Health degree assessment method and system based on equipment fault multi-label system, processing equipment and storage medium

Abstract

The invention relates to a health degree evaluation method and system based on an equipment fault multi-label system, processing equipment and a storage medium. The health degree evaluation method comprises the following steps: performing state label labeling on SCADA (Supervisory Control And Data Acquisition) data; detecting the integrity rate of the data measurement value of each measurement point in the wind turbine generator equipment; measuring value over-limit detection is carried out on the measuring points with the recorded data; carrying out transverse comparison detection on the wind turbine generator equipment in the whole field on the non-overrun measuring points; consistency detection is carried out on the data measurement values of the same measurement point of the parts with the plurality of rotational symmetry; the method comprises the following steps: modeling a key large component of wind turbine generator equipment, calculating a theoretical data measurement value of a measurement point of the key large component, and obtaining a single measurement point dynamic threshold overrun condition of the wind turbine generator equipment according to an actual data measurement value of the measurement point; unifying the number of the abnormal labels of each measuring point of the wind turbine generator equipment at each moment, and determining the health degree of the wind turbine generator equipment, and the method can be widely applied to the technical field of new energy equipment anomaly analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormal analysis of new energy equipment, and in particular to a health assessment method, system, processing equipment and storage medium based on a multi-label system for equipment failure. Background Art

[0002] As the trend of wind turbines becoming larger and larger continues, the importance of abnormal detection of wind turbine equipment is increasing. Abnormal operation of equipment will bring serious safety hazards to wind turbines. The actual operation modes of wind turbines are diverse, and it is difficult to fully cover the abnormal state of wind turbines; on the other hand, due to the limited operating points of the equipment, it is difficult to fully cover the operating information through the existing measuring points.

[0003] At present, the existing fault detection methods generally detect a certain type of fault mode, which is difficult to cover the granularity of each moment and each measurement point, and it is also difficult to deeply detect the intrinsic correlation between different faults.

[0004] Therefore, it is crucial to realize the labeling of various basic measurement point anomaly labels based on each moment and each measurement point to improve the granularity of anomaly detection coverage. It can not only effectively detect the quality of data, but also provide a basis for data validity for subsequent unit health assessment and specific mode fault diagnosis models. Summary of the invention

[0005] In view of the above problems, the purpose of the present invention is to provide a health assessment method, system, processing device and storage medium based on a multi-label system for equipment failures, which can improve the granularity of anomaly detection coverage.

[0006] To achieve the above objectives, the present invention adopts the following technical solutions: In a first aspect, a health assessment method based on a multi-label system for equipment failures is provided, which is characterized by comprising: Label SCADA data with status tags; Test the completeness of the data measured at each measuring point in the wind turbine equipment, and determine the measuring points where data is recorded and the data gaps at the measuring points; Perform a measurement value over-limit detection on the measurement points where data is recorded, take the measurement points with valid data measurement values ​​as measurement points that are not over-limited, and obtain the over-limit status of the measurement points of the wind turbine equipment; Conduct lateral comparison inspection of all wind turbine equipment on the measuring points that are not exceeded, and obtain the lateral comparison of single measuring points of wind turbine equipment; Perform consistency detection on the data measurement values ​​of the same measurement point of components with multiple rotational symmetries to obtain the consistency of the component measurement points of the wind turbine equipment; Model the key large components of the wind turbine equipment, calculate the theoretical data measurement values ​​of the measurement points of the key large components, and obtain the dynamic threshold exceeding limit of a single measurement point of the wind turbine equipment based on the actual data measurement values ​​of the measurement points; Based on SCADA data, the number of abnormal labels at each measuring point of the wind turbine equipment at each moment is unified to determine the health of the wind turbine equipment according to the missing data of the measuring points, the exceeding of the measuring points, the horizontal comparison of the single measuring points, the consistency of the component measuring points and the exceeding of the dynamic threshold of the single measuring point.

[0007] Furthermore, the specific process of the completeness rate detection is as follows: At each timestamp, count which measuring points have recorded data at each timestamp, and calculate the percentage of measuring points with recorded data to all measuring points, which is taken as the completeness rate of the data measurement value of the measuring point at the timestamp.

