Variable pitch system state monitoring method and related device

Abstract

The invention belongs to the technical field of state monitoring, and discloses a variable pitch system state monitoring method and a related device. The variable-pitch system state monitoring method comprises the steps of collecting sensor data of a variable-pitch system in real time, cleaning the sensor data, collecting historical sensor data of the variable-pitch system, constructing a feature matrix according to the cleaned sensor data and the historical sensor data, performing feature extraction on the feature matrix by adopting a deep learning model, and obtaining the state of the variable-pitch system. The extracted features are fused, a fused feature vector is generated, based on the fused feature vector, an integrated learning algorithm is adopted to construct and train a variable pitch system state monitoring model, and the state of the variable pitch system is monitored in real time; according to the invention, abnormal conditions of the system can be found efficiently, accurately and timely, measures are taken in advance for maintenance and repair, further expansion of faults is avoided, the risk of equipment damage is reduced, and the reliability and safety of the variable pitch system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of condition monitoring, and in particular to a variable pitch system condition monitoring method and related devices. Background Art

[0002] Pitch system condition monitoring plays a crucial role in the wind power industry and is a key technology for ensuring the efficient and stable operation of wind turbines. Wind turbines operate in complex and ever-changing natural environments. Their pitch systems precisely adjust blade angles to maximize wind energy capture. During this process, the system's performance directly impacts power generation efficiency and equipment safety. Due to the long-term effects of wind loads, gravity, and inertia, pitch systems are prone to wear, fatigue, and even failure, which can impact the operation of the entire wind turbine.

[0003] However, the variable pitch system is exposed to harsh natural environments for a long time, such as strong winds, dust, rain and snow, and needs to withstand complex and changeable loads, including wind loads, gravity, inertia, etc. These factors greatly increase the risk of failure; once the variable pitch system fails, it will not only affect the power generation efficiency of the wind turbine, but may also pose a serious threat to equipment safety.

[0004] Therefore, there is an urgent need for a highly accurate and efficient condition monitoring method to perform real-time and comprehensive monitoring of the pitch system. Summary of the Invention

[0005] The purpose of the present invention is to provide a variable pitch system status monitoring method and related devices to overcome the problems existing in the prior art. The present invention can efficiently, accurately and timely detect abnormal conditions of the system, take measures in advance for maintenance and repair, avoid further expansion of faults, reduce the risk of equipment damage, and improve the reliability and safety of the variable pitch system.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a method for monitoring the state of a pitch system, comprising the following steps: Collect sensor data of the pitch system in real time and clean the sensor data; Collect historical sensor data of the pitch system and construct a feature matrix based on the cleaned sensor data and historical sensor data; A deep learning model is used to extract features from the feature matrix, and the extracted features are fused to generate a fused feature vector; Based on the fusion feature vector, an integrated learning algorithm is used to build and train a variable pitch system status monitoring model to monitor the status of the variable pitch system in real time. Furthermore, the sensor data specifically includes: blade angle sensor data, servo motor current data, bearing temperature data and vibration sensor data; Furthermore, the cleaning of the sensor data specifically includes: The sensor data is standardized and missing value processed in sequence to obtain normal data and abnormal data. A support vector single classification data cleaning model is constructed based on the normal data. The support vector single classification data cleaning model is trained based on the normal data and abnormal data. The trained support vector single classification data cleaning model is used to clean the sensor data. Furthermore, constructing a feature matrix based on the preprocessed sensor data and historical sensor data specifically includes: Arrange the preprocessed sensor data and historical sensor data in chronological order so that different sensor data at the same time point correspond to each other, forming an ordered data set. Then, filter the data set to obtain the screening features. The screening features are used as the columns of the matrix, and each time point is used as the row of the matrix. The preprocessed sensor data and historical sensor data are sequentially filled into the matrix to obtain the feature matrix. Furthermore, the deep learning model is used to extract features from the feature matrix, and the extracted features are fused to generate a fused feature vector, which specifically includes: Select a suitable deep learning model based on the feature matrix, input the feature matrix into the deep learning model, and the deep learning model processes and transforms the input feature matrix through neurons. Then, several feature representations are extracted from the transformed feature matrix, and the several feature representations are fused. The fused features are normalized to obtain a fused feature vector. Furthermore, the fusion of several feature representations specifically includes: Concatenate several features, or assign weights to several features and then sum them up; Furthermore, the method of constructing and training a pitch system state monitoring model based on the fused feature vector and adopting an integrated learning algorithm to monitor the state of the pitch system in real time specifically includes: The fused feature vector data is divided into a training set and a test set, and several base learners are selected. Each base learner is trained using the training set to obtain prediction results, which include normal or abnormal results. The prediction results are then evaluated using the test set. The variable pitch system state monitoring model is optimized based on the evaluation results to obtain a trained variable pitch system state monitoring model. The fused feature vector is input into the trained variable pitch system state monitoring model to obtain real-time state monitoring results of the variable pitch system, which include normal or abnormal results.

