A wind turbine generator unit multi-source monitoring data fusion processing method

Abstract

The application discloses a kind of wind generating set multi-source monitoring data fusion processing method, it is related to power generation data processing technical field, for solving the problem of insufficient adaptation ability to the temperature evolution difference under different operating conditions, infrared camera equipment is obtained the average gray value of cabin bottom, synchronous call wireless temperature measurement label and optical fiber temperature characteristic quantity generation operating condition label code;Similar operating condition records are retrieved from historical database based on operating condition label code, extract historical temperature evolution increment and build historical temperature rise feature, calculate similar evolution coefficient in combination with current optical fiber temperature rise variation;Similar evolution coefficient is used to correct smoke temperature detector reference alarm threshold, and fusion concentration coefficient is calculated by fusing smoke temperature and gas concentration response data;Distributed optical fiber temperature rise peak value coordinate and infrared thermal mapping coordinate are respectively obtained, and the final heat source position is determined by spatial deviation fusion positioning, and write in data correlation index table, realize the cross-sensor correlation of multi-source monitoring data.

Description

Technical Field

[0001] This invention relates to the field of power generation data processing technology, and more specifically, to a method for fusion processing of multi-source monitoring data from wind turbine generator sets. Background Technology

[0002] Wind turbine generators operate for extended periods in environments characterized by high altitudes, strong winds, vibration, and complex temperature and humidity. Their nacelles integrate multiple heat-generating components, including generators, converters, cable connection units, and auxiliary control equipment. When local contact problems, insulation aging, load fluctuations, or abnormal heat dissipation occur, a sustained temperature rise often occurs, which may be accompanied by smoke release, changes in gas composition, and alterations in surface thermal radiation characteristics. Therefore, the industry typically employs various monitoring methods, such as infrared monitoring, wireless temperature measurement, and smoke temperature detection, to monitor the generator's operating status and improve the ability to identify early anomalies.

[0003] The existing technology has the following shortcomings:

[0004] Currently, existing monitoring methods typically rely on independent data acquisition and single-channel threshold determination, making it difficult to form dynamic mapping and collaborative verification based on historical operating states among multi-source heterogeneous data. This results in the identification of early gradual temperature rise anomalies depending on a single feature trigger and insufficient adaptability to temperature evolution differences under different operating conditions. Consequently, early warning judgment results are prone to both delayed response and false triggering. Therefore, a multi-source monitoring data fusion processing method for wind turbine generators is proposed.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide a method for fusion processing of multi-source monitoring data of wind turbine generator sets. This method addresses the problems mentioned in the background art by employing historical temperature rise evolution analysis driven by operating condition tag association, dynamic correction of early warning thresholds, and a multi-source thermal feature fusion and positioning mechanism.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for fusion processing of multi-source monitoring data from wind turbine generator sets, comprising the following steps:

[0008] Step S1: Use infrared camera equipment to monitor the wind turbine generator set, collect the average gray value of the bottom of the nacelle and determine whether the data association mechanism is triggered. In the data association mechanism, retrieve the wireless temperature measurement tag and the fiber optic temperature feature quantity and generate the operating condition tag code.

[0009] Step S2: Access the historical database based on the operating condition tag code to retrieve the target operating condition record, obtain the temperature evolution increment of the target operating condition record and calculate the historical temperature rise characteristics, collect the fiber optic temperature rise value at the current moment, and evaluate the similarity evolution coefficient in combination with the historical temperature rise characteristics.

[0010] Step S3: Obtain the baseline alarm threshold of the smoke temperature detector, correct the baseline alarm threshold using the similarity evolution coefficient, detect the gas response data of the smoke sensing unit and calculate the fusion concentration coefficient, and determine whether to trigger the warning signal based on the corrected baseline alarm threshold.

[0011] Step S4: After triggering the warning signal, collect the temperature rise peak coordinates of the distributed optical fiber and the thermal mapping coordinates of the infrared camera. Combine the temperature rise peak coordinates and thermal mapping coordinates to perform deviation fusion positioning and lock the heat source location. Write the heat source location into the data association index table.

[0012] In a preferred embodiment, in step S1, an infrared camera continuously images and collects images of the bottom area of ​​the wind turbine nacelle.

[0013] After the output image of the infrared camera is processed into grayscale, the arithmetic mean of the grayscale values ​​of all pixels in the bottom area of ​​the computer cabin is used to obtain the average grayscale value of the bottom of the cabin.

[0014] The grayscale change is obtained by performing a difference operation on the average grayscale value based on the time series.

[0015] When the amount of grayscale change exceeds the preset trigger threshold, the data association mechanism is triggered.

[0016] Otherwise, the data association mechanism will not be triggered.