[0008] Furthermore, the measuring point where the data is recorded is detected to have a measurement value exceeding the limit, and the measuring point where the data measurement value is valid is used as a measuring point that has not exceeded the limit, and the measuring point exceeding the limit of the wind turbine equipment is obtained, including: Pre-set the upper and lower limits of the measurement standard value for each measuring point; At each timestamp, for the measuring point where data is recorded, determine whether the data measurement value of the measuring point is valid based on the actual data measurement value and the upper and lower limits of the corresponding measurement standard value that are set in advance, and calculate the percentage of measuring points with valid data measurement values ​​at the timestamp to all measuring points, which is the effectiveness of the measuring point at the timestamp.

[0009] Furthermore, the horizontal comparison test of all wind turbine generator sets is performed on the measurement points that are not exceeded to obtain the horizontal comparison of the single measurement points of the wind turbine generator sets, including: Perform horizontal comparison and detection on each measuring point of all wind turbine equipment at each time stamp, determine the trend value of all wind turbine equipment on site based on statistical methods, and determine the upper and lower limits of the trend value according to a certain fault tolerance range; If the data measurement value of the measuring point is higher than the upper limit of the trend value or lower than the lower limit of the trend value, it is considered that the data measurement value of the measuring point of the wind turbine equipment at the timestamp is too high or too low, and there is an abnormal risk, thereby obtaining a horizontal comparison of a single measuring point of the wind turbine equipment.

[0010] Furthermore, the key large components of the wind turbine equipment are modeled, the theoretical data measurement values ​​of the measurement points of the key large components are calculated, and the single measurement point dynamic threshold exceeding condition of the wind turbine equipment is obtained according to the actual data measurement values ​​of the measurement points, including: Select historical data of a period of time of key components of wind turbine equipment as training data, use operating condition variables as input variables, and use the measurement points of designated key components as output variables to train machine learning or deep learning models; Selecting historical data of another period of time of key large components of wind turbine equipment as test data, and calculating theoretical data measurement values ​​based on input variables and training models; Calculate the residual value based on the actual data measurement values ​​and theoretical data measurement values ​​of the measurement points of the key large components of the wind turbine equipment; The residual thresholds of the measuring points of the key large components of the wind turbine equipment are preset, and the calculated residual values ​​are detected for exceeding the limit according to the preset residual thresholds to obtain the dynamic threshold exceeding the limit of the single measuring point of the wind turbine equipment.

[0011] Furthermore, according to the missing data of the measuring points of the wind turbine equipment, the exceeding of the limit of the measuring points, the horizontal comparison of the single measuring point, the consistency of the component measuring points and the exceeding of the dynamic threshold of the single measuring point, the number of abnormal tags at each measuring point of the wind turbine equipment at each time is unified to determine the health of the wind turbine equipment, including: The health score of each measuring point at that moment is obtained by taking a weighted average of the missing data, over-limit of each measuring point, horizontal comparison of a single measuring point, consistency of component measuring points, and over-limit of a single measuring point dynamic threshold at each measuring point at each moment. Take a weighted average of the health scores of each measuring point to obtain the health of the component level and unit level at each moment; The health of the component level and unit level at each moment is aggregated and counted according to the time dimension to obtain the health of different time granularities such as 10 minutes, hours, and days.

[0012] Furthermore, the method further comprises: Based on the status tags of SCADA data, the obtained detection results and health status are evaluated.

[0013] Furthermore, the status tag based on the SCADA data evaluates the obtained detection results and health, including: If the status label of the SCADA data is fault, and any data in all the test results and health status reaches abnormality within the specified time window before that, it is considered a successful alarm; otherwise, it is considered a missed alarm; If the status tag of the SCADA data does not show abnormality within the specified time window after each test result and health degree abnormality, it is considered to be a false alarm.

[0014] In a second aspect, a health assessment system based on a multi-label system for equipment failure is provided, comprising: Data labeling module, used to label SCADA data with status tags; The completeness detection module is used to detect the completeness of the data measurement value of each measuring point in the wind turbine equipment, and determine the measuring points where data is recorded and the vacancies of the measuring point data; The measurement value over-limit detection module is used to perform measurement value over-limit detection on the measurement points where data is recorded, and to take the measurement points with valid data measurement values ​​as the measurement points that are not over-limited, and to obtain the over-limit situation of the measurement points of the wind turbine equipment; The horizontal comparison detection module is used to perform horizontal comparison detection on all wind turbine equipment at the measurement points that are not exceeded, and obtain the horizontal comparison of the single measurement point of the wind turbine equipment; A consistency detection module is used to perform consistency detection on the data measurement values ​​of the same measurement point of a component with multiple rotational symmetries to obtain the consistency of the component measurement points of the wind turbine equipment; The single-measurement-point dynamic threshold over-limit detection module is used to model the key large components of the wind turbine equipment, calculate the theoretical data measurement values ​​of the measurement points of the key large components, and obtain the single-measurement-point dynamic threshold over-limit situation of the wind turbine equipment based on the actual data measurement values ​​of the measurement points; The health calculation module is used to unify the number of abnormal labels at each measuring point of the wind turbine equipment at each moment based on the SCADA data and according to the missing measurement point data, over-limit measurement point data, horizontal comparison of single measurement points, consistency of component measurement points and over-limit dynamic threshold of single measurement points of the wind turbine equipment, so as to determine the health of the wind turbine equipment.