[0007] In a second aspect, the present invention provides a pitch system status monitoring system, comprising: Sensor data acquisition and cleaning module, used to collect sensor data of the pitch system in real time and clean the sensor data; The historical sensor data acquisition and feature matrix construction module is used to collect the historical sensor data of the pitch system and construct the feature matrix based on the cleaned sensor data and historical sensor data; A fusion feature vector generation module is used to extract features from the feature matrix using a deep learning model, fuse the extracted features, and generate a fusion feature vector; The variable pitch system status monitoring module is used to build and train the variable pitch system status monitoring model based on the fusion feature vector and the integrated learning algorithm to monitor the status of the variable pitch system in real time.

[0008] In a third aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0009] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, which implements the steps of the above method when executed by a processor.

[0010] The above technical solution has the following advantages or beneficial effects: In the first aspect, the present invention provides a method for monitoring the state of a variable pitch system. By collecting sensor data in real time and cleaning it, it can effectively remove noise and outliers, ensure that the data input into the subsequent analysis link is accurate and reliable, lay a solid foundation for accurate state monitoring, help avoid misjudgment due to data quality problems, and improve the accuracy of monitoring results; by combining the real-time cleaned sensor data and historical sensor data to construct a feature matrix, it can comprehensively mine the operating characteristics of the variable pitch system in different time dimensions; using a deep learning model to extract features from the feature matrix, it can automatically learn the complex and high-level features in the data, and compared with traditional methods, it can more deeply mine the information behind the data; based on the fused feature vector, an integrated learning algorithm is used to construct and train a variable pitch system state monitoring model. The integrated learning algorithm can give full play to the advantages of different algorithms by combining multiple base learners, and improve the generalization ability and robustness of the model; real-time monitoring of the state of the variable pitch system can efficiently, accurately and promptly detect system abnormalities, take measures in advance for maintenance and repair, avoid further expansion of faults, reduce the risk of equipment damage, and improve the reliability and safety of the variable pitch system.

[0011] Furthermore, the blade angle is one of the key parameters of the variable pitch system, which directly reflects the attitude adjustment of the blade under different working conditions; the servo motor is the power source for driving the blade to rotate, and its current is closely related to the load conditions; the bearing is a key component in the variable pitch system that bears load and rotational motion, and its operating temperature is an important indicator reflecting the operating status of the bearing; vibration is a comprehensive reflection of the operating status of mechanical equipment, and each component in the variable pitch system will produce varying degrees of vibration during operation.

[0012] Furthermore, by standardizing and processing missing values ​​of sensor data, and constructing and training a support vector single classification data cleaning model for cleaning, the data quality can be significantly improved, abnormal data can be accurately identified, and the data cleaning effect can be enhanced, thereby improving the performance of the variable pitch system status monitoring model and providing strong support for the reliable operation of the variable pitch system.

