Ultrasonic In-line Inspection Data Compression Study Of Pipeline Corrosion Defect

As the two methods, data pre-processing and direct data compression, are all to improve the degree of compression of the raw data to solve the problem of data mass storage. As data pre-processing is to filter the data before the data saving, and the direct data compression is to compress the stored data, the two methods can be considered together, we can compress the filtered defect data after pre-processing (filtering the data detected).The data volume of an A-scan waveform mainly depends on the sampling frequency and sampling time, in order to make the data complete and clear we need to use higher sampling frequency and sampling time. It will become a very difficult problem becaouse promoting the sampling frequency and sampling time will lead to a significant increase in the data collecting volume. Being constrained by the space of the pipeline, pipeline inspection robot must be very compact in structure and has a high integration level. Therefore, the cache capacity of pipeline robot's information processing and calculation unit would be greatly constrained. A simple, fast and effective compression algorithm becomes a very important key technology to solve the massive pipeline information storage problem. In order to solve the mass pipeline ultrasonic data storage problem, a domestic plan has been proposed to solve the problem, mainly made up with two methods, data preprocessing and direct data compression.Based on this, in this paper a better pipeline data online compression program is brought up, specifically in 2 steps:1. Using data correlation coefficient theory to filter the raw data. Compared with the thickness feature comparison, the data correlation theory has it own advantage, since it is based on the calculation of the similarity between the detected waveform data and the standard waveform data (The pipe has no distortion and there is no corrosion inside and outside of the pipe wall) to determine and filter the data, rather than on the error-prone thickness feature comparison theory. In theory, It fills up a deficiency effectively that the complex pit corrosion can not be determined by the thickness feature comparison method.2. Compress the filtered defect data by and save the compressed data. The method is effective filling up to the data correlation coefficient method, if only using the data-correlation method to reduce the total data volume, but not storing it, the pipeline corrosion defect can not be determined only by some single numeral because there is no waveform data. And if not using data correlation coefficient theory to determine whether there is defect in pipe, only using PSD particular frequency band storing method, the compression rate is extremely limited. The two methods are interdependent.