[0015] According to a third aspect, a processing device is provided, comprising computer program instructions, wherein the computer program instructions, when executed by the processing device, are used to implement the steps corresponding to the above-mentioned health assessment method based on a multi-label system for equipment failures.

[0016] In a fourth aspect, a computer-readable storage medium is provided, on which computer program instructions are stored, wherein the computer program instructions, when executed by a processor, are used to implement the steps corresponding to the above-mentioned health assessment method based on a multi-label system for equipment failures.

[0017] The present invention adopts the above technical solution, which has the following advantages: 1. The present invention can effectively and intuitively reflect the data anomalies and equipment status annotations of each dimension under each timestamp, as well as all fault warnings of the equipment (especially the superposition of multiple types of anomalies or faults at the same time), providing a reference for further warnings of large components and even the calculation and evaluation of the overall health of the equipment.

[0018] 2. The present invention is based on fault or alarm labels, combined with diagnostic labels. For a specific mode fault, it is convenient to trace the data characteristics of the fault through data, and analyze the similarities and differences of the data performance under each fault, so as to provide data basis for the refinement and improvement of the fault diagnosis model.

[0019] 3. Based on fault labels and model warning labels, the present invention provides data basis for model effect evaluation and model iterative optimization.

[0020] 4. The present invention facilitates correlation analysis, causal analysis and root cause tracing among multiple faults, such as a pitch motor failure.

[0021] In summary, the present invention can be widely applied in the technical field of abnormal analysis of new energy equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Throughout the accompanying drawings, the same reference numerals are used to represent the same components. In the accompanying drawings: Figure 1 It is a schematic diagram of a method flow provided by an embodiment of the present invention; Figure 2 is a schematic diagram of a unit status label provided by an embodiment of the present invention; Figure 3 This is a schematic diagram of a measurement point missing label provided by an embodiment of the present invention; Figure 4 This is a schematic diagram of a measuring point over-limit label provided by an embodiment of the present invention; Figure 5 This is a schematic diagram of a single-measurement-point horizontal comparison abnormal label provided by an embodiment of the present invention; Figure 6 This is a schematic diagram of a component measurement point consistency diagnosis label provided by an embodiment of the present invention; Figure 7 This is a schematic diagram of a single-measurement-point dynamic threshold diagnostic label provided by an embodiment of the present invention; Figure 8 is a schematic diagram of a labeled data structure provided by an embodiment of the present invention; Fig. 9 It is a schematic diagram of alarm accuracy provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0024] It should be understood that the terms used herein are only for the purpose of describing specific example embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "said" as used herein may also be meant to include plural forms. The terms "include", "comprise", "contain", and "have" are inclusive, and therefore specify the existence of stated features, steps, operations, elements and / or parts, but do not exclude the existence or addition of one or more other features, steps, operations, elements, parts, and / or combinations thereof. The method steps, processes, and operations described herein are not interpreted as necessarily requiring them to be performed in the specific order described or illustrated, unless the execution order is clearly indicated. It should also be understood that additional or alternative steps may be used.

[0025] Although the terms first, second, third, etc. can be used in the text to describe multiple elements, components, regions, layers and / or sections, these elements, components, regions, layers and / or sections should not be limited by these terms. These terms can only be used to distinguish an element, component, region, layer or section from another region, layer or section. Unless the context clearly indicates, terms such as "first", "second" and other numerical terms do not imply order or sequence when used in the text. Therefore, the first element, component, region, layer or section discussed below can be referred to as the second element, component, region, layer or section without departing from the teaching of the example embodiments.

