Railroad companies have special cars that grind the rails to specifications, and sparks do fly out of the grinders. A water car at the end of this type of train sprays any small fires that get started in brush. Here's a short video: http://www.youtube.com/watch?v=6pAfMlr4Pko.
Buring a trip through Evanston, Wyoming my wife saw one of these grinding trains and thought it was on fire. So sometimes sparks are intentional.
Back about 1987 I designed a system using spectrum analyzers designed for exacly that purpose, industrial fault detection. I coulkd set the frequency range to examine and then have ten "boxes" around the PSD (Power Spectral Density) trace, and if the amplitude went outside the "box" a TTL signal would pull low and my equipment would know that the part being tested was outside of the specification. I believe that package was produced by Spectral Dynamics, but I don't recall the model number. One of the last testers used it to check in-tank fuel pumps, and I gather that it worked quite well. What I don't know is if they are still using that device to chek pumps. IT was an easily upgradeable test stand, so it could be still checking those pumps now, 32 years later. We did build equipment to last.
It would seem that a similar system could listen for bad bearings, but the train speed would need to be quite closely controlled.
Another interesting railroad technology, is the application of Electronically Controlled Pneumatic (ECP) brakes.
Railroads have been using the same brake technology since the late 1800's! Here's an over simplification: Basically, there's a continuous air line that runs the lenght of the train. The locomotive generates air that charges a small resevoir under each car. When sufficient air is present in the system, the brakes on the train release. To activate the brakes, the engineer releases a small amount of air from the system, which propogates back through the train, applying the brakes an amount related to the amount of air released. With long trains, it takes a long time for this air control to pass through the train, resulting in odd train handling where brakes are applying and releasing at different times in different parts of the train. Worst of all, is that if the engineer makes too many brake applications without recharging the system, the train can lose all braking ability!
With ECP, air is still used to operate the brakes, but an electrical system is used to tell the brakes when to apply and release. With this system, the locomotive can keep charging the air line so that the train won't lose braking power. They can also apply and release the brakes with more precision, and all cars react at the same time, decreasing stopping distances. And, there are individual car brake diagnostics available.
Testing has shown it to work well, but adoption is slow due to the cost and amount of rail cars out there. It's primarily being used to "unit" trains in captive service, but I hope that the technology becomes more widespread eventually.
I remember as a kid, seeing a train go by me spewing sparks and glowing debris from the wheel. I thought it was cool at the time. (I was standing right next to the tracks.) Now, that concept frightens me. It is 2012, and 100 year old ways of "doing things" does have to step aside. Innovations like this are so important.
Hi, Cabe. I bet the sensor system looks for a specific pattern of sound frequencies within a certain bandwidth. Various filtering techniques or a fast Fourier transform would let software "look" for activity at those frequencies.
There is a long history of a variety of bearing failure detection methods in the railroad industry. In the 1980's, I was working for a railroad supplier and was heavily involved in bearing failure detection methods. We developed and manufactured a line of trackside heat detection systems (called 'hot box detectors') which, with a variety of sophisticated computer algorythms, were quite successful at detecting bad bearings on trains as they rolled by the hot box detector. There were two downsides to the hot box detector system - you needed a detector every couple of miles and they only detected a bearing that had already gone bad - when they go, they go quick.
We also dug into the audio detection scheme. Our company spent a considerable amount of time trying to come up with a reliable way of acoustically detecting bad bearings but was unsuccessful so congratulations to RailBAM for their success. Our project was started when our client(s)(i.e. the railroad companies) noted that their track crews could detect, by just listening as trains rolled down the hump yard tracks, a failing bearing. Then they could shunt the bad car and have the bearing fixed before it got out on the road and failed (with usually nasty consequences).
So they suggested we try to come up with a product to automatically achieve the same results. We spent many, many hours at track side recording the trains as they rolled by (digitally) and attempted to correlate what the computer picked up with reports from the crews on 'bad' wheel sets. Unfortunately we were never able to come up with a solid and reliable way to do the detection.
