NDTech Online - Current Newsletter

NDTech Newsletter #14

New IR Camera from Inframetrics

An innovative product from Inframetrics of North Billerica, MA is the SC1000 ThermaCAM™ handheld infrared camera. This camera can operate as a standalone camera or as a dedicated real-time digitial system when interfaced with the company's ThermaGRAM Pro 95 real-time digital image processing system for Windows 95.

The SC1000 was developed for users in the scientific market who require high sensitivity and advanced image processing capabilities to support their infrared imaging and temperature analysis applications. The unit can provide screen temperature measurement accuracy of better than +2% or 2 degrees Celsius. It can operate continuously from a single standard camcorder battery for more than two hours, and an optional battery belt will run the camera for 12 hours.

The focal plane array radiometer weighs six pounds with the power source and has FLASH PCMCIA compatibility for storage of up to 256 images on a single card. Remote control and focus from a PC or handheld controller is possible via a serial remote control port.

Inframetrics, established 1975, supplies other thermal imaging and temperature measurement systems, components, and accessories to scientific, industrial, aerospace, law enforcement, and military markets.

For more information on Inframetrics, contact the company at:

16 Esquire Rd., North Billerica, MA 01862-2598

Telephone: 508-670-5555

Fax: 508-667-2702

Website: http://www.inframetrics.com

Acoustic-Emission Bearing-Fault Diagnostics System

A new hardware and software system that uses a sate-of-the-art, high-frequency Acoustic Emissions (AE) sensore and an innovative AE signal-processing technique, called Point Process Spectral Analysis (PPSA), has been developed to help prevent catastrophic failures and costly down time due to false alarms during bearing testing. In the past, bearing health monitoring and fault diagnosis within turbomachinery and drive-train systems have been a significant technical challenge for the aeronautics and transportation industries.

Previous techniques to detect bearing faults could not distinguish between transientevents related to shaft rotational processes and the signatures associated with defective bearings. Also, these techniques did not involve high-frequency, real-time analysis. To overcome these unique problems, PPSA was developed to meet the high-frequency AE signal processing and fault detection requirements.

Conventional time series representation of an ultra-high frequency AE signal requires all data to be sampled over the entire waveform at a high sampling rate. However, PPSA only uses the times occurence of the transient events, along with their strengths, since these transient events contain the major dynamic information needed for bearing fault detection. As a result, PPSA requires much less data to analyze the frequency-domain behavior than conventional time-series representation. Additionally, PPSA overcomes the basic limitations of the fast-Fourier-transform-based spectrum for detecting signal components, such as widely spaced, narrow transient spike that cannot be approximated effectively by a sum of sinusoids.

PPSA uses the first moment function of time to determine the rate of transient impulse, threshholding the temporal AE waveforms for structural failure detection. Next, a mean-lag-jump product, representing the second moment of a point process, provides a statistical estimation of the correlation between all pairs of event occurrences with common time lags. The frequency-domain representation of the second-moment mean-lag-jump product function is used to create a spectrumlike function of the point-process. This point-process spectrum provides a statistical estimation of the event occurrence rate and intensity distribution as a function of frequency.

The superior detection capability of PPSA over conventional envelope analysis in extracting bearing signatures from AE transient signal in a noisy operational enviroment was demonstrated with two computer-simulation examples and with NASA-providing test data from a bearing test rig. Three test conditions were used: a good bearing, an inner rade defect, and a roller defect. PPSA successfully analyzed the data. The results of these proof-of-concept studies indicated that PPSA can provide high computational efficiency in processing ultrahigh-frequency AE signals and is highly suitable for real-time implementtation. Using this analysis technique qould significantly reduce the digital signal processor requirement in developing a low-cost, commercially viable, on-line bearing-diagnostic system.

This work was done by Jen-Yi Jong of AI Signal Research, Inc., for the Marshall Space Flight Center. For further information, access the Technical Support Package (TSP) free on-line at http://www.nasatech.com under the Machinery and Automation category.

Finding Only the Data You Need

If nothing else, the World Wide Web has provided everyone with almost instantaneous access to a global information database. The problem is that now there's so much information out there that it's become nearly impossible to efficiently sort through all of it to find what you really want and need. WWW-based search engines can help, but because of the internet's broad appeal, most search engines developed for it are too broad for the needs of scientists and engineers.

This has become especially important for researchers who need to quickly find data for their new-product and process-development efforts or who want to build upon the results made by other researchers.

Enter scientific on-line resources, such as the recently announced ScienceBase from Knight-Ridder Information, Mountain View, Calif. ScienceBase is a World Wide Web application designed to give researchers easy and intelligent access to published scientific literature. While ScienceBase doesn't search the Web, users can search through industry periodicals, patent documents, conference papers, scientific journals, supplier and buyer listings, and a number of scientific databases.