[0026] At present, the existing fault detection methods generally detect a certain type of specific fault mode, which is difficult to cover the scope of granularity at each moment and each measuring point level, and it is also difficult to deeply detect the intrinsic correlation between different faults. In order to solve the problems of insufficient granularity of abnormal detection coverage and inability to detect correlations between faults in the prior art, an embodiment of the present invention provides a health assessment method based on a multi-label system for equipment faults, including: labeling SCADA data with status labels; detecting the completeness of the data measurement values ​​of each measuring point in the wind turbine equipment, and determining the measuring points where the data is recorded and the data gaps at the measuring points; performing measurement value over-limit detection on the measuring points where the data is recorded, and taking the measuring points with valid data measurement values ​​as the measuring points that are not over-limited, and obtaining the over-limit situation of the measuring points of the wind turbine equipment; performing horizontal comparison detection on the measuring points that are not over-limited for the entire field of wind turbine equipment, and obtaining the horizontal comparison of the single measuring point of the wind turbine equipment. Comparison situation; consistency detection of the data measurement values ​​of the same measuring point of components with multiple rotational symmetries is performed to obtain the consistency of the component measuring points of the wind turbine equipment; modeling of the key large components of the wind turbine equipment, calculating the theoretical data measurement values ​​of the measuring points of the key large components, and according to the actual data measurement values ​​of the measuring points, the dynamic threshold exceeding limit of the single measuring point of the wind turbine equipment is obtained; based on SCADA data, according to the missing data of the measuring point of the wind turbine equipment, the exceeding limit of the measuring point, the horizontal comparison of the single measuring point, the consistency of the component measuring point and the exceeding limit of the dynamic threshold of the single measuring point, the number of abnormal labels of each measuring point of the wind turbine equipment at each moment is unified to determine the health of the wind turbine equipment. The present invention proposes and realizes the labeling of various basic measuring point abnormal labels based on each measuring point at each moment, realizes the detection of data quality, and the health evaluation of the unit, and provides a basis for data validity for the specific mode fault diagnosis model, while realizing the correlation analysis and root cause tracing between faults.

[0027] Example 1 like Figure 1 As shown, this embodiment provides a health assessment method based on a multi-label system for equipment failures, including the following steps: 1) If Figure 2 As shown, the SCADA (Supervisory Control and Data Acquisition) data is labeled with status tags.

[0028] Specifically, the status labels include operating status, shutdown status and other status, among which the operating status includes normal power generation status, limited power generation status and abnormal alarm status, the shutdown status includes non-fault shutdown status and fault shutdown status, and other status includes startup status, standby status and communication interruption status.

[0029] 2) If Figure 3As shown, the completeness rate of the data measurement value of each measuring point in the wind turbine equipment is detected, and the measuring points where data is recorded and the data gaps of the measuring points are determined.

[0030] Specifically, wind turbine equipment includes multiple systems such as blade system, cabin system, tower system, yaw system, main shaft system, pitch system, gearbox system, generator system, converter system, hydraulic system and control system. Each system includes multiple components, and each component includes multiple measuring points. Each site will record data covering multiple systems, multiple components and multiple measuring points into the SCADA system, and the SCADA collection measuring points of each site will be manually determined.

[0031] Specifically, the specific process of completeness detection is as follows: At each timestamp, some data measuring points record data, while some data measuring points do not record data (i.e., data gaps). We count which measuring points record data at each timestamp, and calculate the percentage of measuring points that record data to all measuring points as the completeness rate of the data measurement values ​​of the measuring points at that timestamp.

[0032] 3) If Figure 4 As shown, the measurement points where data is recorded are tested for over-limit values, and the measurement points where data measurement values ​​are valid are regarded as measurement points that are not over-limited, and the over-limit conditions of the measurement points of the wind turbine equipment are obtained, specifically: 3.1) Pre-set the upper and lower limits of the measurement standard value for each measuring point.

[0033] 3.2) At each timestamp, for the measuring points where data is recorded (i.e., non-vacant measuring points), based on the actual data measurement value and the upper and lower limits of the corresponding measurement standard values ​​that are preset, determine whether the data measurement value of the measuring point is valid, and calculate the percentage of measuring points with valid data measurement values ​​at the timestamp to all measuring points, which is the effectiveness of the measuring point at the timestamp.