[0013] Furthermore, by arranging data in chronological order, filtering features and forming a feature matrix, the temporal correlation of the data can be retained, the effectiveness of the feature matrix can be improved, the generalization ability of the model can be enhanced, and data management and analysis can be facilitated, providing more powerful support for the condition monitoring of the variable pitch system.

[0014] Furthermore, the deep learning model has powerful nonlinear mapping capabilities and can perform complex processing and transformation on the input feature matrix through multiple layers of neurons; fusing several feature representations can integrate the complementary information between different features to form a more comprehensive and representative feature vector; feature fusion avoids the one-sidedness and limitations that may exist in a single feature, so that the fused feature vector can make more full use of the information in the feature matrix; normalizing the fused features and unifying the value ranges of different features into a specific range helps to eliminate the impact of dimensional and scale differences between different features, so that each feature has equal importance in the model training process; monitoring the status of the variable pitch system based on the fused feature vector can more accurately evaluate the operating status of the system.

[0015] Furthermore, feature splicing can fully retain the information of each feature, directly connect different feature vectors to form a longer feature vector without losing any feature details, so that the fused feature vector contains richer original information, providing a comprehensive data basis for subsequent status monitoring models, and helping the model capture more potential system status patterns; the summation after assigning weights takes into account the importance differences of different features; by reasonably allocating weights, the impact of key features on the system status is highlighted, and the role of secondary features is weakened.

[0016] Furthermore, by dividing the data into training and test sets, we can objectively evaluate the model performance. By reasonably dividing the fusion feature vector data, using the training set to train the base learner, and then evaluating it with the test set, we can accurately understand the performance of the model on unseen data, avoid overfitting, and ensure the generalization ability of the model.

[0017] In the second aspect, the present invention provides a variable pitch system status monitoring system, in which the sensor data acquisition and cleaning module can obtain the sensor data of the variable pitch system in real time, and remove noise and outliers through cleaning to ensure data quality; the fusion feature vector generation module uses a deep learning model to extract features from the feature matrix, and automatically learns the complex patterns and deep-level features in the data; the variable pitch system status monitoring module is based on the fusion feature vector and adopts an integrated learning algorithm to construct and train the status monitoring model; the system adopts a modular design, and each module has clear functions and is relatively independent, which is convenient for separate maintenance, debugging and upgrading.

[0018] In a third aspect, the present invention provides a computer device that can efficiently implement the steps of the method of the present invention by executing a specific computer program through a processor. When performing data processing tasks, the computer device can accurately perform numerical calculations and logical judgments, avoiding errors caused by human factors; at the same time, since the computer program has a high degree of stability and reliability, the accuracy and consistency of the data processing results can be ensured.

[0019] In a fourth aspect, the present invention provides a computer-readable storage medium. By programming the steps of the method of the present invention into a computer program and storing it on a computer-readable storage medium, users can easily load these programs onto any compatible computer device and execute them without rewriting or converting the code, thereby greatly improving the convenience and flexibility of program execution. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic flow chart of a method for monitoring the state of a pitch system according to the present invention; Figure 2 Schematic diagram of the structure of the computer device of the present invention. DETAILED DESCRIPTION