[0017] In a preferred embodiment, in step S1, after triggering the data association mechanism, the wireless temperature measurement tag and the fiber optic temperature characteristic quantity are retrieved, wherein the wireless temperature measurement tag is the temperature parameter output by the surface of the converter IGBT device.

[0018] The fiber optic temperature characteristic is the temperature characteristic output by the fiber optic temperature measurement system.

[0019] After normalizing the characteristic quantities of wireless temperature measurement tags and fiber optic temperature measurement, a working condition tag code is constructed.

[0020] In a preferred embodiment, in step S2, a structured matching retrieval is performed from the local historical database of the wind turbine generator set using the operating condition tag code as a unique index key.

[0021] The historical database stores historical operating condition records organized in time series, with each historical operating condition record containing a corresponding historical operating condition tag code;

[0022] During the historical database retrieval process, the working condition tag code generated at the current moment is used as the retrieval benchmark. The corresponding historical working condition tag codes in the historical working condition records are compared and filtered one by one to construct the target working condition record.

[0023] The historical temperature sequence output by the fiber optic temperature measurement system in each target operating condition record is processed by time difference to calculate the temperature evolution increment.

[0024] In a preferred embodiment, in step S2, all temperature evolution increments in the target operating condition record set are statistically aggregated to construct historical temperature rise characteristics.

[0025] At the current moment, the real-time temperature sequence output by the fiber optic temperature measurement system is differentially calculated to obtain the current fiber optic temperature rise value.

[0026] A similarity evolution coefficient is constructed based on historical temperature rise characteristics and the current fiber temperature rise value.

[0027] In a preferred embodiment, in step S3, a baseline alarm threshold is read from the parameter configuration unit of the smoke and temperature detector, the baseline alarm threshold representing the critical level of fire risk triggering by the smoke and temperature detector under standard environmental conditions.

[0028] Based on the similarity evolution coefficient, the baseline alarm threshold is corrected to construct the corrected baseline alarm threshold.

[0029] The gas response data output by the smoke sensing unit is collected. The smoke sensing unit includes a carbon monoxide sensor and a volatile organic compound sensor.

[0030] The carbon monoxide sensor outputs gas response data as the carbon monoxide concentration response value;

[0031] The gas response data output by the volatile organic compound sensor is the volatile organic compound concentration response value.

[0032] In a preferred embodiment, in step S3, each gas response data is normalized by dividing it by the maximum observed value within the historical operating cycle to obtain a normalized gas response value.

[0033] The normalized gas response values ​​are summed to obtain the fusion concentration coefficient;

[0034] The current output temperature value of the smoke temperature detector is collected, and the corrected baseline alarm threshold is jointly determined with the fusion concentration coefficient and the current output temperature value to construct the early warning determination coefficient.

[0035] When the warning judgment coefficient is greater than or equal to 0, the warning signal is triggered and the warning judgment flag X=1 is output.

[0036] When the warning determination coefficient is less than 0, it is determined to be a non-warning state, the warning signal is not triggered, and the warning determination flag X=0 is output.

[0037] In a preferred embodiment, in step S4, the temperature rise peak coordinates are extracted from the fiber optic temperature measurement system, where the temperature rise peak coordinates represent the location of the maximum temperature gradient in the fiber optic temperature measurement channel.

[0038] The thermal mapping coordinates are extracted from the thermal imaging results of the infrared camera. The thermal mapping coordinates are the maximum response points of the gray-scale temperature mapping in the thermal imaging results.

[0039] A unified spatial reference transformation is performed on the temperature rise peak coordinates and heat mapping coordinates to obtain standardized temperature rise peak coordinates and standardized heat mapping coordinates.

[0040] The spatial deviation distance is obtained by calculating the Euclidean distance between the standardized peak temperature rise coordinates and the standardized thermal mapping coordinates.

[0041] In a preferred embodiment, in step S4, a spatial consistency threshold is set. When the spatial deviation distance is less than or equal to the spatial consistency threshold, the standardized temperature rise peak coordinates are taken as the location of the heat source.

[0042] When the spatial deviation distance is greater than the spatial consistency threshold, weighted geometric midpoint fusion is used to perform deviation fusion positioning of the standardized temperature rise peak coordinates and standardized heat mapping coordinates and generate the heat source location.

[0043] The locked heat source location, current operating condition tag code, similarity evolution coefficient, fusion concentration coefficient, and early warning judgment flag are encapsulated into a structured association record and written into the data association index table;

[0044] The data association index table is a structured index database used to store the fusion results of multi-source monitoring data.

