Method for constructing state feature database for operation monitoring of hydroelectric generating set

Abstract

The application relates to the technical field of database construction, in particular to a state feature database construction method for operation monitoring of a hydroelectric generating set, which comprises the following steps: acquiring all types of operation monitoring data at all moments in each operation process of the hydroelectric generating set, determining the state feature and the working condition feature of the operation monitoring data in each operation process of each type; analyzing the similarity between two types of operation monitoring data in each operation process, and determining the correlation between the two types of operation monitoring data; classifying all types of operation monitoring data based on the correlation, constructing the storage priority of each data time sequence of each type and performing block storage, and obtaining a state feature database for operation monitoring of the hydroelectric generating set. The application aims to optimize the storage of operation monitoring data of the hydroelectric generating set, improve the data storage efficiency, and improve the data retrieval accuracy and efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of database construction, in particular to a state feature database construction method for operation monitoring of a hydroelectric generating set. BACKGROUND

[0002] With the continuous advancement of industrialization, the demand for electricity is rapidly growing, and large-scale renewable energy such as photovoltaic and wind power has been put into operation. The power system has an increasingly urgent need for flexible and efficient energy forms. In particular, in terms of load regulation, the power system urgently needs to achieve effective peak shaving and valley filling operations. In this context, as an engineering facility with load regulation capacity, hydropower stations have become an indispensable regulation tool in the power system. As the core energy conversion equipment in the hydropower energy system, the stable and safe operation of the hydroelectric generating set is of great importance. Once a fault or abnormal operation occurs, it may not only affect the power generation efficiency, but also cause damage to the equipment, and even cause more serious accidents. Therefore, ensuring the normal operation of the hydroelectric generating set is of great significance to the safe, stable and efficient operation of the power system.

[0003] Generally, the state of the key components and the overall equipment of the hydroelectric generating set is quantitatively evaluated by establishing a database, and support is provided for fault diagnosis. Traditional databases often use fixed storage methods, which cannot be dynamically adjusted according to the real-time value of the data. This means that even if some data is not important at a particular moment, it will still occupy storage space, causing a large amount of redundant storage. This inefficient storage will cause the capacity of the database to rapidly expand, reduce the efficiency of data access, and affect the overall response capability of the system. SUMMARY

[0004] In view of the above, it is necessary to provide a state feature database construction method for operation monitoring of a hydroelectric generating set to solve the above problems.

[0005] One embodiment of the present application provides a state feature database construction method for operation monitoring of a hydroelectric generating set, which comprises:

[0006] Obtaining all types of operation monitoring data at all times during each operation process of the hydroelectric generating set to form each data time sequence of each type;

[0007] Analyzing the distribution characteristics of the elements in each data time sequence of each type and the distance characteristics between elements with the same value to determine the state characteristic of each data time sequence of each type; analyzing the similarity characteristics between each data time sequence of each type and the remaining data time sequences, and combining the differences between the state characteristics to determine the working condition characteristics of each data time sequence of each type;

[0008] determine the relevance between the running monitoring data of any two types based on the similarity between the data in each running process and the distribution of the working condition characteristics of the data; classify all types of running monitoring data based on the relevance to obtain several correlation categories;

[0009] Based on the relevance between the running monitoring data of each type and the remaining types in each correlation category, and the working condition characteristics of each data time sequence of each type, the storage priority of each data time sequence of each type is constructed and stored in blocks to obtain the state feature database of the hydroelectric generating set running monitoring.

[0010] Preferably, the state feature of each data time sequence of each type is determined, specifically:

[0011] For each data time sequence of each type, the minimum time interval between elements with the same numerical value as each element is calculated as the minimum time distance of each element, and the cumulative sum of the minimum time distances obtained by calculating all elements is calculated. The cumulative sum and the element confusion degree of each data time sequence are positively fused to obtain the state feature of each data time sequence of each type.

[0012] Preferably, the working condition characteristic of each data time sequence of each type is determined, and the specific process is:

[0013] The distance feature between the data time sequence of two running processes of the same type is analyzed, and the approximation between the data time sequence of the two running processes is determined based on the difference in the state feature.

[0014] For each data time sequence of each type, the numerical proportion of the maximum approximation in all approximations obtained except the maximum approximation is obtained as the working condition characteristic of each data time sequence of each type.

[0015] Preferably, the approximation between the data time sequence of the two running processes is:

[0016] For the same type, the distance measure between any two data time sequences and the state feature difference are positively fused, and the negative correlation mapping result obtained by positive fusion is taken as the approximation between the two data time sequences.

