierarchical structure, and a database containing working condition associated parameter labels is constructed; According to the number of measuring points associated with each index, the index category labels are obtained through classification processing, and a data quality discrimination model is established; The communication state discrimination result of each measuring point is obtained through the establishment of a communication state discrimination model for the silent condition detection processing of each measuring point; If the communication state discrimination result is communication interruption, an alarm information is pushed; If the communication state discrimination result is normal communication, the data quality discrimination model corresponding to all index category labels is triggered in parallel to obtain the data quality discrimination result; If the data quality discrimination result is abnormal, an alarm information is pushed, the standard value of the point data of the current abnormal measuring point is corrected, and the database containing the working condition associated parameter labels is updated based on the corrected standard value of the point data of the current abnormal measuring point; If the data quality discrimination result is normal, a state information is pushed, and the database containing the working condition associated parameter labels is updated based on the point data of all current measuring points.

2. The method of claim 1, wherein the method further comprises: The communication state discrimination model includes a first communication state discrimination model and a second communication state discrimination model; The first communication state discrimination model is used to determine whether the communication state discrimination result is communication interruption or normal communication, and comprises: If the current measuring point satisfies , it is determined that the communication state discrimination result of the current measuring point is communication interruption. If the current measuring point does not satisfy , it is determined that the communication state determination result of the current measuring point is normal communication. wherein, denotes the current time T i the actual measurement value of the measuring point A, denotes the previous time T i-a the actual measurement value of the measuring point A, , is a positive integer and satisfies ; The second communication state discrimination model is used to obtain the communication state discrimination result of each measuring point through silent condition detection processing of each measuring point, and comprises: Based on the point data corresponding to each measuring point, the state jump frequency and duration of the switch quantity, the data refresh period of the state quantity and the numerical fluctuation amplitude of the analog quantity are monitored within a preset time period; The switch quantity silence factor is calculated according to the actual state jump frequency and the rated state jump frequency of the switch quantity, wherein the switch quantity silence factor = 1 - (actual state jump frequency / rated state jump frequency); The state quantity silence factor is calculated according to the actual data refresh period and the rated data refresh period of the state quantity, wherein the state quantity silence factor = 1 - (actual data refresh period / rated data refresh period); The analog quantity silence factor is calculated according to the actual numerical fluctuation amplitude and the rated numerical fluctuation amplitude of the analog quantity, wherein the analog quantity silence factor = 1 - (actual numerical fluctuation amplitude / rated numerical fluctuation amplitude); The comprehensive silence index is calculated according to the switch quantity silence factor, the state quantity silence factor and the analog quantity silence factor, wherein the comprehensive silence index = (switch quantity silence factor + state quantity silence factor + analog quantity silence factor) / 3; When the comprehensive silence index is greater than a preset first threshold value and the duration is greater than a preset second threshold value, it is determined that the communication is interrupted, otherwise it is determined that the communication is normal.

3. The method of claim 1, wherein the method further comprises: The index category labels include L1 class index and L2 class index, wherein If the number of measuring points associated with each index is greater than 2, it is determined as L1 class index; If the number of measuring points associated with each index is less than or equal to 2, the index is determined as an L2-type index.

4. The method of claim 3, wherein, A first data quality discrimination model and a second data quality discrimination model are respectively established according to the L1-type index and the L2-type index; The first data quality discrimination model comprises a distance analysis module and a MAD optimization module; The distance analysis module is configured to calculate Mahalanobis distance of data distribution between measuring points according to measuring point data corresponding to all measuring points under a certain type of index, and screen out abnormal measuring points; The MAD optimization module is configured to introduce a degradation coefficient into a median absolute deviation (MAD) to calculate a data quality discrimination threshold of the measuring point data corresponding to all measuring points under the certain type of index, and discriminate abnormal measuring points exceeding the index data quality discrimination threshold as data quality abnormality to generate a data quality discrimination result; The second data quality discrimination model comprises a working condition parameter extraction and matching module and a weighted range calculation module; The working condition parameter extraction and matching module is configured to extract working condition associated parameters of a current measuring point from a database comprising working condition associated parameter labels, and match a historical database comprising working condition associated parameter labels to obtain a normal value range and a standard deviation of the working condition associated parameters of the current measuring point; The weighted range calculation module is configured to calculate a deviation allowable range of the working condition associated parameters of the current measuring point according to a seasonal correction coefficient and the standard deviation of the working condition associated parameters of the current measuring point, and generate a data quality discrimination result according to the normal value range of the working condition associated parameters of the current measuring point and the deviation allowable range of the working condition associated parameters of the current measuring point.