[0045] The technical effects and advantages of this invention are as follows:

[0046] This invention acquires the average grayscale value of the bottom of the cabin using an infrared camera. When the grayscale characteristics change abnormally, a data association mechanism is triggered, and a working condition tag code is generated simultaneously by calling the wireless temperature measurement tag and the fiber optic temperature characteristic quantity. Based on the working condition tag code, similar working condition records are retrieved from the historical database, historical temperature evolution increments are extracted, and historical temperature rise characteristics are constructed. At the same time, a similarity evolution coefficient is calculated by combining the current fiber optic temperature rise change. The similarity evolution coefficient is used to dynamically correct the baseline alarm threshold of the smoke temperature detector, and a fusion concentration coefficient is calculated by fusing smoke temperature and gas concentration response data. Early warning determination is achieved by correcting the threshold and fusion judgment value. When an early warning signal is triggered, the temperature rise peak coordinates of the distributed fiber optic cable and the thermal mapping coordinates of the infrared camera are acquired respectively. The final heat source location is determined by the spatial deviation fusion positioning method, and the location is written into the data association index table, thereby realizing cross-sensor association, dynamic correction, and closed-loop index management of multi-source monitoring data. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the implementation of a multi-source monitoring data fusion processing method for wind turbine generator sets according to the present invention.

[0048] Figure 2 This is a schematic diagram illustrating the steps of a multi-source monitoring data fusion processing method for wind turbine generator sets according to the present invention. Detailed Implementation

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

[0050] This invention acquires the average grayscale value of the bottom of the cabin using an infrared camera. When the grayscale characteristics change abnormally, a data association mechanism is triggered, and a working condition tag code is generated simultaneously by calling the wireless temperature measurement tag and the fiber optic temperature characteristic quantity. Based on the working condition tag code, similar working condition records are retrieved from the historical database, historical temperature evolution increments are extracted, and historical temperature rise characteristics are constructed. At the same time, a similarity evolution coefficient is calculated by combining the current fiber optic temperature rise change. The similarity evolution coefficient is used to dynamically correct the baseline alarm threshold of the smoke temperature detector, and a fusion concentration coefficient is calculated by fusing smoke temperature and gas concentration response data. Early warning judgment is achieved by correcting the threshold and fusion judgment value. When an early warning signal is triggered, the temperature rise peak coordinates of the distributed fiber optic cable and the thermal mapping coordinates of the infrared camera are acquired respectively. The final heat source location is determined by the spatial deviation fusion positioning method, and the location is written into the data association index table.

[0051] Example 1, such as Figures 1 to 2As shown, a method for fusing and processing multi-source monitoring data of a wind turbine generator set includes the following steps:

[0052] Step S1: Use infrared camera equipment to monitor the wind turbine generator set, collect the average gray value of the bottom of the nacelle and determine whether the data association mechanism is triggered. In the data association mechanism, retrieve the wireless temperature measurement tag and the fiber optic temperature feature quantity and generate the operating condition tag code.

[0053] Step S2: Access the historical database based on the operating condition tag code to retrieve the target operating condition record, obtain the temperature evolution increment of the target operating condition record and calculate the historical temperature rise characteristics, collect the fiber optic temperature rise value at the current moment, and evaluate the similarity evolution coefficient in combination with the historical temperature rise characteristics.

[0054] Step S3: Obtain the baseline alarm threshold of the smoke temperature detector, correct the baseline alarm threshold using the similarity evolution coefficient, detect the gas response data of the smoke sensing unit and calculate the fusion concentration coefficient, and determine whether to trigger the warning signal based on the corrected baseline alarm threshold.

[0055] Step S4: After triggering the warning signal, collect the temperature rise peak coordinates of the distributed optical fiber and the thermal mapping coordinates of the infrared camera. Combine the temperature rise peak coordinates and thermal mapping coordinates to perform deviation fusion positioning and lock the heat source location. Write the heat source location into the data association index table.

[0056] The specific implementation is as follows:

[0057] In step S1, an infrared camera continuously images and acquires data about the bottom area of ​​the wind turbine nacelle. After grayscale processing of the output image from the infrared camera, the arithmetic mean of the grayscale values ​​of all pixels in the bottom area of ​​the nacelle is calculated to obtain the average grayscale value of the nacelle bottom. The calculation expression is as follows:

[0058] ;

[0059] in, At the time of imaging the bottom of the cabin Average gray value, For the first Each pixel at the imaging time grayscale value, The index value of the pixel. For the imaging time, This represents the total number of pixels in the area at the bottom of the cabin.

[0060] It should be noted that the infrared camera equipment is an image acquisition device deployed inside the nacelle of the wind turbine generator set. It is used to continuously image the bottom area of ​​the nacelle to acquire visual data.

