Post-Accident Safety Analysis (otherwise known as Forensic Safety Analysis, Forensic System Safety Analysis, Forensic Safety Assessment, or Forensic System Safety Assessment) is a very sobering process when there has been loss of life.
As analysts, we
- have had to deal with gauntlets of reporters at the client's premises as well as other reporters
- have interfaced with regulators of the associated industry
- have attended meetings with and made presentations to corporate lawyers
- have seen companies crippled
- have seen management and engineers become unemployable in their professions (have you ever heard of software engineering malpractice?)
- have seen subpoenes issued to past employees
- have seen huge law suits
HCRQ provides around the clock, around the world response to accidents - professionally, efficiently and, of course, confidentially.
Prior to any information, verbal or otherwise, changing hands, we sign legal documents with our client beginning with an NDA. Often all aspects of the contract signed between us and them is confidential. We sometimes fly to our first meeting with a client knowing nothing but suspecting a lot.
Often software errors, which contribute to accidents, resemble transient hardware faults. They can be very difficult to diagnose as little if any evidence of the cause of software-related accidents is left behind. And, if you execute the software again using identical inputs, more than likely it will work correctly since internal timings are different. These scenarios are sometimes encountered during testing and, after multiple attempts to reproduce the same result fail, the embarrassing test result is often discarded.
A commonly seen situation is where our clients have a false sense of security regarding the safety of their complex systems which have operated accident free for years.
It is worthwhile noting that these systems, for which we our services have been requested, are certainly not the worse designs we have seen with respect to potential for catastrophic failure.
By the way, have your ever heard of the THERAC-25 cancer therapy machine? Guess who was called in? HCRQ when its name was DLSF Systems.
Have you heard of a nuclear reactor LOCA (Loss Of Coolant Accident)? The brakes, which prevent bridge motion while holding the fueling machine, were accidentally released while the fueling machine was clamped onto a fuel channel end-fitting. The fueling machine dropped a distance of 40 cm. badly damaging the end-fitting. The heavy water leak (initially 1400 kg/h) dropped to 18 kg/h when the heat transport system was de-pressurized. The incident was traced back to a software error which was introduced into one of the protective computer systems approximately 4 years previous. Due to our previous contract work, we were called in to perform a safety analysis of this software. This work resulted in a published paper "Fault Tree Analysis Of Software At Ontario Hydro" coauthored by HCRQ's director of consulting.
Thus far, HCRQ has performed 14 forensic analyses.
Our work in this area reminds us of the African Proverb: "Smooth seas do not make skillful sailors".