[0034] Specifically, if the data measurement value of the measuring point is higher than the upper limit of the measurement standard value or lower than the lower limit of the measurement standard value, the data measurement value of the measuring point is considered invalid, and then the measuring point out-of-limit situation of the wind turbine equipment is obtained.

[0035] 4) If Figure 5 As shown in the figure, the horizontal comparison test of the wind turbine equipment in the whole field is carried out on the measuring points that are not exceeded, and the horizontal comparison of the single measuring point of the wind turbine equipment is obtained, which is specifically: 4.1) Perform a horizontal comparison test on each measuring point of all wind turbine equipment at each timestamp, determine the trend value of all wind turbine equipment on site based on statistical methods (such as box plot method, etc.), and determine the upper and lower limits of the trend value according to a certain fault tolerance range.

[0036] Specifically, when all wind turbine equipment is operating normally, at each timestamp, for certain measuring points, the data measurement values ​​of each wind turbine equipment on site are considered to be similar; if it is found that the data measurement values ​​of most wind turbine equipment are close to a certain value, and the data measurement values ​​of individual wind turbine equipment are significantly different from this value, it is considered that the measuring point of the wind turbine equipment is abnormal at this timestamp.

[0037] 4.2) If the data measurement value of the measuring point is higher than the upper limit of the trend value or lower than the lower limit of the trend value, it is considered that the data measurement value of the measuring point of the wind turbine equipment at the timestamp is too high or too low, and there is an abnormal risk, and then the horizontal comparison of the single measuring point of the wind turbine equipment is obtained.

[0038] 5) If Figure 6 As shown, the consistency of the data measurement values ​​of the same measurement point of the component with multiple rotational symmetries is detected to obtain the consistency of the component measurement points of the wind turbine equipment.

[0039] Specifically, components with multiple rotational symmetries include components of the pitch system and generator windings of the generator system. The characteristics of such components are that each component has three or more sub-components with rotational symmetry. For example, a pitch system includes three completely identical pitch motors, and a three-phase winding generator includes three completely identical winding coils.

[0040] Specifically, under normal circumstances, the data measurement values ​​of the same measurement point of a component with multiple rotational symmetries are close. For example, at the same moment, the pitch angles of the three blades are the same, the temperature of the three blades' variable pitch motors is close, and the temperature of the three-phase winding of the generator is close. If the data measurement value deviation is large, it means that there is an abnormality in the sensor measurement or the component itself, and then the consistency of the component measurement points of the wind turbine equipment is obtained.

[0041] Specifically, statistical methods may be used to detect consistency, such as standard deviation.

[0042] 6) If Figure 7 As shown in the figure, combined with the working condition variables, the key large components of the wind turbine equipment are modeled, the theoretical data measurement values ​​of the key large components are calculated, and the single measurement point dynamic threshold exceeding limit of the wind turbine equipment is obtained according to the residual between the theoretical data measurement value and the actual data measurement value of the measurement point, which is specifically: 6.1) Select historical data of key components of wind turbine equipment over a period of time as training data, use operating variables such as wind speed, rotation speed, pitch angle and power as input variables X, and use the specified measurement points of key components (such as main bearing temperature) as output variables y to train the machine learning or deep learning model M.

[0043] Specifically, the key large components of wind turbine equipment include main bearings, generators and gearboxes.

[0044] 6.2) Select historical data of another period of time of key components of wind turbine equipment as test data, and calculate the theoretical data measurement value y_pred based on the input variable X and the training model M as the dynamic benchmark reference value.

[0045] 6.3) Based on the actual data measurement value y_true and the theoretical data measurement value y_pred of the measurement points of the key large components of the wind turbine equipment, calculate the residual value residual.

[0046] Specifically, residual=y_true-y_pred.

[0047] 6.4) Preset the residual threshold of the measuring point of the key large components of the wind turbine equipment, and perform over-limit detection on the calculated residual value residual according to the preset residual threshold to obtain the over-limit situation of the dynamic threshold of a single measuring point of the wind turbine equipment.

[0048] Specifically, if the residual value residual is higher than the residual threshold, it is considered that an abnormality exists at the measurement point at this moment.

[0049] 7) If Figure 8 As shown in the figure, based on SCADA data, according to the missing data of the wind turbine equipment's measuring points, the exceeding of the measuring point limit, the horizontal comparison of the single measuring point, the consistency of the component measuring points and the exceeding of the dynamic threshold of the single measuring point, the number of abnormal labels at each measuring point of the wind turbine equipment at each time is unified to determine the health of the wind turbine equipment, specifically: 7.1) Based on the SCADA data annotated with status labels, the data missing situation, the exceeding limit situation, the horizontal comparison of single measuring point, the consistency of component measuring points and the exceeding limit situation of dynamic threshold of single measuring point are weighted averaged at each measuring point at each moment to obtain the health score of the measuring point at that moment.

