Method for analyzing cabin vibration monitoring data based on SCADA (Supervisory Control And Data Acquisition) system

Through the multi-point vibration data processing and neural network analysis of the SCADA system, the complexity of vibration monitoring of offshore wind turbine cabins is solved, and the direct analysis of vibration modes in deep sea environments is realized.

CN118940054APending Publication Date: 2024-11-12CHINA POWER CONSRTUCTION GRP GUIYANG SURVEY & DESIGN INST CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411129110.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-16
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In offshore wind power projects, the cabin structure is complex, the causes of vibration are complex, and a single vibration data source is difficult to analyze, and the existing technology is difficult to realize direct processing of online data, resulting in low vibration monitoring efficiency, especially in deep sea environments.

Method used

The SCADA system obtains cabin vibration monitoring data at multiple points, performs standardized processing and FFT analysis, uses 2 to 5 layers of fully connected feedforward neural networks to combine calculation networks, combines similarity and fuzzy algorithm to determine vibration mode and fault type, and The vibration frequency spectrum is displayed visually.

Benefits of technology

It realizes direct online data processing, simplifies multi-point vibration monitoring data analysis, improves the accuracy and efficiency of vibration mode judgment, is suitable for deep sea environments, and can directly return early warning and alarm information.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940054A_ABST
    Figure CN118940054A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of control or regulation systems, and provides a cabin vibration monitoring data analysis method based on an SCADA system, and the method comprises the following steps: S1, obtaining cabin vibration monitoring data of a plurality of point locations from the SCADA system; s2, carrying out standardization processing on the collected data; s3, independently carrying out FFT (Fast Fourier Transform) analysis on the cabin vibration monitoring data of each point location, and extracting a vibration frequency spectrum of each point location; and S4, inputting the vibration frequency spectrums of the plurality of point locations into a merging calculation network, and calculating to obtain an overall vibration frequency spectrum and a vibration evaluation value. According to the invention, online data direct processing can be realized in a relatively simple and direct manner, cabin vibration monitoring data of multiple point locations are directly analyzed to obtain a vibration mode, so that early warning and alarm information can be directly returned to an SCADA system according to the vibration mode, and the method is especially suitable for cabin vibration monitoring deployed in deep and far sea.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to an analysis method based on SCADA system cabin vibration monitoring data, belonging to the technical field of control or regulation systems. Background Art

[0002] The SCADA system is a computer-based DCS and power automation monitoring system with a wide range of applications. It can be used in data collection and monitoring control as well as process control in the fields of electricity, metallurgy, petroleum, chemical industry, gas, railways, etc.

[0003] At present, the SCADA system used in offshore wind power projects can be programmed with SCADA, so it is mainly used for real-time analysis of monitoring data, and related technologies are also widely used in other industries. At present, there are technical solutions for vibration analysis using SCADA systems, such as a method for time synchronization based on elevator car vibration signals and SCADA data disclosed in Chinese patent application number CN202410494578, and a generator fault identification method based on fusion analysis of Scada system data and vibration data disclosed in Chinese patent application number CN202311506757.

[0004] In offshore wind power projects, the structure of the cabin is relatively complex and presents an integrated system. At the same time, the causes of vibration are also complex. The vibration sources include wind, waves and marine life. It is difficult to analyze the real cause of vibration from a single vibration data source. Therefore, the most common idea is to use multiple vibration data sources, which are installed at multiple spatial points. From the hardware point of view, it is possible to analyze the real cause of vibration through vibration data. Since multiple vibration data sources are used to obtain vibration monitoring data at multiple points, the difficulty of analysis increases sharply, which makes the vibration method of a single data source in the past unable to be used directly. At present, the main method is to process online data offline. It is difficult to draw conclusions based on the direct processing of online data, and the effectiveness is poor. In particular, it is basically impossible to achieve offline processing for cabins deployed in deep seas, which makes the vibration monitoring method of existing technologies basically unusable. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an analysis method based on the cabin vibration monitoring data of the SCADA system. The analysis method based on the cabin vibration monitoring data of the SCADA system can realize direct online data processing in a relatively simple and direct manner, and directly analyze the cabin vibration monitoring data of multiple points to obtain the vibration mode, thereby facilitating the direct return of early warning and alarm information to the SCADA system according to the vibration mode, and is particularly suitable for cabin vibration monitoring deployed in deep sea.