[0017] Preferably, the relevance between the running monitoring data of any two types is determined, specifically:

[0018] The common access frequency and the maximum access frequency of the running monitoring data of the two types are obtained; the ratio between the common access frequency and the maximum access frequency is calculated;

[0019] obtaining the similarity between the data time series of each operation of the two types of operation monitoring data, calculating the average of the working condition characteristics of the data time series of each operation of the two types of operation monitoring data;

[0020] obtaining the product of the similarity and the average of each operation, accumulating the products of all operations, multiplying the accumulated products by the ratio, and obtaining the correlation between the two types of operation monitoring data.

[0021] Preferably, the specific process of obtaining a plurality of correlation categories is as follows:

[0022] randomly selecting one type of operation monitoring data, taking all other types of operation monitoring data having a correlation greater than a preset threshold with the selected type of operation monitoring data as a correlation category, continuing to randomly select the remaining types of monitoring data, repeating the process, and obtaining all correlation categories.

[0023] Preferably, the specific process of obtaining a plurality of correlation categories is as follows:

[0024] randomly selecting one type of operation monitoring data, taking all other types of operation monitoring data having a correlation greater than a preset threshold with the selected type of operation monitoring data as a correlation category, and the correlation between any two types of operation monitoring data in a correlation category is greater than a preset threshold; continuing to randomly select the remaining types of monitoring data, repeating the process, and obtaining all correlation categories.

[0025] Preferably, the preset threshold is specifically the average of the correlation between all types of operation monitoring data.

[0026] Preferably, the specific process of constructing the storage priority of each data time series of each type is as follows:

[0027] For each type of operation monitoring data, calculating the accumulated sum of the correlation between each type and all other types of operation monitoring data in the correlation category to which the type belongs, and performing forward fusion of the obtained accumulated sum and the working condition characteristics of each data time series of each type to obtain the storage priority of each data time series of each type.

[0028] Preferably, the specific process of performing block storage is to use an encoding algorithm to perform block storage by taking the storage priority of each data time series of each type as an encoding priority.

[0029] The present application has at least the following beneficial effects:

[0030] The application constructs a state feature database for hydropower unit operation monitoring, significantly optimizes the storage and retrieval performance of hydropower unit operation monitoring data, and specifically, based on state feature and working condition feature dynamic quantitative data value, the value of each data is dynamically evaluated through state feature and working condition feature, and data with large information quantity is preferentially stored, so that useless data can be avoided to occupy storage resources. This method helps to identify and retain key data, reduces redundant storage, and improves the utilization efficiency of storage space; wherein, the working condition feature helps to capture key data of abnormal working conditions, which can help to discover potential risks of equipment failure in time, and through monitoring and analysis of these working condition features, early fault warning and state evaluation can be performed, and through the correlation blocking strategy, the data is stored in blocks according to the storage priority, which can improve the efficiency and security of storage, and reduce unnecessary redundant operations during data retrieval; the storage priority can ensure that data with high correlation and high working condition feature can be quickly acquired, reducing retrieval time; the strongly correlated data is clustered and stored in blocks to ensure fast retrieval of related data. This strategy effectively reduces cross-slice operations during query, improves data access speed and efficiency, especially when handling a large amount of data, it can reduce query delay and improve response speed; finally, the state feature database for hydropower unit operation monitoring is constructed, the data storage and retrieval performance is optimized, and the storage capacity, retrieval efficiency and state evaluation accuracy of the data are improved. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 A flowchart of the state feature database construction method for hydropower unit operation monitoring provided by the application is provided.

[0032] Figure 2 A flowchart of the storage priority acquisition provided by the application is provided. DETAILED DESCRIPTION

[0033] In the description of the embodiments of the application, the words "exemplary", "or", "for example" are used to mean as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of the application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the use of "exemplary", "or", "for example" and the like is intended to present the relevant concept in a specific manner.

[0034] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the application belongs. The terms used in the specification of the application are only for the purpose of describing the specific embodiments and are not intended to limit the application.

[0035] It should be noted that the terms "first", "second" in the present application and its drawings are used to distinguish similar objects, rather than to describe a specific order or sequence. The method disclosed in the embodiments of the present application or the method shown in the flowchart includes one or more steps for implementing the method, and the execution order of the steps can be interchanged with each other without departing from the scope of the present application, and some steps can also be deleted.