[0050] 7.2) Take a weighted average of the health scores of each measuring point to obtain the health of the component level and unit level at each moment.

[0051] 7.3) The health of the component level and unit level at each moment is aggregated and counted according to the time dimension to obtain the health of different time granularities such as 10 minutes, hours, and days.

[0052] Specifically, each step from step 7.1) to step 7.3) generates a data table respectively. All data tables are column storage databases, which can support the functions of adding, deleting and modifying columns, as well as adding, deleting, modifying and querying records. There is also a column specifically recording the time of adding, deleting, modifying and querying. The data table includes the abnormal results of each type of diagnosis at each time stamp of each unit.

[0053] Specifically, the comprehensive diagnosis results at each time stamp of each unit include two parts: label column and variable_list column, which respectively record the abnormal judgment result (0 means no abnormality, 1 means abnormality) and the name of the measurement point involved in the abnormality, such as Figure 2 As shown in the figure, the measuring point names are indicated by measuring point 1, measuring point 2, etc.

[0054] Specifically, the basis for the early warning and analysis of later unit failures includes: pushing out high-risk components and high-risk measurement points of high-risk units or low-risk units based on pre-set rules.

[0055] 8) If Fig. 9 As shown, based on the status label of the SCADA data, the detection results obtained in the above steps 2) to 6) and the health status obtained in step 7) are evaluated.

[0056] Specifically, if the status label of the SCADA data is fault, and any data in all the test results and health status reaches an abnormality within the specified time window (for example, 2 hours) before that, it is considered a successful alarm; otherwise it is considered a missed alarm; if the status label of the SCADA data does not show an abnormality within the specified time window (for example, 2 hours) after the abnormality of each test result and health status, it is considered a false alarm.

[0057] Example 2 This embodiment provides a health evaluation system based on a multi-label system for equipment failures, including: The data labeling module is used to label the SCADA data with status tags.

[0058] The completeness detection module is used to detect the completeness of the data measurement value of each measuring point in the wind turbine equipment, and determine the measuring points where data is recorded and the vacancies of the measuring point data.

[0059] The measurement value over-limit detection module is used to perform measurement value over-limit detection on the measurement points where data is recorded, and to take the measurement points with valid data measurement values ​​as the measurement points that are not over-limited, and to obtain the over-limit status of the measurement points of the wind turbine equipment.

[0060] The horizontal comparison detection module is used to perform horizontal comparison detection of all wind turbine equipment at the measurement points that are not exceeded, and obtain the horizontal comparison of single measurement points of the wind turbine equipment.

[0061] The consistency detection module is used to perform consistency detection on the data measurement values ​​of the same measurement point of a component with multiple rotational symmetries to obtain the consistency of the component measurement points of the wind turbine equipment.

[0062] The single measuring point dynamic threshold over-limit detection module is used to model the key large components of the wind turbine equipment, calculate the theoretical data measurement values ​​of the measuring points of the key large components, and obtain the single measuring point dynamic threshold over-limit situation of the wind turbine equipment based on the actual data measurement values ​​of the measuring points.

[0063] The health calculation module is used to unify the number of abnormal labels at each measuring point of the wind turbine equipment at each moment according to the missing measurement point data, over-limit measurement point conditions, horizontal comparison of single measurement points, consistency of component measurement points and over-limit dynamic threshold conditions of single measurement points of the wind turbine equipment, so as to determine the health of the wind turbine equipment.

[0064] The system provided in this embodiment is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for specific processes and detailed contents, which will not be repeated here.

[0065] Example 3 This embodiment provides a processing device corresponding to the health assessment method based on the equipment failure multi-label system provided in Embodiment 1. The processing device can be applicable to a client processing device, such as a mobile phone, a laptop computer, a tablet computer, a desktop computer, etc., to execute the method of Embodiment 1.

[0066] The processing device includes a processor, a memory, a communication interface and a bus, and the processor, the memory and the communication interface are connected through the bus to complete mutual communication. The memory stores a computer program that can be run on the processing device, and when the processing device runs the computer program, the health assessment method based on the equipment fault multi-label system provided in this embodiment 1 is executed.