[0021] The present invention will be further described in detail below with reference to specific embodiments, which are intended to explain the present invention rather than to limit it. In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention. It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0022] Example: See also Figure 1 The present invention provides a method for monitoring the state of a pitch system, comprising the following steps: Step 1: collect sensor data of the pitch system in real time and clean the sensor data; Specifically, the sensor data includes: blade angle sensor data, servo motor current data, bearing temperature data, and vibration sensor data; Specifically, cleaning the sensor data includes: performing standardization and missing value processing on the sensor data in sequence to obtain normal data and abnormal data, constructing a support vector single classification data cleaning model based on the normal data, taking the normal data and abnormal data as prepared data, dividing the prepared data into a training set and a test set, setting various parameters of the support vector single classification data cleaning model, using the training set to train the support vector single classification data cleaning model, and then using the test set to calculate various indicators of the support vector single classification data cleaning model, evaluating the trained support vector single classification data cleaning model according to the various indicators, and then tuning the parameters and data volume of the support vector single classification data cleaning model according to the evaluation results to obtain a trained support vector single classification data cleaning model, inputting the sensor data into the trained support vector single classification data cleaning model, and identifying multi-source heterogeneous data through the support vector single classification data cleaning model. If it is identified as normal data, the normal data is directly output; if it is identified as abnormal data, the abnormal data is cleaned; Specifically, standardization includes scaling sensor data to a uniform scale. Missing value processing includes filling missing values ​​in the normalized sensor data using a time series-based interpolation algorithm, marking sensor data with missing values ​​that do not exceed a preset threshold as normal data, and marking sensor data with missing values ​​that exceed a preset threshold as abnormal data. Preferably, cleaning the abnormal data includes: using a smoothing correction algorithm based on adjacent data to clean the data, or deleting the abnormal data; Step 2: Collect historical sensor data of the pitch system and construct a feature matrix based on the cleaned sensor data and historical sensor data; Specifically, the preprocessed sensor data and historical sensor data are arranged in chronological order so that different sensor data at the same time point correspond to each other, forming an ordered data set. The data set is then filtered to obtain filtering features. The filtering features are used as the columns of the matrix, and each time point is used as the row of the matrix. The preprocessed sensor data and historical sensor data are sequentially filled into the matrix to obtain a feature matrix. Step 3: Use a deep learning model to extract features from the feature matrix, fuse the extracted features, and generate a fused feature vector; Specifically, a suitable deep learning model is selected according to the feature matrix, and the feature matrix is ​​input into the deep learning model. The deep learning model processes and transforms the input feature matrix through neurons, and then extracts several feature representations from the transformed feature matrix. The several feature representations are fused, and the fused features are normalized to obtain a fused feature vector. Preferably, selecting a suitable deep learning model according to the feature matrix includes: if the feature matrix has a spatial structure, selecting a convolutional neural network (CNN); if the feature matrix has a time series relationship, selecting a recurrent neural network (RNN); Preferably, fusing several feature representations includes: concatenating several features, or summing up several features after assigning weights to them; Step 4: Based on the fused feature vector, an integrated learning algorithm is used to build and train a variable pitch system status monitoring model to monitor the status of the variable pitch system in real time; Specifically, the fused feature vector data is divided into a training set and a test set, several base learners are selected, and each base learner is trained using the training set to obtain a prediction result, which includes normal or abnormal results. The prediction result is then evaluated using the test set, and the variable pitch system state monitoring model is optimized according to the evaluation result to obtain a trained variable pitch system state monitoring model. The fused feature vector is input into the trained variable pitch system state monitoring model to obtain a real-time state monitoring result of the variable pitch system, which includes normal or abnormal results.

[0023] In one embodiment of the present invention, a pitch system condition monitoring system is provided, comprising: Sensor data acquisition and cleaning module, used to collect sensor data of the pitch system in real time and clean the sensor data; The historical sensor data acquisition and feature matrix construction module is used to collect the historical sensor data of the pitch system and construct the feature matrix based on the cleaned sensor data and historical sensor data; A fusion feature vector generation module is used to extract features from the feature matrix using a deep learning model, fuse the extracted features, and generate a fusion feature vector; The variable pitch system status monitoring module is used to build and train the variable pitch system status monitoring model based on the fusion feature vector and the integrated learning algorithm to monitor the status of the variable pitch system in real time.

[0024] See also Figure 2 In one embodiment of the present invention, a computer device is provided, comprising a processor and a memory, wherein the memory is used to store a computer program, the computer program including program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., and is the computing core and control core of the terminal. It is suitable for implementing one or more instructions, specifically for loading and executing one or more instructions in the computer storage medium to implement a corresponding method process or corresponding function. The processor described in this embodiment of the present invention can be used to operate a variable pitch system condition monitoring method.

[0025] In one embodiment of the present invention, a computer-readable storage medium is provided, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space that stores the operating system of the terminal. In addition, the storage space also stores one or more instructions suitable for being loaded and executed by the processor. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of a variable pitch system status monitoring method in the embodiment.

