E.S. Furgason

Flaw-to-grain echo enhancement by split-spectrum processing

Abstract A split-spectrum processing technique for an ultrasonic flaw detection system has been developed which improves the flaw-to-grain echo ratio in large-grained materials. The enhancement is achieved by partitioning a wide-band received spectrum to obtain frequency shifted bands, which are then processed to suppress the grain echoes with respect to the flaw echo, using a novel signal minimization algorithm. Experimental data for titanium and stainless steel samples are presented which show superior flaw detection capabilities for the minimization algorithm with respect to frequency averaging techniques.

Application of random signal correlation techniques to ultrasonic flaw detection

Abstract A system is described and analysed which applies random signal correlation techniques to ultrasonic flaw detection. The use of correlation and time integration techniques gives it a signal-to-noise ratio correlation gain of the order of 10 4 , and the use of a random transmitted signal whose range resolution is independent of its duration permits a peak-to-average transmitted power ratio of the order of unity. It is shown that random signal correlation systems should therefore be capable of much greater range and/or resolution than is possible with current pulse-echo systems.

High-speed digital Golay code flaw detection system

Abstract A new, entirely digital, correlation flaw detection system is described which maintains optimal correlation detection over a much wider range of input signal-to-noise ratio conditions and operating speeds than previous single transducer correlation flaw detection systems. This new system uses special pseudo-random codes called complementary Golay codes to overcome the problem of self-noise which results from incomplete pulse-compression in the correlation process. Results are presented which show that the Golay code system produces better performance in the presence of grains and large reflecting surfaces, while offering much higher operating speed than previous single transducer correlation flaw detection systems.

Golay Codes for Simultaneous Multi-Mode Operation in Phased Arrays

A multi-mode phased array system is discussed which would simultaneously transmit a set of special pseudo-random codes called Golay codes and then isolate the coded echoes, and their associated modes (e.g. scan directions), upon receg tion using matched correlation receivers. Golay codes are special paired pseudo-random codes which have been shown to provide improved single-mode correlation system operation in the presence of clutter and large interfering targets due to the cancellation of range sidelobes. An algorithm is presented for synthesizing pairs of complementary Golay codes which also have the special property of producing zero cross-correlation during simultaneous transmission. A method for creating a larger, eight member, set of low cross-correlation Golay codes consisting of pairs which have zero cross-correlation, is also described.

Ultrasound Imaging Through Highly Reverberant Thin Layers

Nondestructive testing of targets c onsisting of a finite number of thin layers using backscattered ultrasound echoes is highly desirable in material evaluation. These targets can be examined by an ultrasound pulse-echo system from uhich acoustical parameters sensitive to the inhomogeneity of the samples are extracted. In practice, extracting these parameters is difficult if nor impossible because of: l) limitation o f system spatial resolution, and 2) multiple reflection uithin the target. In this study the system resolution has been improved by using transducers uith the largest possible banduidth taking into account the increase in absorption and scattering, and applying a subtraction technique uhen the signal-to-noise ratio is high. In addition, the r everberation echoes are also identified and classified by decoupling the components of the backscattered echoes. These techniques have been adapted for measuring support

The use of correlation systems for real-time ultrasonic imaging

In previous studies, we have shown that correlation systems can overcome the peak-power limitations of conventional pulse-echo systems in non-moving target applications. This study examines the applicability of correlation systems to practical real-time imaging situations involving moving targets and clutter. The study was carried out through simulations of the generalized ambiguity function and a derived system signal-to-noise ratio formula. Results indicate that whether a correlation system is better than a conventional system depends on the target velocity and the clutter-to-noise ratio, and that special pseudo-random Golay codes provide the best overall- .correlation system performance. Although the special Golay code properties of self-noise and crosscorrelation cancellation are diminished as the velocity of a target increases, the degradation of system performance was found to be minimal for all but the fastest moving targets in medical imaging applications.

Optimal Operation of Ultrasonic Correlation Systems

Correlation systems for use in ultrasonic nondestructive evaluation have been slowly evolving toward practical, general purpose data acquisition systems. Since these systems represent a significant departure from the design of conventional ultrasonic detection systems, the basic operating principles of ultrasonic correlation systems are reviewed in order to help identify the principal strengths and weaknesses of previously reported system implementations. The dependence of system performance on transmit signal code characteristics, modulation technique, and system clock frequency is evaluated and a general formula for the signal-to-noise ratio enhancement available from correlation type detection systems is developed which includes the effects of grain clutter. The results of recent research directed toward development of

Acoustic Emission from Partial Discharge in Encapsulated Voids

BACKGROUND The techniques currently used to assess partial discharge activity in power transformers rely on electrical measurements performed at the terminals. Although such measurements can detect aggregate discharge activity, they give little information regarding the location of individual discharge sites and only a crude estimate of the actual discharge activity. In modern encapsulated power transformers, where size and weight limitations lead to stringent requirements of dielectric integrity, such terminal measurements should be supplemented by other techniques in order to provide better quality control of production and better incipient fault detection. An acoustic emission detection system is being developed to locate and evaluate individual partial discharge sites in encapsulated transformers. Experiments in the frequency range of 130 to 170 kHz have been carried out in a specially designed, intFgrated test cell which show the relationship between conventional electrical measurements and acoustic emissions for a range of void sizes. The significance of these two sets of measurements in the evaluation of a practical acoustic emission detection system will be discussed.

Random Signal Flaw Detection

Abstract : This report describes four areas of work. In the first area random signal beam forming techniques are developed which make it possible to transmit several independent beams from a phased array using the same number of incoherent noise sources. Following this ultrasound beam scanning techniques are demonstrated in which random signal correlation systems and RF correlation systems using quadrature phase techniques are used in methods for scanning the intensity profile of ultrasound transducer beams using respectively broadband signals and narrow band signals. In addition two types of deconvolution procedure are described and demonstrated which have been successful in improving the effective resolution of ultrasound correlation systems when viewing surface features. Finally the design and construction of a random signal correlation system is described which uses an electronic delay line. (Author)

Robust Deconvolution of Correlation Functions

A new generalized cross-correlator (GCC) is presented which can provide greatly improved resolution in delay estimation measurements. Generalized crosscorrelation is a technique in which the input signals are prefiltered, in some optimal manner, before crosscorrelating. The prefilters for this GCC are robust (in the minimax sense) linear filters which have found use in linear equalization. The presence of the prefilters, in effect, deconvolves the cross-correlation function. An example of the use of this new GCC is provided by locating acoustic emissions from partial discharge in a high voltage, solid cast transformer. Partial discharge is an electrical phenomena caused by defects in the insulation systems of transformers and is observed as both electrical and acoustic emissions. Time delay estimation techniques used to locate the emissions sites in power transformers are complicated by the multipath signals that are present. Experimental measurements of the acoustic emissions are used to illustrate the effectiveness of this new GCC in combating the problems associated with multipath signals.