[0067] In some implementations, the memory may be a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage.

[0068] In some other implementations, the processor may be a central processing unit (CPU), a digital signal processor (DSP), or other general-purpose processors of various types, which are not limited herein.

[0069] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0070] Those skilled in the art will understand that the structure of the above-mentioned computing device is only a partial structure related to the solution of the present invention, and does not constitute a limitation on the computing device to which the solution of the present invention is applied. The specific computing device may include more or fewer components, or combine certain components, or have a different arrangement of components.

[0071] Example 4 This embodiment provides a computer program product corresponding to the health assessment method based on a multiple-label system for equipment failures provided in Embodiment 1. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing the health assessment method based on a multiple-label system for equipment failures described in Embodiment 1.

[0072] Computer readable storage media can be tangible devices that hold and store instructions used by instruction execution devices. Computer readable storage media can be, for example, but not limited to, electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any combination thereof.

[0073] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0074] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0075] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0076] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0077] The above embodiments are only used to illustrate the present invention, wherein the structure, connection mode and manufacturing process of each component may be changed. Any equivalent transformations and improvements based on the technical solution of the present invention should not be excluded from the protection scope of the present invention.

Claims

1. A health assessment method based on a multi-label system for equipment failures, characterized in that: include: Label SCADA data with status tags; Test the completeness of the data measured at each measuring point in the wind turbine equipment, and determine the measuring points where data is recorded and the data gaps at the measuring points; Perform a measurement value over-limit detection on the measurement points where data is recorded, take the measurement points with valid data measurement values ​​as measurement points that are not over-limited, and obtain the over-limit status of the measurement points of the wind turbine equipment; Conduct lateral comparison inspection of all wind turbine equipment on the measuring points that are not exceeded, and obtain the lateral comparison of single measuring points of wind turbine equipment; Perform consistency detection on the data measurement values ​​of the same measurement point of components with multiple rotational symmetries to obtain the consistency of the component measurement points of the wind turbine equipment; Model the key large components of the wind turbine equipment, calculate the theoretical data measurement values ​​of the measurement points of the key large components, and obtain the dynamic threshold exceeding limit of a single measurement point of the wind turbine equipment based on the actual data measurement values ​​of the measurement points; Based on SCADA data, the number of abnormal labels at each measuring point of the wind turbine equipment at each moment is unified to determine the health of the wind turbine equipment according to the missing data of the measuring points, the exceeding of the measuring points, the horizontal comparison of the single measuring points, the consistency of the component measuring points and the exceeding of the dynamic threshold of the single measuring point.

2. A health assessment method based on a multi-label system for equipment failures as claimed in claim 1, characterized in that: The method of performing over-limit detection on the measuring points where data is recorded, taking the measuring points where the data measurement values ​​are valid as the measuring points that are not over-limited, and obtaining the over-limit situation of the measuring points of the wind turbine equipment includes: Pre-set the upper and lower limits of the measurement standard value for each measuring point; At each timestamp, for the measuring point where data is recorded, determine whether the data measurement value of the measuring point is valid based on the actual data measurement value and the upper and lower limits of the corresponding measurement standard value that are set in advance, and calculate the percentage of measuring points with valid data measurement values ​​at the timestamp to all measuring points, which is the effectiveness of the measuring point at the timestamp.

3. The health evaluation method based on the equipment failure multi-label system according to claim 1 is characterized in that: The horizontal comparison test of all wind turbine generator sets on the measuring points that are not exceeded to obtain the horizontal comparison of the single measuring points of the wind turbine generator sets includes: Perform horizontal comparison and detection on each measuring point of all wind turbine equipment at each time stamp, determine the trend value of all wind turbine equipment on site based on statistical methods, and determine the upper and lower limits of the trend value according to a certain fault tolerance range; If the data measurement value of the measuring point is higher than the upper limit of the trend value or lower than the lower limit of the trend value, it is considered that the data measurement value of the measuring point of the wind turbine equipment at the timestamp is too high or too low, and there is an abnormal risk, thereby obtaining a horizontal comparison of a single measuring point of the wind turbine equipment.