[0036] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0037] The present application proposes a state feature database construction method for hydroelectric generating set operation monitoring, which is applied to the technical field of database construction, and reference is made to the attached Figure 1 , the method comprises the following steps:

[0038] S1: Obtain all types of operation monitoring data at all times during each operation process of the hydroelectric generating set to form each data time sequence of each type.

[0039] The present application obtains data related to the operation of the hydroelectric generating set through a sensor monitoring network, a hydroelectric plant monitoring system, and externally accessed hydrology and meteorology, which includes unit body operation data: bearing temperature, unit speed, axial displacement, lubricating oil pressure, lubricating oil level, bearing runout, rack vibration, etc.; electrical operation data: stator / rotor current, voltage, power, power factor, frequency, partial discharge (PD), generator air gap monitoring, stator winding temperature distribution, etc.; hydraulic system data: water head (static water head, dynamic water head), flow, volute pressure, tail pipe pressure fluctuation, guide vane opening, paddle angle, etc.; environmental and working condition data: reservoir water level, downstream water level, water temperature, silt content, etc.

[0040] The operation monitoring data collected during each operation process of the hydroelectric generating set is stored to form an original database in the operation monitoring of the hydroelectric generating set. At the same time, access log records of the original database are obtained, so that when the fault and state analysis of the hydroelectric generating set is performed, the overall state of the equipment and fault diagnosis can be determined by comparing the matching condition of the operation monitoring data and the database.

[0041] S2: Analyze the distribution characteristics of the elements in each data time sequence of each type and the distance characteristics between the same numerical elements to determine the state characteristic of each data time sequence of each type; analyze the similarity characteristics between each data time sequence of each type and the remaining data time sequences, and determine the working condition characteristics of each data time sequence of each type in combination with the differences between the state characteristics.

[0042] In constructing the state feature database of the hydroelectric generating set, the original database storage is faced with core problems such as data explosion, read-write performance bottleneck, uneven value density, difficulty in multi-source heterogeneous management and life cycle loss, which leads to low efficiency and accuracy of the database in providing decision assistance for fault diagnosis and state maintenance of the hydroelectric generating set. Therefore, in the present application, the hydroelectric generating set operation monitoring data is further analyzed, the unit feature data is extracted, the data storage of the original database is optimized, the construction of the state feature database is completed, the data retrieval accuracy and efficiency are improved, and the overall state evaluation and fault diagnosis of the equipment are assisted.

[0043] All types of operation monitoring data at all times during each operation process of the hydroelectric generating set are combined to form each data time sequence of each type, as an example, the bearing runout data collected at all times during each operation process of the hydroelectric generating set are combined to form a data time sequence of the bearing runout.

[0044] Firstly, the state feature of each data time sequence of each type is reflected by the repetition condition of each data time sequence of each type, specifically: for each data time sequence of each type, the minimum time interval between elements with the same numerical value is calculated as the minimum time distance of each element, the cumulative sum of the minimum time distances obtained by all elements is calculated, and the cumulative sum is positively fused with the element confusion degree of each data time sequence to obtain the state feature of each data time sequence of each type. It should be noted that if there is no other element with the same numerical value in the data time sequence, the minimum time interval between the element with the closest numerical value and the element is taken as the minimum time interval of the element.

[0045] In the present embodiment, the state feature of each data time sequence of each type is denoted as , and the formula form is: In the formula, Y represents the number of elements of each data time sequence of each type; represents the information entropy of each data time sequence of each type, which is used to measure the confusion degree of the sequence elements, the greater the information entropy, the higher the uncertainty of the data value distribution in the data time sequence, the more the information amount, and the greater the state feature; represents the minimum time distance between the i-th element and the element with the same data value, as an example, the obtained data time sequence is , wherein the minimum time distance of the element is the time interval between the element with the sequence number 1 and the element with the sequence number 3; the greater the cumulative sum of the minimum time distances, the smaller the possibility of data repetition, the smaller the repetition, and the more the information amount.

[0046] It should be understood that the greater the accumulation of the minimum time distance, the higher the repeatability between the data points, and the more concentrated the distribution of the data points with the same value, indicating that the change range of the data is small, the information amount is small, and the state characteristic value is also low. At this time, the reducibility of the data is greater, which means that the data storage priority in the time sequence is lower; on the contrary, the greater the state characteristic, the lower the repeatability of the data points, the more key information of the equipment operation condition, the greater the information amount, and the smaller the reducibility of the data, and the storage priority should be higher.

