Advances in NDT

In many types of NDT, the sensitivity of the technique depends upon the ability to distinguish the significant part of the signal from the general background due to electronic noise, or inherent background signal from the material being examined.. With the power and speed of modern computers, signal processing methods and modelling numerical methods, remarkable developments took place in research on NDT techniques Significant improvements have been made both in the NDT equipment and in the specific techniques used (16). Thus, defects that may not have been revealed by NDT performed five, ten or 40 years may be detected by more sophisticated NDT equipment or techniques currently available.

The results of reliability studies indicate that the probability of detecting a defect with ultrasonics increases with the degree of sophistication of the system. (manual ultrasonics, without sophistication, can be expected to reject an equal or greater percentage of the discontinuities present than will radiography). Manual ultrasonic systems relaying on 20dB or 6dB drop are known to be inaccurate. The incorporation of computer assisted processing into ultrasonic systems has allowed the easy implementation of potentially better methods for defect detection and sizing such as SAFT, ALOK, TOFD etc. (17) Similarly, use of multiple NDT sensors, NDT techniques and computer assisted processing in modern NDI systems have reduced costs by increasing both the speed and reliability of inspectionHuman ComponentNDT inspections are performed over extended periods of time. In addition to keeping track of areas inspected and to be inspected, the operator must remain alert to possible signs of a discontinuity.

As a consequence, inspection reliability depends significantly on the operator.The three main components of human performance are the person, the activity and the environment. One can expect ideal performance when the person is highly skilled and motivated, the activity is familiar and satisfying and the environmental conditions are favourable. These three requirements are not likely to be satisfied simultaneously (18).Manual scanning requires skill and probe movement control. In addition to scanning, the human operator must exhibit vigilance in observing the flaw detector screen for over long, unbroken, periods of time to detect small changes in the information. Maximum human performance can not be expected to persist over time. As a consequence, inspection reliability depends on the performance and judgement of the operator carrying out that inspection.One question that arises is that why not just remove the human from the inspection routine ?

In many instances this is possible, especially in manufacturing processes. An interesting example is quality assurance of hardened components. One industrially important heat treatment is case hardening of steel, which is carried out to improve surface properties. For consistent quality it is important to ensure a constant case depth. This is in general is measured by conventional optical microscopy which slows down the production process and does not ensure consistency of the case depth quality. The main effect of the treatment is to produce a hard, highly strained martensitic layer which, unlike the ferrite/pearlite, is unable to support the 90° domain wall motion that generates MAE. Hence, MAE measurements can be used to measure the case depth on-line.

In many instances, the case depth deduced from the MAE measurements was in good agreement with the case depth deduced by optical microscopy. Primarily in a manufacturing/quality control/quality assurance setting, when inspections can be automated, a great deal of consistency gained over human operators, provided engineering factors are given the appropriate emphasis. However, automated inspection, by definition, require each and every component to be identical. When each component is different from the other, the automated inspection becomes impractical (19).The first step in the development of an inspection procedure is to anticipate the types of discontinuities that may be present and determine whether they may ultimately interfere with the service requirements of the test piece. Discontinuities may arise from (raw) material selection, manufacturing process, handling, geometric configurations, service loads and environmental conditions

(20). They may be localised or they may span a larger volume of the test piece.During manufacturing stages, NDT can be used to detect defects at various stages during production, by means of off-line measurements, for example, before and after each set of welds, on the parent plate before welding, on the billets before rolling or forging or on the steel slabs after casting etc. While there are clear benefits in this approach, it has the drawback that off-line measurements interrupt the process stream and reduce the productivity

(21). A better way would be monitor for incipient defect formation on-line. On-line monitoring and control of the welding process has the potential to improve weld quality and increase productivity in the automated welding. Weld monitoring and control can be achieved by the integration of real time non-destructive evaluation techniques with the welding process. In-production weld inspection can improve weld quality and may provide a significant cost reduction. The welding parameters can usually be adjusted to prevent defects from forming. Furthermore, if welding defects do occur the flaws can be found and repaired before they are covered by subsequent welding passes, leading to a decrease in the level of post-weld inspection and repair. Good quality welds rely on the correct weld pool size, geometry and position relative to the weld preparation.