While this may not sound much different than other database search engines other than having a broader number of available resources, ScienceBase's main advantage lies in a built-in search intelligence that executes expert search commands behind the scenes to retrieve the desired information. You don't need to be a search expert or take any training courses as you do for other sophisticated database search resources. ScienceBase automatically selects the appropriate databases, looks for related items, refines your search according to your selection criteria, and then presents the data in a logical format. To speed things along, it even searches multiple databases simultaneously.

"We worked with our information-technology professionals to get around the database vagaries that exist from one database to another," says Larry French, director of product development at Knight-Ridder. Since the indexing conventions used by each of the 90 databases available to ScienceBase could be different from one to the other, the search engine developers found they had to modify the queries posed to each database to optimize the user's specific search. The developers also had to establish an intelligent way to automatically select the best databases to search for each query.

To provide its users with the best value, ScienceBase only draws from established reputable databases, including Elsevier Science Publishers, Derwent, Merck, NTIS, Business Wire, CAS, USPTO, The Royal Society of Chemistry, and 82 others.

Database queries are easily accomplished. The searcher starts by selecting a category of interest, then clicks on a hotlink to see the topics available. Subject categories contain intelligent query templatesthat you customize; they also perform quick search survey options that you define.

After you enter your topic interest, you can make additional refinements from a pull-down menu and determine the number of items you want to view. You click on the "Submit Query" button and let ScienceBase do the rest. Cost of each article or record is displayed on each abstract-no charges are incurred until you make a selection.

"We wanted to minimize the amount of unnecesary information or false hits that a user would get in a search," says French. "Everything a user gets should be right on."

For more information:

Telephone: 415-254-7000

Website: http://krscience.dialog.com

Pulse-Echo Ultrasonic Imaging Method for Eliminating Sample Thickness Variation Effects (U.S. Patent No. 5,629,865) Inventor: Don J. Roth, Lewis Research Center

This method for nondestructive ultra-sonic evaluation of materials measure velocity using a single transducer pulse-echo immersion systems, automatic scanning, and digital imaging, providing a video image of the sample in color or gray scale that is a map of a material property such as porosity fraction. In the immersion method, the material to be evaluated is surrounded by a liquid and positioned over an acoustic reflector that is also immersed in the liquid An ultrasonic wave of a known frequency is transmitted through the liquid and four separate echoes are recorded and evaluated at each scan point. Each echo is recieved as an analog waveform that is digitized and stored in a computer. Before the evaluation scans, nonlevelness and sample thickness variations are accounted for and eliminated by pre-scans to insure that the recieved reflections or echoes are within their set time windows to provide a complete waveform for evaluation and cross-correlation to acurrately obtain the time-delay data used in calculating the velocity values.

Machine-Vision System Maps Surface Flaws

A machine-vision system maps surface flaws generated in fretting of a metal panel or other structural components. The flaws include pits, gouges, scratches, small protuberances and other features that deviate >from the otherwise smooth or planar surfaces. The machine-vision system should be useful in many mass-production situations in which there is a need for quick and accurate detections of flaws on machined, polished and otherwise smooth surfaces.

The machine-vision system includes a black-and-white video camera aimed perpendicularly to the surface to be inspected. The surface is illuminated, from opposite sides of the line of sight of the camera, by two rectangular, woven fiber-optic panels that provide evenly distributed light that is semicollimated and incident in two planes at acute angles (typically 50 degrees )from the surface. The camera and the sources of light are mounted on a robotic manipulator that scans them across the surface, stopping at inspection locations that are selected consistently with inspection requirements and with the dimensions of the field of view of the camera. The image data acquired by the camera and the position data generated by the robotic manipulator are digitized and sent to a desktop computer for processing.

The relatively smooth unflawed portion of the inspected surface reflects the illumination away from the line of sight of the camera. Because flaws reflect and diffract the illumination, they appear as bright spots in the video image. The brightness of a pixel is digitized on a gray scale of 0 (black) to 255 (white). In processing, the image data are first Fourier-transformed to filter out background information. Next, each pixel is binarized by setting its value equal to 0 if its gray level is <32 or 1 if its gray level is >32. The binarized pixel values, taken together with the pixel-coordinate information, and the position data from the manipulator, constitute the data for a primitive binary map of flaws on the surface.

This work was done by Mark Mueller of Thiokol Corp. Inquiries concerning rights for the commercial use of this invention should be addressed to the Patent Counsel, Marshall Space Flight Center; (205)544-0021.