[0047] Further, the data change characteristics between each data time sequence of each type and the rest of the data time sequence are analyzed to obtain the working condition characteristic of each data time sequence of each type, which is used to identify working condition data in the event of failure or abnormal working condition of the hydroelectric generating set during operation. Specifically: for the same type, the distance between any two data time sequences is measured and the state characteristic difference is positively fused, and the negative correlation mapping result obtained by the positive fusion is taken as the approximation between the two data time sequences; for each data time sequence of each type, the numerical proportion of the maximum approximation in all approximations obtained except the maximum approximation is obtained as the working condition characteristic of each data time sequence of each type.

[0048] In this embodiment, the approximation between data time sequence A and data time sequence B of the same type is denoted as , and the formula form is: In the formula, , respectively represent the state characteristic of data time sequence A and data time sequence B of each type, represents the state characteristic difference between data time sequence A and data time sequence B, and represents the difference in information amount between the two data time sequences. The greater the value, the smaller the approximation; represents the DTW distance function; represents a preset parameter greater than zero to prevent the denominator from being 0. In this embodiment, the value is 0.001, which can be adjusted by the implementer according to the actual situation.

[0049] The working condition characteristic of data time sequence A of each type is denoted as , and the formula form is: ; In the formula, represents the maximum approximation obtained by data time sequence A in the same type; represents the approximation between data time sequence A and the nth data time sequence in the same type except the data time sequence corresponding to the maximum approximation; represents the number of all data time sequences of each type.

[0050] It should be understood that the greater the working condition characteristic is, the more the data time series conforms to the characteristics of the characteristic working condition, and the greater the contingency and regularity. Contingency refers to a situation that occurs less frequently in the historical operation of the hydroelectric generating set, and regularity refers to the fact that the same type of data time series contains similar information in different operation processes, which may be manifested as typical characteristics of a device when a certain fault occurs. Therefore, the greater the working condition characteristic is, the higher the possibility that the data time series contains characteristic working condition data; and the smaller the working condition characteristic is, the lower the possibility that the data time series contains characteristic working condition data.

[0051] S3: determining the relevance between the two types of operation monitoring data based on the similarity between the data in each operation process and the distribution of the working condition characteristic; and classifying all types of operation monitoring data based on the relevance to obtain a plurality of association categories.

[0052] If data storage is only based on working condition characteristics, although the data storage efficiency and security under the characteristic working condition can be effectively improved, this method ignores the change relevance between the data under the same working condition. This may result in low efficiency when searching for data in the database. Therefore, the present application constructs the relevance between different types of operation monitoring data to improve the search efficiency.

[0053] In the operation monitoring process of the hydroelectric generating set, the changes in different types of operation monitoring data often have certain relevance. For example, when the water head decreases, the guide vane opening needs to be increased to maintain the same power, which may cause vortex band vibration in the draft tube and increase the harmonic component of the stator current. Based on this, the present application constructs the relevance between any two types of operation monitoring data based on the change conditions between different types of operation monitoring data in the original database and the access conditions of different types of data.

[0054] Specifically, the common access frequency and the maximum access frequency of any two types of operation monitoring data in all operation processes are obtained; the ratio between the common access frequency and the maximum access frequency is calculated; the similarity between the data time series obtained by any two types of operation monitoring data in each operation process is obtained; the average value of the working condition characteristic of the data time series obtained by any two types of operation monitoring data in each operation process is calculated; the product of the similarity and the average value of each operation process is obtained, the products obtained in all operation processes are accumulated, and the product is multiplied by the ratio to obtain the relevance between the two types of operation monitoring data. In this embodiment, the similarity between the sequences is calculated using the Pearson correlation coefficient.

[0055] It should be understood that the greater the product of the similarity degree of each run and the average value means the stronger the change relevance of the two types of data under the same characteristic working condition, and the greater the relevance. The higher the relevance between the two types of operation monitoring data, the greater the relevance of the two types of data when the equipment is abnormal in the process of evaluating or diagnosing the state of the hydroelectric generating set equipment. Therefore, it is more necessary to store and simultaneously refer to the two types of data as the same type, thereby helping to improve the accuracy and efficiency of the state evaluation and fault diagnosis of the hydroelectric generating set.

[0056] The relevance between different types of operation monitoring data is obtained, and different types of operation monitoring data are adaptively clustered through the relevance of the operation monitoring data, so that the collected different types of operation monitoring data are divided into different relevance categories, so that the operation monitoring data of different relevance categories are stored in a database block, i.e., the data of the same relevance category in the database is stored on the same server, i.e., the data of different relevance categories is stored on different servers, so that when the database is used to evaluate or diagnose the state of the hydroelectric generating set equipment, the probability of cross-plate data query is reduced, and the query and writing efficiency and accuracy of the database are improved.