[0061] The average gray value is used to characterize the combined imaging features of light reflection and thermal radiation in the bottom area of ​​the cabin. The larger the value, the stronger the reflection intensity or thermal radiation response in the area, thus reflecting the higher the possibility of changes in the local surface condition.

[0062] The grayscale change is obtained by performing a difference operation on the average grayscale value based on the time series, and its expression is as follows:

[0063] ;

[0064] in, This represents the change in grayscale. At the time of imaging the bottom of the cabin Average gray value, At the time of imaging the bottom of the cabin Average gray value, For adjacent sampling periods.

[0065] The grayscale change represents the intensity of visual disturbance in the bottom area of ​​the cabin per unit time. The larger the value, the more drastic the change in the environmental state at the bottom of the cabin, which may correspond to abnormal factors such as changes in oil stain reflection, local high temperature or smoke obstruction.

[0066] When the amount of grayscale change exceeds the preset trigger threshold, the data association mechanism is triggered.

[0067] Otherwise, the data association mechanism will not be triggered.

[0068] It should be noted that the preset trigger threshold is determined by the statistical quantile value of grayscale change under historical normal operating conditions, and is used to characterize the baseline limit of sensitivity to visual anomalies.

[0069] After the data association mechanism is triggered, the wireless temperature measurement tag and the fiber optic temperature characteristic quantity are retrieved. The wireless temperature measurement tag is the temperature parameter output on the surface of the converter IGBT device, which is used to reflect the local heat accumulation state of the power electronic device. The larger the value, the higher the internal conduction loss or heat dissipation of the converter IGBT device.

[0070] The fiber optic temperature characteristic is the temperature characteristic output by the fiber optic temperature measurement system. It originates from the temperature demodulation results of the corresponding spatial location in the distributed fiber optic temperature measurement loop and is used to reflect the continuous temperature field distribution at the bottom of the cabin. The larger the value, the higher the overall heat accumulation level in that space.

[0071] It should be noted that the IGBT device in the converter is an insulated gate bipolar transistor power module in the power conversion unit of the wind turbine generator set. Wireless temperature measurement tags are deployed on its surface to collect the junction temperature or case temperature parameters of the device in real time. The fiber optic temperature measurement system is a temperature measurement system based on the distributed Raman scattering effect. It utilizes the temperature sensitivity of the backscattered Raman light generated when the incident light pulse propagates in the optical fiber to demodulate and calculate the continuous temperature along the optical fiber path in the internal space of the nacelle and output the spatial temperature characteristic sequence. The distributed fiber optic temperature measurement loop is a continuous fiber optic sensing network structure deployed along the bottom of the nacelle and key equipment. It realizes the acquisition of spatial temperature distribution by the change of scattered signal along the length direction. This loop is used to reflect the propagation path of heat sources inside the nacelle along the spatial structure.

[0072] Furthermore, after normalizing the characteristic quantities of the wireless temperature measurement tag and the fiber optic temperature measurement, a working condition tag code is constructed, which is calculated as follows:

[0073] ;

[0074] in, For operating condition label codes, For wireless temperature measurement tags, This is a characteristic quantity of fiber temperature. and These are preset weighting coefficients used to eliminate scale differences between different temperature measurement mechanisms, enabling wireless temperature tag data and fiber optic temperature measurement data to be linearly mapped within the same numerical space. This indicates a floor operation, used to convert continuous temperature fusion results into discrete integer encoding.

[0075] It should be noted that the preset weighting coefficient is a set of parameters used to unify the scale of wireless temperature measurement tags and fiber optic temperature measurement features during the construction of the working condition tag code. It is used to control the contribution ratio of different sensor data in the fusion calculation. The larger the value, the stronger the influence of the corresponding sensor on the generation of the working condition tag code. It is determined by calibration through the variance stability of historical normal operation data.

[0076] The operating condition tag code is used to uniquely identify the multi-source thermal-visual coupling state of the bottom area of ​​the cabin at the current moment. The larger the value, the higher the overall heat accumulation level of the bottom area of ​​the cabin at that moment and the heavier the heat load of the local power devices. This provides a unified index basis for subsequent historical operating condition matching and similarity retrieval, and realizes the structured association mapping between data from different sensors.

[0077] In step S2, the operating condition tag code is used as the unique index key to perform a structured matching retrieval from the local historical database of the wind turbine generator set. The historical database stores historical operating condition records organized in time series, and each historical operating condition record contains a corresponding historical operating condition tag code.

[0078] It should be noted that the historical database is a structured data storage unit that stores historical operating data of the unit.

