Research On The Ultrasonic Detection And Its Signal Processing Of Heat Exchanger Fouling

Heat exchanger fouling problems generally exist in electric power, chemical, metallurgical, aviation and other industrial production fields. Many direct detection methods of heat exchanger fouling have been applied, but most of them can not realize the real-time online detection; thermal method is used widely, but it also has disadvantages such as low reliability. Finding a highly reliable, simple, practicable, real-time online accurate quantitative detection method of heat exchanger fouling is always a hotspot in the fouling research field.Aiming at this problem, this dissertation adopts the ultrasonic time-domain reflectometry technology to detect fouling. Meanwhile, the ultrasonic detection signal denoising and time of flight(TOF) acquisition also were studied in this dissertation. The main research work as the following:Based on the analysis of related fouling theory and ultrasonic detection method, overall structure design of heat exchanger fouling ultrasonic detection system has been set up. The tests of sonic speed calibration and ultrasonic detection of heat exchanger fouling on plate and pipe were conducted. The tests prove that using ultrasonic time-domain reflectometry method to detect the heat exchanger fouling has better feasibility.Due to characteristics of the heat exchanger fouling and interference of the detection environment, the ultrasonic detection echo signal usually contains lots of noise and the further study is effected. This dissertation proposed the 2-D wave atoms hard threshold denoising method combined with non-local mean filtering on the basis of the characteristics of heat exchanger fouling ulrasonic detection signal. The simulation results show this method can realize the ultrasonic signal denosing, and has better results on keeping the signal detail information.Accurate extraction of TOF is the premise of ultrasonic time-domain reflectometry reliability. The method to extract TOF of ultrasonic detection signal without serious waveform aliasing was studied. The result shows the improved Fruit Fly Optimization Matching Pursuit Algorithm is the better method to extract the echo TOF information of heat exchanger fouling ultrasonic detection. The experiments verified that this algorithm has accurate waveform extraction ability and fast response speed.Because the heat exchanger fouling thickness is too small or there exists the ultrasonic high-order echo interference, the heat exchanger fouling ultrasonic detection signal may include a serious waveform aliasing in time-domain. The previous method is not suitable for serious aliasing ultrasonic detection signal processing. This dissertation proposed to extract TOF by using the combination Winener deconvolution and high order cumulant MUSIC spectrum estimation when serious aliasing ultrasonic detection signal exists. Increment iteration method was used to correct the accuracy of the reference signal. The simulation experiment shows this method has better measurement precision, anti-noise ability and practicability.Using the proposed ultrasonic detection signal processing method, ultrasonic detection experiments data accrued early was analyzed and processed. The processing results were analyzed combining with the related fouling treatment theories. The results show that this heat exchanger fouling ultrasonic detection method and related signal processing method have better precision and practicability.