By Mayank Chhaya-
Professor Bharat Thakkar, a leading Indian American expert on reliability of systems and quality control, says as India grapples with the tragic Odisha train disaster, metal fatigue in the railway tracks deserves a close examination as a possible cause.
Thakkar, who has a Ph. D. in mechanical engineering and has taught at major universities in the U.S. with several papers to his name in the field of reliability and quality control, says that it is important to overhaul the Indian Railways’ maintenance regimen.
He offers specific calculations to establish that poor maintenance of the tracks, 98 percent of which according to a report in the Times of India were built between 1870 and 1930, to illustrate how metal fatigues in the tracks could be a likely cause.
Dr. Thakkar was the first expert to have said unambiguously within a day that India’s Vikram moon lander likely crashed at about 184 miles per hour (mph), a speed at which nothing of value could have survived. The lunar lander crashed on the Moon on September 7, 2019.
In an interview with Indica News Dr. Thakkar, says the Indian government needs to mount a major effort possibly under a single department to ensure high quality maintenance of all physical assets and infrastructure.
Indica News (IN): As a leading reliability and quality expert what is your immediate sense about the likely cause of the Odisha train crash?
Bharat Thakkar (BT): Railroad tracks should be designed for impact. Impact is similar to driving a nail in the wood versus putting hammer weight on the nail. Impact is lot of energy in a dynamic situation. Fatigue in metals and stress concentration in structure mask (hide) deteriorating tracks. Frequent checks are required. Only one bad location in hundreds of miles of tracks can cause an accident. Therefore, if the Times of India report suggests that 98% of tracks were built during the British Raj, then our tracks are close to 100 years or more old and not necessarily up to the code.
IN: You have argued that maintenance issues coupled with decades-long wear and tear of the rail tracks could be the primary reason for the crash?
BT: Our rail system’s worst case can be head-on collisions. Electronic malfunctioning is also similarly reviewed as how the wear and tear occurs there. Electronic systems must be recent. So, one cannot easily suspect that as the cause. That leaves me with a question mark over the tracks.
IN: Tell me about the set of facts you base your judgment on? I believe you have made some specific calculations. Can you share those calculations?
BT: I present the following assumptions as a case for wear and tear is most likely cause:
- The tracks are used once a day (which is very conservative)
- The distance between wooden planks is about 12 inches. (Reasonable)
- Fatigue experienced by tracks is 7 million cycles. (If the tracks are 100 years old) plus the fatigue experienced by variation in temperatures daily and seasonal.
- The hardest and strongest steel is used. (Conservative)
- Railroad tracks are considered to be beams on elastic foundation (ground strength build by pebbles under the tracks, the lack of better foundation or missing stone pebbles makes the ground weak.)
- The stress on the tracks should be below the endurance limit which is 30,000 psi (pound per square inch) for high-strength steel.
- Cracks take little energy to spread similar to tearing a piece of paper. Once crack is created (95% of energy) it takes only 5% energy to propagate.
- If cracks are the cause, then maintenance issue is valid.
- Combination of fatigue and stress concentration can together create accidents similar to the Morbi bridge disaster. In that report, it was advised that all structures be checked. Time degrades the strength of materials, which is a well-known fact in engineering and in scriptures.
IN: You have a stellar track record of having accurately worked out the math of India’s Vikram lunar landing failure and said clearly that the lander could have survived the crash. Have you taken a similar approach to the Odisha crash?
BT: Be it the Vikram or the Morbi bridge, dynamics of situations were present in both cases and are present in other such cases. We should not design on the basis static situations. All systems can fail some time in life, but they should be fail-safe, similar to a flat tire on highway with 65 mph versus a leak that takes time before all the air is out. This is called a graceful failure where no catastrophe happens. There should be a subsystem that has a function to detect errors and take corrective actions.
IN: Can you describe the fatigue factor that affects any products that operate under high stress as the tracks? After all, 98 % of India’s tracks were reportedly laid between 1870 and 1930. We do not know what kind upgradation/replacement and maintenance have happened over the decades?
BT: Fatigue is a nuisance and not well understood in metals. We humans face the same situation where our bodies do not function after so many years of life. We need to find a permanent solution. We need to design the traffic light such that green-green situation will never happen. It requires error-proof system design. Even if humans make errors, the system will hunt down errors and stop massacre from happening. Our systems have to be not only reliable but robust. It should be able to take punches of the worst kind.
IN: In a broader sense, would you make the case that the Indian government needs to set up a special department that continuously studies and fixes the problems of reliability of large systems such as railways?
BT: The government should look after all monuments including, bridges, railroads, transportation under one department. It would be a 1000-employee venture to look into the problem. Robustness can be achieved by having a redundant system in parallel with the main system. Redundancy reduces the probability of both systems down is close to zero percent. Redundancy is a common practice in designing important systems. This is similar to a truck having multiple wheels to avoid tire failure and imbalance caused by the single failure.
