Fitness for Service: An Assessment of Safety—and HealthWritten by Gary T. Fry, Vice President, Fry Technical Services, Inc.
TIMEOUT FOR TECH, RAILWAY AGE SEPTEMBER 2022 ISSUE: Fitness-for-service assessments of an engineered system are designed to address a basic question: Is the system acceptably safe to use for its intended purpose at a given point in time?
Welcome to “Timeout for Tech.” Each month, we examine a technology topic that professionals in the railway industry have asked to learn more about. This month, we are focusing on fitness-for-service assessments of engineered systems.
Many fitness-for-service procedures are somewhat analogous to a person having a routine physical examination from a doctor. In fact, much of the underlying mathematics is essentially identical when assessing an engineered system for safe use or a person’s risk factors for serious disease. Figure 1 is offered in response to the idea that assessing the safe use of an engineered system has analogies to assessing the health of a person. Let’s get started.
In a routine physical exam, various tests are performed, and the results are compared against predefined ranges of values considered “healthy.” The tests that are performed during the exam, and the value ranges for each test, are usually selected statistically to limit the probability that an important risk factor for serious disease is not detected. Note, however, that this probability can never be identically zero—in theory or in practice. The predefined “healthy” value ranges for each test, and the target risk-factor detection probabilities, are the results of comprehensive studies, expert discussion and expert decision-making. The results could be all “healthy,” and I wish that for all of my readers! Some results could come back indicating a concern that needs attention. Some results could come back indicating a serious health issue that needs immediate intervention. Now what about engineered system safety?
Fitness-for-service assessments of an engineered system are designed to address a basic question: Is the system acceptably safe to use for its intended purpose at a given point in time? Continuing with our analogy, we are asking, “Is the system healthy?” To answer these questions, we examine the system and the condition of its components and compare our observations and measurements against predefined tolerances. The measurements that are performed, and the predefined tolerances for each measurement, are usually selected statistically to limit the probability that the system’s components will fail before the next assessment cycle begins. Note, however, that this probability can never be identically zero—in theory or in practice. The predefined measurement tolerances and target failure probabilities are the results of comprehensive studies, expert discussion and expert decision-making. At the end of the process, we decide as to the system’s safety for its intended purpose, and clear the system for normal operation, flag it for reduced operation and further evaluation, or take it offline for repair or replacement.
With that introduction, and continuing with our analogy, let’s look at the process of examining the performance of an engineered system. There exists a field of engineering consulting, technology development, and research called Structural Health Monitoring or SHM. The common application of SHM involves deploying a strategic inventory of electronic sensors on an engineered system along with data processing algorithms, which often incorporate artificial intelligence (AI) approaches. Most often, SHM is optimized to measure the dynamic performance of a system in use, and alert when unusual or extreme events occur that could lead to system failure. Some experts refer to SHM as giving an engineered system an ability to register “pain.” A student of mine once described SHM as “teaching a silent child to cry” so that his parents can better respond to his needs. Anyone who has purchased a car in recent years has observed something similar to these concepts.
For the past 20 years, our automobiles have come complete with dozens of electronic sensors, onboard computers, software systems and data visualization systems. For example, as seen in Figure 2, indicator lights sometimes appear on my dashboard that are simple and intuitive. With some of these indicators, such as “Check Engine” or the ominous red exclamation point warning that my brake system is somehow under the weather, I am going to set up an appointment with a repair shop.
When I leave my car at the shop, one of the first things a mechanic will do is connect a computer to the car’s On-Board Diagnostic or OBD port. Though coming from the exact same data sources that resulted in my simple indicator light being activated, data presented to the mechanic will be far more detailed and designed specifically to help the mechanic sort out the issues my car is having. Essentially, the car is participating “actively” in its own fitness-for-service assessment—like a patient answering questions from a doctor.
There are quite a few applications of “intelligent sensing” systems in the railway industry that provide data streams that aid in fitness-for-service assessments. And more are under development every year.
For example, on the track side we have:
- Track geometry measurement systems.
- Rail profile measurement systems.
- Rail defect detection systems.
- Ultrasonic systems (for subsurface cracks).
- Electromagnetic field imaging systems (for surface cracks and spalls).
- Tie inspection systems.
- Ballast inspection systems.
- Optical systems.
- Light detection and ranging (LIDAR).
- Ground penetrating radar (GPR).
There are also several wayside systems available to monitor passing rolling stock:
- Truck performance detectors.
- Acoustic bearing detectors.
- Thermal bearing detectors.
- Thermal wheel detectors.
- Wheel impact load detectors.
- Wheel profile measurement systems.
- Automated cracked/broken wheel detectors.
- Ultrasonic systems (for subsurface fatigue cracks).
- Electromagnetic field imaging systems (for surface cracks and spalls).
- Optical systems (for partially failed rims).
Certainly, cutting-edge technology helps us make more accurate and precise assessments of real-time system condition—notably when it is in use and under load. But there will always be fundamental questions related to establishing the threshold limits of allowable system performance and degradation. And this brings us to the most difficult and uncomfortable part of every fitness-for-service discussion: determining a target for acceptable probability of failure. It is most sensitive in scenarios where human life is at risk. So, I pose the question: What is an acceptable probability of failure when lives are at stake? Many of us are prone to say, “Zero! Failure is not an option!”
Failure of essentially zero engineered systems is not possible, certainly not today. Our imperfect scientific and engineering knowledge, our imperfect manufacturing abilities, our nearly complete inability to control Mother Nature, our imperfect decision-making paradigms, the reality of human error, and other inherent imperfections combine in complex ways and prevent the total elimination of failure risk.
To mitigate the most serious consequences of failure, we have nothing available above our best possible efforts. We must commit them to all tasks at all times, model carefully what that means in terms of failure risk, and report the result in good faith as our acceptable probability of failure. We can’t do better than that, and no one deserves less when life-safety is at stake.
Dr. Fry is the Vice President of Fry Technical Services, Inc. (https://www.frytechservices.com/). He has 30 years of experience in research and consulting on the fatigue and fracture behavior of structural metals and weldments. His research results have been incorporated into international codes of practice used in the design of structural components and systems, including structural welds, railway and highway bridges, and high-rise commercial buildings in seismic risk zones. He has extensive experience performing in situ testing of railway bridges under live loading of trains, including high-speed passenger trains and heavy-axle-load freight trains. His research, publications and consulting have advanced the state of the art in structural health monitoring and structural impairment detection.