[0057] After normalization processing of all the obtained relevance, statistical analysis is performed to obtain a relevance data set, and the average value of all data in the relevance data set is taken as a preset threshold, and the method for obtaining the preset threshold can be adjusted by the implementer.

[0058] In this embodiment, one type of operation monitoring data is randomly selected, and the Qth type of operation monitoring data is taken as an example. The relevance between the Qth type of operation monitoring data and the remaining types of operation monitoring data is obtained. If the relevance is greater than or equal to a preset threshold, it is judged that the Qth type of operation monitoring data has a high relevance with the data of this category, and can be regarded as the same relevance category, otherwise, it is not the same relevance data. The above steps are repeated until the number of types of data in the relevance category to which the Qth type of operation monitoring data belongs does not change. Then, a type of data that has not been divided is randomly selected, and the above steps are repeated until the number of types of data in all relevance categories does not change.

[0059] As an example, there are , , , , 5 types of operation monitoring data, and one type is randomly selected, such as , the relevance between and the remaining four types of operation monitoring data is obtained. If the relevance is greater than the preset threshold, it belongs to the same relevance category as . and If they belong to the same category, then they become , , , Then, choose any of the remaining categories, for example, select... , obtain The correlation with the other two types of operational monitoring data; if the correlation is greater than a preset threshold, then... If they belong to the same category, they are not in the same category. Repeat this process until all categories of data no longer change.

[0060] In another embodiment, when obtaining the associated category, it is also necessary to satisfy the condition that the correlation between any two types of operation monitoring data in the same associated category is greater than a preset threshold.

[0061] It should be noted that if there are still discrete operational monitoring data of a certain type at this time, the average correlation between the operational monitoring data of that type and all types of data in each related category is obtained, and the related category with the largest average correlation is selected as the category to which the operational monitoring data of that type belongs.

[0062] S4: Based on the correlation between each type of operational monitoring data and the other types in each associated category, and the operating condition characteristics of each data time series of each type, construct the storage priority of each data time series of each type and perform block storage to obtain the status characteristic database of hydropower unit operation monitoring.

[0063] The above steps adaptively divide different types of operational monitoring data from hydropower units into multiple related categories. Each related category is stored in a different block in the database, completing the block storage of the database. This improves data storage efficiency and security while also enhancing data retrieval efficiency and accuracy.

[0064] In this application, the storage priority of each data time series is constructed based on the correlation between each type of operational monitoring data and the other types of operational monitoring data in each associated category, as well as the operating condition characteristics of each data time series of operational monitoring data. The formula is as follows: In the formula: F represents the storage priority of each data time series of each type; Z represents the operating condition characteristics of each data time series of each type; S represents the number of data types in the associated category to which each data time series of each type belongs; This indicates the correlation between the operational monitoring data of the time series type and the operational monitoring data of the s-th type in the associated category.

[0065] The flowchart for obtaining storage priority is as follows: Figure 2 As shown.

[0066] It should be understood that the greater the storage priority, the more likely the data time sequence contains the characteristic information of the operating condition of the hydroelectric generating unit, and the smaller the storage priority, the less likely the data time sequence contains the characteristic information of the operating condition.

[0067] The storage priority of each data time sequence in each type of operating monitoring data is obtained, and the storage priority is encoded as an encoding priority in Huffman coding, so as to complete the block storage of the database, wherein the Huffman coding technology is a known means, and will not be described here again. The above steps are repeated until each type of operating monitoring data is stored, the block optimization storage of the hydroelectric generating unit operating monitoring original database is completed, and the state characteristic database of the hydroelectric generating unit operating monitoring is obtained.

