FEATURES OF DIAGNOSING THE TECHNICAL CONDITION OF TECHNOLOGICAL PIPELINES BY LOW-FREQUENCY DIRECTED WAVES ON THE BASIS OF THEORY PATTERN RECOGNITION

Abstract

Based on the analysis of theoretical foundations and practical aspects of methods and basic principles of image recognition systems in various fields of science and technology, the essence of the basic concepts of image recognition methodology in relation to its use in assessing the technical condition of industrial facilities by non-destructive testing and technical diagnostics. The essence of the concept of "image" and its use in sonar to obtain images of acoustic images in the detection of submarines and registration of seismic signals in the search for oil fields. The mathematical problem of classification of images based on separate functions is given, which consists in division of all space of signs of images to corresponding classes with application of elements of threshold logic on dividing borders between classes.

Based on the developed classification of reflectors of low-frequency ultrasonic directional wave, the classification of the set of control images of reflectors of the process pipeline is substantiated, which are divided into corresponding subsets of control images belonging to three clusters (groups). Components of clusters are subsets of control images that reflect, respectively, technological, structural and operational reflectors of the technological pipeline. The subsets of control images of each cluster divided into symmetric and asymmetric, respectively. To apply the theory of pattern recognition in the procedure of diagnosing the technical condition of the technological pipeline by low-frequency control systems with directional waves, the basic concepts of control and diagnostic images, features of these images in relation to operational reflectors of the pipeline.

Based on the analysis of experimental studies presented in foreign scientific journals, the defining boundaries of diagnostic image features, which adopted in low-frequency control systems by directional waves of long pipelines and which correspond to metal losses in the cross section of the pipe wall in three classes. Such separating limits are set at 9% and 12% of metal losses in the cross section of the pipe wall.

The analysis of the procedure of control and diagnostic process in determining the technical condition of technological pipelines using low-frequency ultrasonic control systems by directional waves, which is based on the operation of identifying metal losses of the pipe wall during the operation of the pipeline. The structural scheme of the procedure of control-diagnostic process developed, which based on the rule of threshold identification of defects with the use of logical binary information coding system, which realized by means of threshold elements.

The method of preliminary presentation of subsets of control images obtained from technological and structural reflectors of the pipeline and subsets of operational control images divided into three classes described in the development of the scale format of numerous superimposed A-scans on the «Teletest» display, control of the pipeline section.

Features of identification of metal losses in the cross section of the pipe wall due to the defect analyzed. It is shown that according to the control of the technical condition of the pipeline using low-frequency control system by directional waves it is impossible to substantiate information about the depth of the pipe wall, which can only give physical methods of measuring wall thickness using thickness gauges. A scheme of the pipe wall metal loss estimation algorithm has been developed, which illustrates that the conclusion about the presence of a defect and the degree of damage to the pipe wall using a low-frequency directional wave control system is based on the amplitude of the reflected signal. However, such an assessment is not direct, ie qualitative. The amount of thinning of the pipe wall, measured by a thickness gauge, is direct, ie quantitative. There is no direct correlation between them, because different physical methods and means of evaluation are used.