4. The health evaluation method based on the equipment failure multi-label system according to claim 1 is characterized in that: The key large components of the wind turbine equipment are modeled, the theoretical data measurement values ​​of the measurement points of the key large components are calculated, and the single measurement point dynamic threshold exceeding condition of the wind turbine equipment is obtained according to the actual data measurement values ​​of the measurement points, including: Select historical data of a period of time of key components of wind turbine equipment as training data, use operating condition variables as input variables, and use the measurement points of designated key components as output variables to train machine learning or deep learning models; Selecting historical data of another period of time of key large components of wind turbine equipment as test data, and calculating theoretical data measurement values ​​based on input variables and training models; Calculate the residual value based on the actual data measurement values ​​and theoretical data measurement values ​​of the measurement points of the key large components of the wind turbine equipment; The residual thresholds of the measuring points of the key large components of the wind turbine equipment are preset, and the calculated residual values ​​are detected for exceeding the limit according to the preset residual thresholds to obtain the dynamic threshold exceeding the limit of the single measuring point of the wind turbine equipment.

5. The health evaluation method based on the equipment failure multi-label system according to claim 1 is characterized in that: According to the missing data of the measuring points of the wind turbine equipment, the exceeding of the measuring points, the horizontal comparison of the single measuring points, the consistency of the component measuring points and the exceeding of the dynamic threshold of the single measuring point, the number of abnormal tags at each measuring point of the wind turbine equipment at each time is unified to determine the health of the wind turbine equipment, including: The health score of each measuring point at that moment is obtained by taking a weighted average of the missing data, over-limit of each measuring point, horizontal comparison of a single measuring point, consistency of component measuring points, and over-limit of a single measuring point dynamic threshold at each measuring point at each moment. Take a weighted average of the health scores of each measuring point to obtain the health of the component and unit at each moment; The health of the component level and unit level at each moment is aggregated and counted according to the time dimension to obtain the health of different time granularities such as 10 minutes, hours, and days.

6. A health assessment method based on a multi-label system for equipment failures as claimed in claim 1, characterized in that: The method further includes: Based on the status tags of SCADA data, the obtained detection results and health status are evaluated.

7. A health assessment method based on a multi-label system for equipment failures as claimed in claim 6, characterized in that: The status tag based on the SCADA data evaluates the obtained detection results and health, including: If the status label of the SCADA data is fault, and any data in all the test results and health status reaches abnormality within the specified time window before that, it is considered a successful alarm; otherwise, it is considered a missed alarm; If the status tag of the SCADA data does not show abnormality within the specified time window after each test result and health degree abnormality, it is considered to be a false alarm.

8. A health assessment system based on a multi-label system for equipment failures, characterized in that: include: Data labeling module, used to label SCADA data with status tags; The completeness detection module is used to detect the completeness of the data measurement value of each measuring point in the wind turbine equipment, and determine the measuring points where data is recorded and the vacancies of the measuring point data; The measurement value over-limit detection module is used to perform measurement value over-limit detection on the measurement points where data is recorded, and to take the measurement points with valid data measurement values ​​as the measurement points that are not over-limited, and to obtain the over-limit situation of the measurement points of the wind turbine equipment; The horizontal comparison detection module is used to perform horizontal comparison detection on all wind turbine equipment at the measurement points that are not exceeded, and obtain the horizontal comparison of the single measurement point of the wind turbine equipment; A consistency detection module is used to perform consistency detection on the data measurement values ​​of the same measurement point of a component with multiple rotational symmetries to obtain the consistency of the component measurement points of the wind turbine equipment; The single-measurement-point dynamic threshold over-limit detection module is used to model the key large components of the wind turbine equipment, calculate the theoretical data measurement values ​​of the measurement points of the key large components, and obtain the single-measurement-point dynamic threshold over-limit situation of the wind turbine equipment based on the actual data measurement values ​​of the measurement points; The health calculation module is used to unify the number of abnormal labels at each measuring point of the wind turbine equipment at each moment based on the SCADA data and according to the missing measurement point data, over-limit measurement point data, horizontal comparison of single measurement points, consistency of component measurement points and over-limit dynamic threshold of single measurement points of the wind turbine equipment, so as to determine the health of the wind turbine equipment.

9. A processing device, characterized in that: It comprises computer program instructions, wherein when the computer program instructions are executed by a processing device, they are used to implement the steps corresponding to the health assessment method based on a multi-label system for equipment failures according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, wherein the computer program instructions are used to implement the steps corresponding to the health assessment method based on a multi-label system for equipment failures according to any one of claims 1 to 7 when executed by a processor.

Images