Auto-body OCMM Measuring Data Automatic Alarm Study And Application

With a wide range of applications on body-line detection system and the rising requirements of car body manufacturing quality, the application method study of auto-body dimensional online detecting data is also getting more and more attention from industry and academia. Effective analysis and accurate alarm to car auto-body dimensional data of online detecting both to improve the important means of the body manufacturing quality, but also put forward specific objective evidence of corrective actions. In this paper, study on analysis and application of online detecting data. The focus is to build an efficient, accurate and useful automatic alarm method for auto-body online detecting data.This paper obeys the philosophy of"Data Driving Quality", and follows the route of"Measurement----Evaluation----Control"in auto body manufacturing quality control. New data alarming approach in auto-body assembly process is proposed here, which enriches the theory on auto-body manufacturing quality control and does favor to improve the auto body manufacturing quality effectively. The research work in this paper was focused on the several points as follows:1. Analysis of auto-body dimensional online detecting system and characteristics of data. According to the characteristics of online detection system, using the theory of measuring points and the body size of the required functional testing point line BIW selection and optimization; on-line detection accuracy of the data analysis measuring point and measuring point coordinates of the selected calibration coordinate a unified body department; the characteristics of online detection of abnormal body size of several main types of classification and sources of data anomalies.2. The establishment of control charts based association analysis and distribution pattern matching method for the online test data alarm. On-line analysis of the characteristics of test data summarized based on the proposed line detection method of abnormal data, automatic alarm, and applied validation; to the association of different control chart analysis of data from the body in white line detection of different abnormalities in the screening Alarm data; on the issue of cyclical fluctuations in the distribution pattern matching analysis, screening out of data anomalies and the police. Formed a set of alarm method based on control chart association analysis and distribution pattern matching for online detection data.3. Based on time series pattern recognition diagnosis of body size deviation. Proposed on-line test data pattern recognition theory based on time series of body size deviation diagnostic methods, and to be verified. Body in white-line detection system error diagnosis, classification of the error sources, error state, by the size of the corresponding deviations of data, the establishment of the corresponding AR model, the geometric distance function obtained through the geometric distance as a reference, and to establish appropriate deviation collection. When the daily production process in the workshop can be detected on-line real-time data modeling, obtained geometric distance, and then under different bias compare the geometric distance to determine deviation of the white body size of the problem source, rapid diagnostic bias source.4. Online detection method for the case of the alarm application. After the drop tank for the white part of the body-line change detection data fluctuating problems associated with the use of control chart-based analysis method for online detection alarm the alarm on the original test data pattern recognition-based diagnostic method of body size deviation analysis. First, the data pretreatment method of online data through the alarm the alarm, after the fall to determine the type of tank size fluctuations, and then use time series model off the tank after the different theories on the causes of fluctuations in the type of time series modeling, and the help of principal component analysis to determine the size of fluctuations in the sink after the fall the main error source, so that data off-line detection of fluctuations in the tank to be solved to prove that association analysis based on control chart for body in white-line test data alert methods and white based on pattern recognition body size deviation diagnosis method.