[0026] Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0027] The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as the combination of processes and / or blocks in the flowcharts and / or block diagrams, 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 the instructions for implementing the processes in the flowcharts and / or block diagrams. 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.

[0028] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work 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 The function specified in one or more boxes.

[0029] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device 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.

[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring the state of a pitch system, characterized in that: The following steps are involved: Collect sensor data of the pitch system in real time and clean the sensor data; Collect historical sensor data of the pitch system and construct a feature matrix based on the cleaned sensor data and historical sensor data; A deep learning model is used to extract features from the feature matrix, and the extracted features are fused to generate a fused feature vector; Based on the fusion feature vector, an integrated learning algorithm is used to build and train a variable pitch system status monitoring model to monitor the status of the variable pitch system in real time.

2. A pitch system status monitoring method according to claim 1, characterized in that: The sensor data specifically includes: blade angle sensor data, servo motor current data, bearing temperature data and vibration sensor data.

3. A pitch system status monitoring method according to claim 1, characterized in that: The cleaning of the sensor data specifically includes: The sensor data is standardized and missing value processed in turn to obtain normal data and abnormal data. A support vector single classification data cleaning model is constructed based on the normal data. The support vector single classification data cleaning model is trained based on the normal data and abnormal data. The trained support vector single classification data cleaning model is used to clean the sensor data.

4. A method for monitoring the state of a pitch system according to claim 1, characterized in that: The constructing of a feature matrix based on the preprocessed sensor data and the historical sensor data specifically includes: Arrange the preprocessed sensor data and historical sensor data in chronological order so that different sensor data at the same time point correspond to each other, forming an ordered data set. Then, filter the data set to obtain the filtering features. The filtering features are used as the columns of the matrix, and each time point is used as the row of the matrix. The preprocessed sensor data and historical sensor data are filled into the matrix in turn to obtain the feature matrix.

5. A method for monitoring the state of a pitch system according to claim 1, characterized in that: The method uses a deep learning model to extract features from the feature matrix, fuses the extracted features, and generates a fused feature vector, specifically including: A suitable deep learning model is selected according to the feature matrix, and the feature matrix is ​​input into the deep learning model. The deep learning model processes and transforms the input feature matrix through neurons, and then extracts several feature representations from the transformed feature matrix, fuses the several feature representations, and normalizes the fused features to obtain a fused feature vector.

6. A method for monitoring the state of a pitch system according to claim 5, characterized in that: The fusion of several feature representations specifically includes: Concatenate several features, or assign weights to several features and then sum them up.

7. A method for monitoring the state of a pitch system according to claim 1, characterized in that: The method uses an integrated learning algorithm based on the fusion feature vector to build and train a variable pitch system state monitoring model to monitor the state of the variable pitch system in real time, specifically including: The fused feature vector data is divided into a training set and a test set, and several base learners are selected. Each base learner is trained using the training set to obtain prediction results, which include normal or abnormal results. The prediction results are then evaluated using the test set. The variable pitch system state monitoring model is optimized based on the evaluation results to obtain a trained variable pitch system state monitoring model. The fused feature vector is input into the trained variable pitch system state monitoring model to obtain real-time state monitoring results of the variable pitch system, which include normal or abnormal results.

8. A variable pitch system status monitoring system, characterized in that: include: Sensor data acquisition and cleaning module, used to collect sensor data of the pitch system in real time and clean the sensor data; The historical sensor data acquisition and feature matrix construction module is used to collect the historical sensor data of the pitch system and construct the feature matrix based on the cleaned sensor data and historical sensor data; A fusion feature vector generation module is used to extract features from the feature matrix using a deep learning model, fuse the extracted features, and generate a fusion feature vector; The variable pitch system status monitoring module is used to build and train the variable pitch system status monitoring model based on the fusion feature vector and the integrated learning algorithm to monitor the status of the variable pitch system in real time.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images