5. The method of claim 4, wherein, The Mahalanobis distance of data distribution between measuring points is calculated according to measuring point data corresponding to all measuring points under a certain type of index, and abnormal measuring points are screened out, comprising: Z-score standardization is performed on actual measurement values of the measuring point data corresponding to all measuring points under a certain type of index to obtain standardized measuring point data corresponding to all measuring points; A measuring point inter-matrix is constructed based on the standardized measuring point data corresponding to all measuring points to calculate Mahalanobis distance values of all measuring points; If the Mahalanobis distance value of the current measuring point data is greater than a preset distance value alarm threshold, the current measuring point is marked as an abnormal measuring point.

6. The method of claim 4, wherein the method further comprises: A degradation coefficient is introduced into a median absolute deviation (MAD) to calculate a data quality discrimination threshold of the measuring point data corresponding to all measuring points under a certain type of index, and abnormal measuring points exceeding the index data quality discrimination threshold are discriminated as data quality abnormality to generate a data quality discrimination result, comprising: A degradation coefficient is calculated based on a device running time and a preset device service life, and a calculation formula of the degradation coefficient is represented as: ； In the formula, This represents the unit's degradation coefficient. This represents the unit's degradation coefficient in the previous major overhaul cycle for a certain type of indicator. The maximum absolute value, Indicates the first appearance The corresponding time, This represents the unit deterioration coefficient at the time of resumption of production in the previous major overhaul cycle, representing a certain type of indicator. initial value, Indicates the first appearance The corresponding time, T represents the overhaul cycle. , They are respectively Weighting coefficients; A sliding time window of a measuring point under a device health state is extracted, and a median absolute deviation (MAD) of the measuring point data corresponding to all measuring points under a certain type of index in the sliding time window is calculated; The degradation coefficient is introduced into the median absolute deviation (MAD) to calculate a data quality discrimination threshold of the measuring point data corresponding to all measuring points under a certain type of index, and a calculation formula of the data quality discrimination threshold is represented as: ； In the formula, represents a data quality discrimination threshold value, represents the median absolute deviation of the data of the measuring point corresponding to the measuring point. Real-time quality discrimination is performed on the measuring point data corresponding to all measuring points under a certain type of index, comprising: When the absolute deviation of the median of the measurement point data corresponding to a certain measurement point under a certain type of index is greater than the data quality discrimination threshold, it is determined that the data quality discrimination result is abnormal. When the absolute deviation of the median of the measurement point data corresponding to a certain measurement point under a certain type of index is less than or equal to the data quality discrimination threshold, it is determined that the data quality discrimination result is normal.

7. The method of claim 4, wherein the method further comprises: extracting the working condition correlation parameter of the current measurement point from the database containing the working condition correlation parameter label and matching the historical database containing the working condition correlation parameter label to obtain the normal value range and standard deviation of the working condition correlation parameter of the current measurement point, including: extracting the working condition correlation parameter of the current measurement point from the database containing the working condition correlation parameter label; performing similarity matching in the historical database containing the working condition correlation parameter based on a preset matching rule according to the working condition correlation parameter of the current measurement point, and screening the historical working condition correlation parameter of the current measurement point; calculating the normal value range and standard deviation of the working condition correlation parameter of the current measurement point according to the historical working condition correlation parameter of the current measurement point.

8. The method of claim 7, wherein the method further comprises: calculating the deviation allowable range of the working condition correlation parameter of the current measurement point according to the seasonal correction coefficient and the standard deviation of the working condition correlation parameter of the current measurement point, and generating the data quality discrimination result according to the normal value range of the working condition correlation parameter of the current measurement point and the deviation allowable range of the working condition correlation parameter of the current measurement point, including: calculating the seasonal correction coefficient according to the difference between the current environmental temperature and the standard temperature; calculating the deviation allowable range of the working condition correlation parameter of the current measurement point according to the standard deviation of the working condition correlation parameter of the current measurement point and the seasonal correction coefficient; if the actual measurement value of the working condition correlation parameter of the current measurement point is within the normal value range, it is determined that the data quality discrimination result is normal; if the actual measurement value of the working condition correlation parameter of the current measurement point is within the deviation allowable range but exceeds the normal value range or the actual measurement value exceeds the deviation allowable range, it is determined that the data quality discrimination result is abnormal.

9. The method of claim 1, wherein, correcting the standard value of the measurement point data of the current abnormal measurement point, including: correcting the standard value of the measurement point data of the current abnormal measurement point based on the actual measurement value of all measurement points at the current time from the last time of unit overhaul, the time interval from the last time of unit overhaul to the current time, and the last time of unit overhaul to obtain the corrected standard value of the measurement point data of the current abnormal measurement point.

10. The method of claim 9, wherein, The corrected standard value of the measurement point data of the current abnormal measurement point is represented as: ； In the formula, represents the standard value of the measuring point data of the current abnormal measuring point after correction, represents the standard value of the measuring point data of the current abnormal measuring point, represents the time interval from the last time of the unit overhaul to the current time, represents the time interval from the last time of the unit overhaul to the current time, represents the actual measurement value of the measuring point data at the current time corresponding to the th measuring point.

Images