The entire subject is quite fascinating - well, at least to me!
Being a geeky engineer and a railfan, I know a little about this stuff.
Here in the USA, railroads primarily employ three types of defect detection.
- Hot box/bearing detectors where thermal detectors look at wheel bearing temperatures.
- Dragging equipment detectors where sensors detect if anything on the equipment is dragging or hanging too low.
- Wheel Impact Load Detectors (WILD) where they measure the acoustics / impact of the passing equipment.
All of these detectors can be interfaced to a AEI reader which reads the RFID tags present on all rail equipment, and I think all of the equipment is equipped with a computerized radio transmit system which transmits the status of the train to the engineer/conductor after the train passes. (If you have a radio scanner, it's interesting to listed to the train speed and axle count as a train passes.)
I find that the WILD systems are the most interesting, as the data is used to predict when wheels will start to cause a problem. Most rail cars are owned privately, but there are standard repair agreements that let a host railroad perform maintenance on a car if it's needed. Replacing wheels are expensive and obviously car owners wish to avoid this, but out of round wheels (we've all heard a freight train with a banging wheel) cause major damage to the railroad infrastructure and can cause derailments.
In addition to the three primary system types mentioned above, there are a few (about sixteen nationwide I believe) Trackside Acoustic Detection Systems (TADS) installed. Similar to the system in the article, TADS uses acoustic signatures to indentify potential wheel bearing problems. Neat stuff!
I don't think many people are aware the railroads were probably the earliest adopter of RFID technology with their AEI system. Much better than the old ACI system which used colored bar codes, which failed miserably in the dirty railroad environment.
As to wheel repair as mentioned by a previous poster, the feds limit what type of work that can be done to wheels. I think that certain failures such as cracking, can condemn a wheelset.
The original bearing method for railcars was poured babbit, hand scraped and lubricated with oil-soaked rags stuffed into a metal box above the outboard end of the axle. The failure detector was the brakeman in the caboose looking for smoke as the babbit melted out and the steel-to-steel created molten metal that ignited the oil-soaked rags. Hot-boxes are still an issue even with modern roller bearings and yearly we fight fires along the rail tracks because a bearing failed and spewed molten metal along 10 miles of track igniting ties, brush and debris along the trackside. Thermal detection methods work to detect failed bearings but a method of finding incipient failures would be a significant improvement. A 120 car train with 2 two wheel trucks per car is nearly a thousand bearings. Swapping out a car with a defective bearing in the middle of a train is nearly impossible so early detection is important.
In the early seventies I started work with a previous employer and inherited the engineering responsibility for their line of cylindrical roller freight train bearings. Many bearing failures were spotted by hotbox indicators, and after quite a few of these, the failure mode was established. The thin walled, straight Inner Races would turn on the axle journals and wear against the fillet ring and end cap that doubled as thrust flanges. This reduced the tension on the 3 bolts that screwed into the axle ends, holding the whole bearing assembly together. The bolts unscrewed even though they were supposed to be "locked" by bent tabs of a large triangular washer for all 3 bolts. The axle cap would then fall off, and since the bearing's inner race was turning on the axle journal, it would also fall off. Next, the bearing's outer race and rollers only had to drop down 3/8" (thickness of the inner race) and run directly on the axle journal. Grease would be lost and the axle being relatively soft resulted in a bad bearing making heat that was picked up by the hot box indicators. Sometimes, when a maintanence crew would go to the car, they would find the whole end of the axle worn off. This fiasco resulted in the company abandoning the railroad bearing business.
Truchard will be presented the award at the 2014 Golden Mousetrap Awards ceremony during the co-located events Pacific Design & Manufacturing, MD&M West, WestPack, PLASTEC West, Electronics West, ATX West, and AeroCon.
In a bid to boost the viability of lithium-based electric car batteries, a team at Lawrence Berkeley National Laboratory has developed a chemistry that could possibly double an EV’s driving range while cutting its battery cost in half.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.