Railway Track & Structures - February 2009 - (Page 23)

Detect rail defects Innovative technologies and advanced equipment are ﬁnding the smallest of cracks before they become big problems. S Dapco by Mischa Wanek-Libman, managing editor uppliers are focusing their railﬂaw detection efforts on covering more miles and getting real time results in order to ﬁnd ever-smaller defects before they morph into a large crisis. Dapco Industries, Inc., has 40 years of experience and commitment to the design and development of ultrasonic rail-ﬂaw detection systems and vehicles. According to Dapco, the company has invested tens of millions of dollars in research and development in all facets of rail inspection to ensure their systems are the most advanced and reliable systems available. In summer of 2007, Dapco introduced a rail-ﬂaw inspection technology that detected gauge corner detail fractures, a defect thought to be responsible for 70-90 percent of service failures. The company has equipped its entire ﬂeet of rail-ﬂaw detection vehicles with this new technology and, according to Dapco, in certain cases, on speciﬁc routes, the number of defects detected has increased by 30-35 percent. Dapco’s newest R&D efforts have been centered on detecting defects associated speciﬁcally with concrete ties. The company has partnered with several of its Class 1 railroad partners to cooperatively develop a new technology to detect these rail defects that are associated speciﬁcally with concrete ties. Not all R&D at Dapco is focused on new products, resolving existing challenges is also a key focus. www.rtands.com “Without exception, the one challenge that each of our clients faces is trying to increase the number of miles tested, within increasingly smaller track windows,” said Bob Coakley, marketing and sales manager at Dapco. “Railroads would like to increase the speed of test or at least the average miles tested per hour of track time.” Coakley points out that there are a number of variables that can affect the average miles tested per hour, but the one Dapco feels can have the largest impact on that metric is the ratio of false positives. “As you reduce the number of false positives, you reduce the number of times the rail-ﬂaw detection vehicle needs to stop and hand verify defects and thereby increase the average speed of the vehicle,” said Coakley. Dapco’s RFD vehicles are equipped with the Dapco RTS-400 software system. The system has artiﬁcial intelligence that allows the software to automatically recognize common rail conditions and to Herzog’s new F750-based test vehicle. Below, a trio of defects found by Herzog’s detection services. recognize and classify defects. According to Coakley, the power of the Dapco system is that it is an Adaptive Learning System and, as new defects are detected and analyzed, the “defect signatures” can be added to the Pattern Recognition Library, allowing the system to recognize these new defects automatically. “In 2008, we launched a new initiative aimed speciﬁcally at reducing false positives by evaluating and improving the artiﬁcial intelligence of our system. We have already seen a reduction in our false positive rates and we are projecting that this program will help reduce our false positives by more than 25 percent,” said Coakley. “Traditional rail-ﬂaw detection has focused on interior rail ﬂaws, like interior Railway Track & Structures February 2009 23 ENSCO http://www.rtands.com

Table of Contents Feed for the Digital Edition of Railway Track & Structures - February 2009

Railway Track and Structures - February 2009 Contents On Track Industry Today Supplier News AREMA News NRC News TTCI R&D Railroads Stepping Up Use of Technology to Locate Rail Flaws Rail Lubrication Realizing Great Potential M/W Challenges: Track Settlement at Bridge Approaches Supplier Profiles Products and Literature People Calendar Advertisers Index Sales Representatives Website Directory Professional Directory Classified Advertising Chicago Perspective

Railway Track & Structures - February 2009

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.