| Titanium steel composite plate has been widely used in electric power, chemical industry, shipbuilding and other fields, the current domestic electric power design institute design of chimney steel inner cylinder most choose titanium steel composite plate, because of its good corrosion resistance and low material cost. Titanium steel composite plate in power but in order to protect the atmosphere, plant or Electric Power Research Institute widely used wet desulfurization of flue gas, dilute sulphuric acid solution which would produce a certain concentration, and the corrosion of titanium steel composite board, if the titanium steel composite board has many defects of welding in welding, will cause the safety of hidden dangers, so the detection of weld steel composite plate is imperative. Now, positioning and quantitative method of weld defect is relatively mature, and the qualitative analysis of the defect has not achieved satisfactory results. Now, therefore, the qualitative recognition of weld defects is important, effective operation and service life of power plant chimney.In this paper, the nondestructive testing and ultrasonic testing technology of the principle of development, has made the analysis and the summary, and connecting with the actual industrial background noise, respectively based on probabilistic neural network(PNN) is proposed feature detection identification technology and weak signal detection identification technology based on stochastic resonance. Multiple feature detection identification technology based on PNN cascade atoms library much echo feature extracting method has greatly increased the recognition accuracy, and PNN network frequency characteristic of the detected based on stochastic resonance detection identification technology provide a priori information, the realization of can greatly improve the detection rate and accuracy. These two kinds of technology for titanium pipe steel composite board qualitative analysis provides a feasible method, ultrasonic testing has important practical significance.Chimney of multipath effect caused by ultrasonic, a weak mobile handheld ultrasound device causes the doppler frequency shift, and industrial sites of background noise include pulse noise and so on, makes the received echo signal signal to noise ratio is reduced, and the stochastic resonance in weak signal extraction has a unique advantage. Due to the lack of penetration, crack, bubble, slag four ground loss echo of different frequency components, through the above method to obtain the frequency under the condition of a priori information, this article in view of the weak in strong noise with multiple frequency components, and not meet the characteristics of the adiabatic approximation assumptions, establish a frequency sweep type stochastic resonance detection system, and puts forward a kind of big parameters multi-frequency signal detecting method is used to identify the chimney four kinds of damage.Experiments in this paper, the use of CTS-9002 ultrasonic flaw detector, a given weld flaw detection test on the spot, the qualitative analysis for post weld provide rich data. Experimental results show that under the actual background noise detection, the two techniques are successfully detect the lack of penetration, crack, bubble, slag inclusion, four kinds of injuries, clear information for chimney repair work. |