NDE sensors provide information on the state of the weld pool. This information, together with critical welding parameters such as current, voltage, torch position and travel speed, is used to adjust the welding process to maintain desired stable process with little operator intervention.The major advantages of the in-process method are cost and speed. Conventional NDT is relatively slow, because of the need to position the sensor on the spot and make measurements for each spot weld. As a result, 100% testing may be more time consuming than manufacturing. For example, for components with large numbers of spot welds, in-process techniques provide a measure of weld quality in fractions of a second and can be automated.For on-line material monitoring, when the parameters affecting the NDT are limited and identified, it is possible to establish a viable empirical approach, provided experimental relationships can be established between readily measured NDT parameters and the desired materials properties.

The final step is process control, preferable by means of a closed feed-back loop in real-time. Provided a reliable relationship is found between measured NDT parameter and process variables, the technique can be used to improve the product quality and process efficiency (22).The on-line approach offers the earliest possible warning of problems so that a remedial action can be taken in the most cost effective manner, with hopefully minimum scrap. It will be also possible to use the information about the presence of defects to adjust the process parameters or the feed-stock, in order to minimise or even prevent further defect formation either open-loop or closed loop control. It has been demonstrated that NDT techniques such as ultrasonics and magnetics are sensitive not only to defects but also to certain material properties such as such as homogeneity, grain size, texture, elastic modules, plasticity, hardness, stress and temperature. For example, If there exists 1:1 relationships between measured parameters such as ultrasonic velocity and material properties and in turn, between material properties and mechanical properties such as strength, fracture toughness, then the application would be straight forward.

A reliable on-line inspection methodology can be accomplished with a sensing heat suitable for application to the production line without adversely affecting product quality or productivity. Ideally, the NDT sensing head should be non-contact, robust and capable of processing the data sufficiently rapidly to guide the remedial action. This is especially important if the goal is closed-loop control, when a high degree of automation is desired.Fortunately, a wide range of NDT technologies developed for more traditional post-manufacture NDT, can be used, in principle, during the manufacturing process and also in-serviceIt is likely that when an automated process indicates an anomaly, human inspectors will be sent for manual verification before rejecting the component. At this moment it is essential to note that the human inspector has two key qualities than an automated system does not: The ability to adopt to the requirements of the individual items being inspected, and the ability to judge whether an indication is in fact a discontinuity or not. UT Vs AutomationUltrasonic waves have great potential for use in-process NDT (23). First UT velocity and attenuation give important information on material property changes during processing. Secondly, they give indirect information about temperature, pressure, flow etc. Thirdly, ultrasonic waves can be used on-line for inspection of part quality. Measurement of process parameters and materials properties makes possible the control process variables to achieve the required material properties.RT Vs AutomationThe advantages of real-time radiography are on-line testing of defect formation in the weld and study of metal fusion, filler-metal-to-base metal interaction, metal transfer and mass flow in the welding pool and the application of this information to welding process control (24). The welding current can be automatically controlled as a function of defect-feature extraction from computer processing of weld images. While the infrared and optical sensing and control systems have the disadvantage that the information on weld quality is indirect, closed-loop intelligent process controls have been demonstrated to characterise the weld penetration through the depth of the weld pool. The information extracted from real-time radiographic images about weld quality, supplemented by sensor data on weld current and voltage, is used for weld power-supply control (25).Pros & Cons of AutomationAutomation is desirable when high confidence, highly repeatable inspections must be performed in a timely fashion while recording and analysing a larger amount of raw data. Today, successful completion of an automation task can be strongly influenced by rapidly changing computer technology and the human element involved in the transition from manual to automated procedure. For these reasons, the development and integration of automated inspection technology into routine use should proceed on a phased basis. The pioneering status of current technology means that the development and application of an automated inspection system incurs significant costs and risks. Therefore, the decision to automate should be made after assuring that it is the optimum solution for the particular problem of concern. Listed below are the reasons that often justify pursuing the automated approach:1) Full coverage of the item to be inspected is demonstrated, recorded and repeatable2) Automated data acquisition and analysis permits working at higher sensitivity because the consequent increase amount of data can be handled rapidly with modern computational hardware.3) Collecting and storing position annotated inspection signal information in a computer compatible format greatly increases signal interpretation options.4) For nuclear systems one large benefit is the substantial reduction in radiation exposure.The elements which must be fully integrated to produce an automated inspection system are 1) Electronic hardware 2) Software3) Transducer position4) Signal transmitter and receiver5) TransducerThe challenge to the developer is to combine these elements into a well integrated system that performs inspection rapidly, efficiently and with great reliability. Two useful aspects for successful development of automated inspection systems are 1) involving the end user early and periodically during the development cycle and 2) system performance verification by someone independent of both developer and end user.