[0006] The present invention is achieved through the following technical solutions.

[0007] The present invention provides an analysis method based on SCADA system cabin vibration monitoring data, comprising the following steps: S1: Obtain cabin vibration monitoring data at multiple points from the SCADA system; S2: Standardize the collected data; S3: Perform FFT analysis on the cabin vibration monitoring data at each point separately to extract the vibration frequency spectrum of each point; the combined calculation network is a 2-5 layer fully-connected feedforward neural network; S4: Input the vibration frequency spectra of multiple points into the combined calculation network to calculate the overall vibration frequency spectrum and vibration evaluation value; S5: A similarity algorithm is used to determine the vibration mode according to the overall vibration frequency spectrum in the historical data and the overall vibration frequency spectrum calculated in the previous step, and a fuzzy algorithm is used to calculate the vibration fault type according to the vibration evaluation value.

[0008] The following steps are also included: S6: The vibration frequency spectrum is displayed in the time domain as a single multi-curve graph using a visualization method.

[0009] The similarity algorithm in step S5 is the Pearson correlation coefficient algorithm, which uses ergodic calculation to select the vibration mode corresponding to the overall vibration frequency spectrum in the historical data with the highest similarity calculation result as the judgment result.

[0010] The combined computing network is a three-layer fully-connected feedforward neural network.

[0011] In the step S1, the cabin vibration monitoring data at multiple points are aligned on the time axis in the time domain.

[0012] In step S3, when performing FFT analysis, the frequencies whose maximum amplitude is less than a preset threshold are filtered out; the preset threshold is set according to the cabin structure.

[0013] The preset threshold is 1 / 20 of the fault threshold, and the fault threshold is the minimum vibration amplitude that causes a fault and is set according to the cabin structure.

[0014] In the historical data, the overall vibration frequency spectrum and the vibration mode are stored in pairs.

[0015] In step S5, the membership function of the fuzzy algorithm is a single variable piecewise function.

[0016] The beneficial effects of the present invention are that it can realize direct online data processing in a relatively simple and direct manner, and directly analyze the cabin vibration monitoring data at multiple points to obtain the vibration mode, thereby facilitating the direct return of early warning and alarm information to the SCADA system according to the vibration mode, and is particularly suitable for cabin vibration monitoring deployed in deep sea. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of a process of at least one embodiment of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is further described below, but the scope of protection claimed is not limited to the description.

[0019] A first embodiment of the present invention relates to Figure 1 The analysis method of cabin vibration monitoring data based on the SCADA system shown includes the following steps: S1: Obtain cabin vibration monitoring data at multiple points from the SCADA system; S2: Standardize the collected data; S3: Perform FFT analysis on the cabin vibration monitoring data at each point separately to extract the vibration frequency spectrum of each point; S4: Input the vibration frequency spectra of multiple points into the combined calculation network to calculate the overall vibration frequency spectrum and vibration evaluation value; the combined calculation network is a 2-5 layer fully-connected feedforward neural network; S5: A similarity algorithm is used to determine the vibration mode according to the overall vibration frequency spectrum in the historical data and the overall vibration frequency spectrum calculated in the previous step, and a fuzzy algorithm is used to calculate the vibration fault type according to the vibration evaluation value.

[0020] In the cabin vibration monitoring deployed in the deep sea, it is necessary to avoid the ambiguity of calculation and ensure the accuracy of calculation, and to avoid large data transmission and reception, and try to localize the calculation process. Therefore, medium and large neural network models are basically unavailable, and remote interactive calculation based on cloud platform is also unavailable. This implementation method is based on FFT analysis that is convenient for local calculation. Compared with the combined calculation network with simpler and less computational complexity, it can effectively localize the calculation process, with small overall calculation complexity, small data transmission and reception (mainly only need to send vibration fault type and receive control instructions), and high accuracy of calculation results (the accuracy of fuzzy calculation is higher than that of complex neural network).