[0079] During the historical database retrieval process, the operating condition tag code generated at the current moment is used as the retrieval benchmark. Each historical operating condition tag code in the historical operating condition records is compared and filtered to construct the target operating condition records. Specifically, for any historical operating condition record, the absolute value of the deviation between its historical operating condition tag code and the current operating condition tag code is calculated. When the absolute value of the deviation is less than or equal to a preset tag code matching tolerance threshold, the corresponding historical operating condition record is determined to meet the similarity constraint condition of the current operating condition, and it is included in the target operating condition record set.

[0080] It should be noted that the tag code matching tolerance threshold is used to characterize the allowable deviation range of the working condition code during the historical working condition retrieval process. It is determined by the discrete interval statistical value of the historical working condition distribution. The smaller the value, the stricter the historical working condition matching constraint, thereby improving the retrieval accuracy.

[0081] After obtaining the target operating condition record set, the historical temperature sequence output by the fiber optic temperature measurement system in each target operating condition record is subjected to time difference processing to calculate the temperature evolution increment. The temperature evolution increment is used to characterize the rate of temperature change over time in the historical operating condition. The larger the value, the more intense the heat diffusion or heat accumulation process in the region corresponding to the historical operating condition.

[0082] Furthermore, statistical aggregation of all temperature evolution increments in the target operating condition record set is performed to construct historical temperature rise characteristics, the calculation expression of which is:

[0083] ;

[0084] in, This is a characteristic of historical warming. For the target working condition record set, Record the total number of records for the target operating condition. The index value recorded for the target operating condition. This represents the temperature evolution increment.

[0085] Historical temperature rise characteristics represent the intensity of average temperature evolution under the same or similar operating condition tag code constraints. The larger the value, the more obvious the overall thermal growth trend of the cabin area under this type of operating condition, thus reflecting the poor thermal stability of this type of operating condition.

[0086] At the current moment, differential operation is performed on the real-time temperature sequence output by the fiber optic temperature measurement system to obtain the fiber optic temperature rise value at the current moment. The fiber optic temperature rise value at the current moment represents the instantaneous temperature rise rate of the fiber optic temperature measurement system in the bottom area of ​​the cabin at the current moment. The larger the value, the faster the local heat accumulation rate at the current moment.

[0087] Based on historical temperature rise characteristics and the current fiber temperature rise value, a similarity evolution coefficient is constructed, which is calculated as follows:

[0088] ;

[0089] in, The similarity evolution coefficient, This is a characteristic of historical warming. This represents the current temperature rise of the optical fiber. It is a very small positive number, used to avoid the denominator being zero and to improve numerical stability.

[0090] The similarity evolution coefficient characterizes the degree of consistency between the current temperature rise evolution process and the historical temperature rise evolution process of similar operating conditions. The closer the value is to 1, the more consistent the current temperature rise evolution trend is with the historical operating conditions. A value greater than 1 indicates that the overall temperature rise evolution intensity of the historical operating conditions is higher than that of the current operating conditions. A value less than 1 indicates that the temperature rise evolution of the current operating conditions is more intense and there is a stronger abnormal expansion trend.

[0091] Through the above processing, we can achieve accurate matching of historical operating conditions driven by operating condition tag codes, extraction of temperature evolution increments, and quantifiable evaluation of the degree of similarity evolution, providing continuous input features for subsequent threshold adaptive correction and multi-source fusion judgment.

[0092] In step S3, the baseline alarm threshold is read from the parameter configuration unit of the smoke and temperature detector. The baseline alarm threshold represents the critical level of the smoke and temperature detector to trigger fire risk under standard environmental conditions. The lower the value, the more sensitive it is to smoke and temperature anomalies, and the easier it is to trigger the early warning judgment.

[0093] It should be noted that the smoke and temperature detector is a detection device that integrates temperature and smoke sensing capabilities. The parameter configuration unit is the parameter storage and retrieval module inside the smoke and temperature detector, which is used to store the baseline alarm threshold and control parameters.

[0094] Based on the similarity evolution coefficient, the baseline alarm threshold is corrected to construct the corrected baseline alarm threshold, and its calculation expression is as follows:

[0095] ;

[0096] in, The corrected baseline alarm threshold. This is the original baseline alarm threshold. The similarity evolution coefficient, This is a preset threshold adjustment coefficient used to control the influence of the similarity evolution coefficient on the alarm threshold.

[0097] It should be noted that the preset threshold adjustment coefficient is a control parameter used to correct the baseline alarm threshold. It is used to adjust the influence of the similarity evolution coefficient on the threshold offset. The larger the value, the stronger the influence of historical operating conditions on the current alarm threshold. It is calibrated and determined by the criterion of minimizing historical false alarm rate and missed alarm rate.

