Abnormal state detection method and detection system of mechanical parts
A technology of mechanical components and detection methods, applied in the field of data analysis, can solve problems such as false positive detection errors, and achieve the effects of improving detection, avoiding detection anomalies, and improving signal-to-noise ratio.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0036]This embodiment provides a method for abnormal detection of mechanical parts, such as figure 1 and figure 2 As shown in Fig. 1 , the first part to be tested is the bearing set on the motor, and the detected value is the temperature value. For the abnormality of the generator bearing temperature, the temperature rise of the front and rear bearings of the generator can be used as a feature.
[0037] The whole method includes the training step of calculating the comparison reference amount and the prediction step of using the data calculated in the training step as a reference and basis for calculation, wherein the training step is specifically:
[0038] (1) Obtain the monitoring data of the equipment sensor, and select a group of data of earlier equipment and larger data volume as the training data.
[0039] (2) Then the characteristic variable is discretized with a fixed width or a non-fixed width. With a width of 10, the value of the bearing temperature range of [20,1...
Embodiment 2
[0052] A detection system applying the above detection method, including a data acquisition module, a data transmission module and a data analysis module, the data acquisition module is a sensor arranged on a mechanical component, and the characteristic value is collected by the sensor and passed through the data transmission module Pass it to the data analysis module for calculation and judge whether it is abnormal according to the preset threshold.
Embodiment 3
[0054] This embodiment is an example of using simulated data to preview and deduce the simulated calculation process, taking the training process as an example:
[0055] 1. Assuming data data, wherein the characteristic value is temperature rise, two temperature monitoring points are set on the target component, and then the temperature data is monitored in real time.
[0056] time T 1
T 2
2018-01-01 01:00:00 51.2 48.3 2018-01-01 01:00:01 52.1 49.0 2018-01-01 01:00:02 54.3 49.3
[0057] Here is just an example of the first three sets of data, that is to say, the temperature data is collected and recorded every second, and then through T 1 -T 2 = temperature rise to calculate.
[0058] 2. Discretize the temperature rise data. In this example, the temperature rise dispersion width (bin) is taken as 0.5, such as temperature rise 1.3, after discretization is 1, temperature rise 2.83, after discretization is 3, temperature rise 1.57, after...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