[0068] The flowcharts and block diagrams in the drawings show the possible implementation architecture, function and operation of the system, method and computer program product according to the embodiments of the present application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment or a part of code containing one or more executable instructions for implementing a specified logic function. In some alternative implementations, the functions annotated in the blocks can also occur in an order different from that annotated in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and sometimes they can be executed in reverse order, depending on the functions involved. In the description corresponding to the flowcharts and block diagrams in the drawings, the operations or steps corresponding to different blocks can also occur in an order different from that disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and sometimes they can be executed in reverse order, depending on the functions involved. Each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0069] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for constructing a condition feature database for monitoring the operation of a hydroelectric generating unit, characterized in that, The method comprises the following steps: acquiring all types of operation monitoring data at all times during each operation process of the hydroelectric generating set to form each data time sequence of each type; analyzing the distribution characteristics of elements in each data time sequence of each type and the distance characteristics between elements with the same numerical value to determine the state characteristic of each data time sequence of each type; analyzing the similarity characteristics between each data time sequence of each type and the remaining data time sequences, and combining the differences between the state characteristics to determine the working condition characteristic of each data time sequence of each type; based on the similarity degree between the operation monitoring data of any two types in each operation process, and combining the distribution of the working condition characteristics, determining the correlation between the operation monitoring data of the any two types; classifying all types of operation monitoring data based on the correlation to obtain several correlation categories; based on the correlation between the operation monitoring data of each type and the remaining types in each correlation category, and the working condition characteristic of each data time sequence of each type, constructing the storage priority of each data time sequence of each type and performing block storage to obtain a state feature database of hydroelectric generating set operation monitoring; the determination of the state characteristic of each data time sequence of each type is specifically: for each data time sequence of each type, calculating the minimum time interval between elements with the same numerical value as the minimum time distance of each element, calculating the cumulative sum of the minimum time distances obtained by all elements, and positively fusing the cumulative sum with the element confusion degree of each data time sequence to obtain the state characteristic of each data time sequence of each type; the determination of the working condition characteristic of each data time sequence of each type is specifically: analyze the distance characteristics between the data time sequences of two operation processes of the same type, combine the differences between the state characteristics, and determine the approximation between the data time sequences of the two operation processes; for each data time sequence of each type, obtain the numerical proportion of the maximum approximation in all approximations obtained except the maximum approximation as the working condition characteristic of each data time sequence of each type; the approximation between the data time sequences of the two operation processes is specifically: for the same type, positively fuse the distance measurement between any two data time sequences and the difference between the state characteristics, and map the obtained positively fused negative correlation as the approximation between the any two data time sequences.

2. The method for constructing a condition feature database for operation monitoring of a hydroelectric generating unit according to claim 1, characterized in that, the determination of the correlation between the operation monitoring data of the any two types is specifically: acquiring the common access frequency and the maximum access frequency of the operation monitoring data of the any two types in all operation processes; calculating the ratio between the common access frequency and the maximum access frequency; acquiring the similarity degree between the data time sequences obtained in each operation process of the operation monitoring data of the any two types; calculating the average value of the working condition characteristics of the data time sequences obtained in each operation process of the operation monitoring data of the any two types; The product of the similarity degree of each run and the average value is obtained, the products obtained from all runs are accumulated, and the correlation between the two types of operation monitoring data is obtained by multiplying the ratio.

3. The method for constructing a condition signature database for operation monitoring of a hydroelectric generating unit as claimed in claim 1, characterized in that, The specific process of obtaining the plurality of correlation categories is: Randomly selecting one type of operation monitoring data, taking all other types of operation monitoring data with a correlation greater than a preset threshold with the selected type of operation monitoring data as a correlation category, and continuing to randomly select the remaining types of monitoring data and repeating the process to obtain all correlation categories.

4. The method for constructing a condition feature database for operation monitoring of a hydroelectric generating unit according to claim 1, characterized in that, The specific process of obtaining the plurality of correlation categories is: Randomly selecting one type of operation monitoring data, taking all other types of operation monitoring data with a correlation greater than a preset threshold with the selected type of operation monitoring data as a correlation category, and a correlation between any two types of operation monitoring data in a correlation category is greater than a preset threshold; Continue to randomly select the remaining types of monitoring data, and repeat the process to obtain all correlation categories.

5. The method for constructing a condition signature database for operation monitoring of a hydroelectric generating unit as claimed in claim 4, characterized in that, The preset threshold is specifically an average value of the correlation between any two types of operation monitoring data.

6. The method for constructing a condition feature database for operation monitoring of a hydroelectric generating unit according to claim 1, characterized in that, The specific process of constructing the storage priority of each data time sequence of each type is: For each type of operation monitoring data, the cumulative sum of the correlation between each type and the remaining types of operation monitoring data in the associated category is calculated, the obtained cumulative sum is positively fused with the operating condition characteristics of each data time sequence of each type, and the storage priority of each data time sequence of each type is obtained.

7. The method for constructing a condition feature database for operation monitoring of a hydroelectric generating unit as claimed in claim 1, characterized in that, The specific block storage is achieved by taking the storage priority of each data time sequence of each type as an encoding priority and using an encoding algorithm for block storage.

Images