[0098] The similarity evolution coefficient reflects the ratio of the current temperature rise rate to the temperature rise rate under similar historical operating conditions. When the similarity evolution coefficient is equal to or greater than 1, it indicates that the thermal evolution intensity of the current operating condition is comparable to or lower than that of similar historical operating conditions. In this case, the corrected baseline alarm threshold is increased or maintained relative to the original value to reduce the sensitivity to false alarms under normal fluctuating operating conditions. When the similarity evolution coefficient is less than 1, it indicates that the current temperature rise rate is significantly higher than the average level of similar historical operating conditions, indicating an increased risk of thermal runaway. In this case, the corrected baseline alarm threshold is decreased relative to the original value to improve the early warning sensitivity to abnormal operating conditions.

[0099] Subsequently, the gas response data output by the smoke sensing unit was collected. The smoke sensing unit includes a carbon monoxide sensor and a volatile organic compound (VOC) sensor. The gas response data output by the carbon monoxide sensor is the carbon monoxide concentration response value, which reflects the characteristics of incomplete combustion or early smoldering. The larger the value, the stronger the potential combustion reaction in the cabin. The gas response data output by the VOC sensor is the VOC concentration response value, which reflects the degree of thermal decomposition of insulating materials or lubricating oil. The larger the value, the more significant the degree of thermal decomposition.

[0100] It should be noted that the smoke sensing unit is a composite detection module composed of multiple gas sensors, which is used to collect information on changes in the concentration of combustion product gases; the carbon monoxide sensor is used to detect changes in the carbon monoxide concentration in the cabin environment, which is used to reflect the intensity of incomplete combustion or smoldering processes; the volatile organic compound sensor is used to detect the concentration of volatile organic compounds in the cabin, which is used to reflect the degree of thermal decomposition of insulating materials, lubricating oil or organic components.

[0101] Normalize each gas response data by dividing it by the maximum observed value within the historical operating cycle to obtain the normalized gas response value. Add the normalized gas response values ​​to obtain the fusion concentration coefficient. The fusion concentration coefficient characterizes the degree of multi-gas synergistic enhancement of the smoke sensing unit. The larger the value, the stronger the synchronicity of the simultaneous increase of carbon monoxide and volatile organic compounds, and the higher the confidence level of the corresponding fire or pyrolysis reaction.

[0102] Furthermore, the current output temperature value of the smoke and temperature detector is collected, and the corrected baseline alarm threshold is jointly determined by the fusion concentration coefficient and the current output temperature value to construct an early warning determination coefficient:

[0103] ;

[0104] in, This is the early warning judgment coefficient. This is the current output temperature value. The corrected baseline alarm threshold. The fusion concentration coefficient, The preset gas fusion enhancement coefficient is used to characterize the compensation weight of gas concentration on the temperature determination result.

[0105] It should be noted that the preset gas fusion enhancement coefficient is a parameter used to enhance and compensate for the influence of gas concentration during the early warning judgment process. It is used to adjust the weight of smoke gas information in the comprehensive risk calculation. The larger the value, the more significant the contribution of gas anomaly to fire risk judgment. It is determined by the optimal point of the ROC curve of historical fire samples.

[0106] The warning judgment coefficient represents the combined risk level of abnormal smoke and temperature and abnormal gas. The higher the value, the higher the risk of fire or thermal runaway in the cabin, and the greater the possibility of triggering a warning response.

[0107] When the warning judgment coefficient is greater than or equal to 0, the warning signal is triggered and the warning judgment flag X=1 is output.

[0108] When the warning determination coefficient is less than 0, it is determined to be a non-warning state, the warning signal is not triggered, and the warning determination flag X=0 is output.

[0109] In step S4, after the warning signal is triggered, the spatial positioning fusion process is started synchronously. The temperature rise peak coordinates are extracted from the fiber optic temperature measurement system. The temperature rise peak coordinates represent the location of the maximum temperature gradient in the fiber optic temperature measurement channel. The value comes from the extreme point location result of the temperature spatial distribution along the distributed fiber optic temperature measurement loop. It is used to reflect the concentrated area of ​​heat propagation along the structural path inside the cabin. The closer the coordinate position is to the core area of ​​the equipment, the higher the risk of heat accumulation.

[0110] Simultaneously, thermal mapping coordinates are extracted from the thermal imaging results of the infrared camera equipment. The thermal mapping coordinates are determined by the maximum response point of gray-scale temperature mapping in the thermal imaging results, which is used to reflect the area with the strongest thermal radiation on the surface of the cabin. The more concentrated the values ​​are, the more obvious the local surface thermal anomaly is.