[0021] The second embodiment of the present invention is substantially the same as the first embodiment, mainly in that the similarity algorithm in step S5 is a Pearson correlation coefficient algorithm, which uses a traversal calculation to select the vibration mode corresponding to the overall vibration frequency spectrum in the historical data with the highest similarity calculation result as the judgment result.

[0022] Furthermore, in step S5, the membership function of the fuzzy algorithm is a single variable piecewise function.

[0023] Furthermore, in step S1, the cabin vibration monitoring data at multiple points are aligned on the time axis of the time domain.

[0024] Preferably, the combined computing network is a three-layer fully-connected feedforward neural network. The third embodiment of the present invention is substantially the same as the first embodiment, mainly in that it further includes the following steps: S6: The vibration frequency spectrum is displayed in the time domain as a single multi-curve graph using a visualization method.

[0025] Furthermore, in the historical data, the overall vibration frequency spectrum and the vibration mode are stored in pairs.

[0026] Furthermore, in step S3, when performing FFT analysis, the frequencies whose maximum amplitude is less than a preset threshold are filtered out; the preset threshold is set according to the cabin structure.

[0027] Preferably, the preset threshold is 1 / 20 of the fault threshold, and the fault threshold is the minimum vibration amplitude that causes a fault and is set according to the cabin structure.

Claims

1. An analysis method based on cabin vibration monitoring data of a SCADA system, characterized in that: The following steps are involved: S1: Obtain cabin vibration monitoring data at multiple points from the SCADA system; S2: Standardize the collected data; S3: Perform FFT analysis on the cabin vibration monitoring data at each point separately to extract the vibration frequency spectrum of each point; S4: Input the vibration frequency spectra of multiple points into the combined calculation network to calculate the overall vibration frequency spectrum and vibration evaluation value; the combined calculation network is a 2-5 layer fully-connected feedforward neural network; S5: A similarity algorithm is used to determine the vibration mode according to the overall vibration frequency spectrum in the historical data and the overall vibration frequency spectrum calculated in the previous step, and a fuzzy algorithm is used to calculate the vibration fault type according to the vibration evaluation value.

2. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 1, characterized in that, The following steps are also included: S6: The vibration frequency spectrum is displayed in the time domain as a single multi-curve graph using a visualization method.

3. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 1, characterized in that, The similarity algorithm in step S5 is the Pearson correlation coefficient algorithm, which adopts ergodic calculation and selects the vibration mode corresponding to the overall vibration frequency spectrum in the historical data with the highest similarity calculation result as the judgment result.

4. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 1, characterized in that, The combined computing network is a three-layer fully-connected feedforward neural network.

5. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 1, characterized in that, In the step S1, the cabin vibration monitoring data at multiple points are aligned on the time axis in the time domain.

6. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 1, characterized in that, In step S3, when performing FFT analysis, the frequencies whose maximum amplitude is less than a preset threshold are filtered out; the preset threshold is set according to the cabin structure.

7. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 6, characterized in that, The preset threshold is 1 / 20 of the fault threshold, and the fault threshold is the minimum vibration amplitude that causes a fault and is set according to the cabin structure.

8. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 1, characterized in that, In the historical data, the overall vibration frequency spectrum and the vibration mode are stored in pairs.

9. The analysis method based on SCADA system cabin vibration monitoring data as claimed in claim 1, characterized in that, In step S5, the membership function of the fuzzy algorithm is a single variable piecewise function.

Citation Information

Patent Citations

  • D-S evidence theory-based high-voltage circuit breaker mechanical fault diagnosis method

    CN110119713A

  • Wind turbine generator transmission chain fault early warning method based on big data analysis

    CN110926809A

  • System and method for intelligent ship propeller rudder rotation vibration detection

    CN110979592A

  • Rolling bearing fault diagnosis method based on VMD and deep convolutional neural network

    CN112903294A