[0111] The temperature rise peak coordinates and thermal mapping coordinates are transformed to a unified spatial reference, so that they are mapped to the same cabin three-dimensional coordinate system, resulting in standardized temperature rise peak coordinates and standardized thermal mapping coordinates.

[0112] The spatial deviation distance is obtained by calculating the Euclidean distance between the standardized peak temperature rise coordinates and the standardized thermal mapping coordinates in a unified coordinate system. The spatial deviation distance characterizes the consistency between the fiber optic temperature measurement and positioning results and the infrared thermal imaging positioning results. The larger the value, the more significant the difference in the spatial judgment of the heat source between the two sensing mechanisms, thus reflecting a higher degree of non-uniformity in the local thermal field distribution.

[0113] A spatial consistency threshold is set to characterize the allowable multi-source positioning error boundary of the system, which is determined by the quantile statistics of the historical known heat source location error distribution.

[0114] When the spatial deviation distance is less than or equal to the spatial consistency threshold, it indicates that the two types of sensors have consistent spatial positioning results of the heat source. In this case, the standardized temperature rise peak coordinates are taken as the location of the heat source.

[0115] When the spatial deviation distance exceeds the spatial consistency threshold, it indicates a significant deviation between the two types of positioning results. In this case, weighted geometric midpoint fusion is used to perform deviation fusion positioning of the standardized temperature rise peak coordinates and standardized heat mapping coordinates to generate the heat source location. The expression is as follows:

[0116] ;

[0117] in, Location of the heat source To standardize the coordinates of peak temperature rise, These are standardized heat-mapped coordinates.

[0118] The heat source location is used to make a compromise estimate of the spatial location of the heat source when the two types of measurement results are inconsistent, thereby reducing the impact of a single sensor error on the positioning result.

[0119] Finally, the locked heat source location, along with the current operating condition label code, similarity evolution coefficient, fusion concentration coefficient, and early warning judgment flag, is encapsulated into a structured association record and written to the data association index table. The data association index table is a structured index database used to store the fusion results of multi-source monitoring data. Its fields include the operating condition label code, heat source spatial coordinates, similarity evolution coefficient, fusion concentration coefficient, early warning timestamp, and early warning judgment flag, used to establish a mapping relationship between the operating condition label code, heat source spatial location, and multi-source risk characteristics.

[0120] In subsequent monitoring cycles, when the newly generated operating condition tag code matches an existing record in the index table, the system will prioritize calling the heat source location and risk characteristics in that record as a historical reference baseline to assist in the consistency verification between the current early warning judgment and spatial positioning, thereby realizing the closed-loop feedback utilization of historical early warning events and rapid backtracking across sensor operating conditions.

[0121] Finally, it should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0122] Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0123] In this document, the singular forms “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that terms such as “comprising / including” or “having” specify the presence of the stated features, integrals, steps, operations, components, parts, or combinations thereof, but do not preclude the possibility of the presence or addition of one or more other features, integrals, steps, operations, components, parts, or combinations thereof. Meanwhile, the term “and / or” as used in this specification includes any and all combinations of the associated listed items.

[0124] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0125] The above description of the disclosed embodiments will enable those skilled in the art to make or use various modifications to these embodiments. It will be readily apparent to those skilled in the art that the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for fusing and processing multi-source monitoring data from wind turbine generator sets, characterized in that: Includes the following steps: Step S1: Use infrared camera equipment to monitor the wind turbine generator set, collect the average gray value of the bottom of the nacelle and determine whether the data association mechanism is triggered. In the data association mechanism, retrieve the wireless temperature measurement tag and the fiber optic temperature characteristic quantity and generate the operating condition tag code. Step S2: Access the historical database based on the operating condition tag code to retrieve the target operating condition record, obtain the temperature evolution increment of the target operating condition record and calculate the historical temperature rise characteristics, collect the fiber optic temperature rise value at the current moment, and evaluate the similarity evolution coefficient in combination with the historical temperature rise characteristics. Step S3: Obtain the baseline alarm threshold of the smoke temperature detector, correct the baseline alarm threshold using the similarity evolution coefficient, detect the gas response data of the smoke sensing unit and calculate the fusion concentration coefficient, and determine whether to trigger the warning signal based on the corrected baseline alarm threshold. Step S4: After triggering the warning signal, collect the temperature rise peak coordinates of the distributed optical fiber and the thermal mapping coordinates of the infrared camera. Combine the temperature rise peak coordinates and thermal mapping coordinates to perform deviation fusion positioning and lock the heat source location. Write the heat source location into the data association index table.

2. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 1, characterized in that: In step S1, an infrared camera continuously captures images of the bottom area of ​​the wind turbine nacelle. After the output image of the infrared camera is processed into grayscale, the arithmetic mean of the grayscale values ​​of all pixels in the bottom area of ​​the computer cabin is used to obtain the average grayscale value of the bottom of the cabin. The grayscale change is obtained by performing a difference operation on the average grayscale value based on the time series. When the amount of grayscale change exceeds the preset trigger threshold, the data association mechanism is triggered. Otherwise, the data association mechanism will not be triggered.

3. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 1, characterized in that: In step S1, after the data association mechanism is triggered, the wireless temperature measurement tag and the fiber optic temperature characteristic quantity are retrieved, wherein the wireless temperature measurement tag is the temperature parameter output on the surface of the converter IGBT device. The fiber optic temperature characteristic is the temperature characteristic output by the fiber optic temperature measurement system. After normalizing the characteristic quantities of wireless temperature measurement tags and fiber optic temperature measurement, a working condition tag code is constructed.

4. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 3, characterized in that: In step S2, a structured matching retrieval is performed from the local historical database of the wind turbine generator set, using the operating condition tag code as the unique index key. The historical database stores historical operating condition records organized in time series, with each historical operating condition record containing a corresponding historical operating condition tag code; During the historical database retrieval process, the working condition tag code generated at the current moment is used as the retrieval benchmark. The corresponding historical working condition tag codes in the historical working condition records are compared and filtered one by one to construct the target working condition record. The historical temperature sequence output by the fiber optic temperature measurement system in each target operating condition record is processed by time difference to calculate the temperature evolution increment.

5. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 4, characterized in that: In step S2, all temperature evolution increments in the target operating condition record set are statistically aggregated to construct historical temperature rise characteristics; At the current moment, the real-time temperature sequence output by the fiber optic temperature measurement system is differentially calculated to obtain the current fiber optic temperature rise value. A similarity evolution coefficient is constructed based on historical temperature rise characteristics and the current fiber temperature rise value.

6. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 5, characterized in that: In step S3, a baseline alarm threshold is read from the parameter configuration unit of the smoke and temperature detector. The baseline alarm threshold represents the critical level at which the smoke and temperature detector triggers fire risk under standard environmental conditions. Based on the similarity evolution coefficient, the baseline alarm threshold is corrected to construct the corrected baseline alarm threshold. The gas response data output by the smoke sensing unit is collected. The smoke sensing unit includes a carbon monoxide sensor and a volatile organic compound sensor. The carbon monoxide sensor outputs gas response data as the carbon monoxide concentration response value; The gas response data output by the volatile organic compound sensor is the volatile organic compound concentration response value.

7. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 6, characterized in that: In step S3, each gas response data is normalized by dividing it by the maximum observed value within the historical operating cycle to obtain the normalized gas response value. The normalized gas response values ​​are summed to obtain the fusion concentration coefficient; The current output temperature value of the smoke temperature detector is collected, and the corrected baseline alarm threshold is jointly determined with the fusion concentration coefficient and the current output temperature value to construct the early warning determination coefficient. When the warning judgment coefficient is greater than or equal to 0, the warning signal is triggered and the warning judgment flag X=1 is output. When the warning determination coefficient is less than 0, it is determined to be a non-warning state, the warning signal is not triggered, and the warning determination flag X=0 is output.

8. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 1, characterized in that: In step S4, the temperature rise peak coordinates are extracted from the fiber optic temperature measurement system. The temperature rise peak coordinates represent the location of the maximum temperature gradient in the fiber optic temperature measurement channel. The thermal mapping coordinates are extracted from the thermal imaging results of the infrared camera. The thermal mapping coordinates are the maximum response points of the gray-scale temperature mapping in the thermal imaging results. A unified spatial reference transformation is performed on the temperature rise peak coordinates and heat mapping coordinates to obtain standardized temperature rise peak coordinates and standardized heat mapping coordinates. The spatial deviation distance is obtained by calculating the Euclidean distance between the standardized peak temperature rise coordinates and the standardized thermal mapping coordinates.

9. The method for fusion processing of multi-source monitoring data of wind turbine generator sets according to claim 8, characterized in that: In step S4, a spatial consistency threshold is set. When the spatial deviation distance is less than or equal to the spatial consistency threshold, the standardized temperature rise peak coordinates are taken as the location of the heat source. When the spatial deviation distance is greater than the spatial consistency threshold, weighted geometric midpoint fusion is used to perform deviation fusion positioning of the standardized temperature rise peak coordinates and standardized heat mapping coordinates and generate the heat source location. The locked heat source location, current operating condition tag code, similarity evolution coefficient, fusion concentration coefficient, and early warning judgment flag are encapsulated into a structured association record and written into the data association index table; The data association index table is a structured index database used to store the fusion results of multi-source monitoring data.

Images