Lessons learnt from process safety accidents - an article from insurancenewsnet.com

A good article is found on the above title, narrating lessons from five less known accidents. The lessons are,

1. containment of water from fire fighting operations to avoid contamination of water sources in the rundown 
2. to consider all auxiliary equipment also as important as main plant equipment for understanding hazards
3. not to ignore hazards and need for monitoring even during shutdown 
4. positive isolation, purging and checking for hazardous atmospheres (using flame is not a correct check for presence of flammable atmosphere)
5. follow written approval system to override interlocks
6. follow safety permit procedures
7. identify the correct equipment before taking up work
8. communicate with clear identification while handing over the equipment for maintenance
9. ensure physical check by another agency before permitting equipment operation/maintenance
10. hazards from dust, propagation of dust explosions
11. preparation of emergency response management, with detailed instructions for every type of emergency
12. good house keeping
13. establish process safety management procedures
14. inculcate good safety culture, starting from the top management

Some statements on safety

Following are some of the statements in "Process Safety Analysis - An Introduction"  by Bob Skelton published by Institution of Chemical Engineers, UK.

1. A good safety culture ensures that both the spirit and the letter of the law are fulfilled.
2. Attitude to safety should be highly visible and shared at all levels within the company.
3. A well managed company is almost invariably not only a profitable company but a safe company.
4. Changes in existing plant are costlier than that introduced in design stage.
5. Design should be such that operator intervention is not needed for at least 30 minutes after an incident. Experience has shown that operators can not always be relied upon to make the correct decisions under immediate post-accident conditions.
6. Safety in design must be both proactive and reactive. Changes, once a plant is built, are very expensive compared with changes at the design stage.  It is not sufficient and cost efficient to make safety review after completing the design and then BOLT ON safety devices. It will not be cost effective. Engineered safety is BOLT ON safety. Engineered protective devices can fail and never place too much reliance on BOLT ON safety.
7. Commissioning is one of the most hazardous parts of any process plant operation. Not only do design errors which escaped previous checks manifest themselves but problems due to construction errors also become obvious. In addition commissioning generates hazards of its own as the plant moves from construction to operating status. It is essential that a formal set of checks be carried out before process fluids are introduced for the the first time.
8. Fire and explosions can be prevented  by not exceeding 25% of LEL. Flammable atmospheres can be avoided by ensuring that fuel lines and tanks are pressurized so the flammable material leaks out rather than air leaking in. good ventilation of vessels and plant areas can maintain safe working conditions.
9. Dust explosions are best prevented by good housekeeping - that is, by keeping the concentrations of dust down and perhaps keeping the dust damp. Inerting by dilution with non-combustible dust is another effective technique, frequently used in coal mines.
10. The risk is serious in case of static electricity, if the relative humidity is below about 60%.
11. Explosives manufacturing facilities are usually designed so that the buildings are separated by safe distance, surrounded by earth mounds so that any explosion will go upwards rather than affect other plants in the area. In addition there is usually a limit on the number of people allowed in a building.
12. Fire fighting water causes more damage than the fire itself, when polluted water is let into rivers. There may be a conflict between accepting the atmospheric pollution caused by letting the fire burn out and the water-borne pollution caused by fighting it.
13. Non-process hazards account more than 70% of all accidents in process plant.
14. Many of the worst accidents in the process industries are the result of bad maintenance practice. Ex: Piper Alpha and Flixborough
15. As many people die by asphyxiation as from toxic gases.
16. A good health and safety policy is always cost effective; most organisations grossly under estimate the cost of accidents, often by an order of magnitude. The organisation should be such that the attitude to safety is highly visible and shared at all levels within the company. Active participation is encouraged to promote the objectives of not just preventing accidents and industrial illness but motivating and empowering everyone to work safely.
17. A safety culture, once established, must be maintained, any any tendency to careless practices stamped out at once. Experience shows that 80% of accidents tend to happen to 20% of the workforce - the young and the old being particularly vulnerable.  Many accidents are caused by operators not fully appreciating the significance of small, but nevertheless important changes.
18. A good system of accident reporting is proactive and reactive, whereas most tend to be purely reactive.
19. Effective safety at all stages of a project - from inception to demolition - can only be achieved if there is a commitment at all levels. The senior management must see health and safety as being just as important as profitability and they must make certain that all their workers are aware of this fact.
20. A well managed company is almost invariably not only a profitable company but a safe company.
21. In hazard analysis, a distinction must be made between routine operator action and operator intervention in an emergency.  For routine operator action, the operator can usually take time and is under no great stress. Safety assessments involve the prediction of the likelihood of errors when the operator is taking corrective action against alarms. The time for corrective action may short, the operator is liable to be under some stress and so the probability of errors is greater.
22. Total elimination of human error will never be possible. Use must be made of the science of ergonomics to ensure that everything possible is done to enhance the strengths of human operators whilst at the same time allowing for the weaknesses.
23. The most important rule is, 'inherent safety is better than engineered safety', ' what you have not got can not leak'. Even elaborate safety devices can't reduce risk to zero due to the escape of a noxious substance, but replacing a noxious substance by a more benign one could well eliminate that risk altogether.

 

Process safety and stuxnet worm?

Of late a number of articles are appearing on impact of stuxnet worm on process safety and it is an eye opener for me. Though, I am bothered about safety of my computer at home and office, it never occurred to me that this can lead to accidents in plants which are operated from computers. Possibly, this is due to the fact that only limited sections are operated from computers in the facilities which I saw all along.
Now that I am aware of this stuxnet worm and its cousins that can affect process safety, I though what we can do to prevent such process safety hazards and following are some that came to my mind.
1. As the worm has to enter through internet / removable disks, these computers should not be connected to internet.
2. There should not be any provision to use pen drives, floppy drives, CDs.
3. Realtime protection from viruses, worms, etc should be available. For this realtime protection program loading, the server system should be under lock and key control of a senior officer. If it is affected, then this man only is the channel for viruses/worms to enter the system.
4. A realtime standby system operating on a different mode should be available (diversity in redundancy).
5. Backup should be taken in every shift.
6. The persons manning the control room should change their dress totally in the change room and enter with company provided dress before entering control room / server room.
7. Control room, server room, etc should be under surveillance.
8. As it is seen in the movie Die Hard (2?) of taking over all control systems by connecting from distance, the cables connecting servers with plant data monitoring and control cables should be secured and any attempt to interfere should be sensed and alarmed.
9. Like railway gangmen check the rail tracks regularly, cable system should be checked physically also by authorized staff.

Above are some thoughts which came to me for ensuring process safety from worms like stuxnet. I have to understand about how plants are operated from computers and what are their defence systems.

Process Safety Management - for whom?

Everybody says process safety management (PSM) without knowing or having commitment. Many may be thinking it as day-to-day plant operation. In this regard, to identify, evaluate and control hazards in chemical plants, OSHA has given guidelines and can be read here.
It must be understood that by doing PSM, we are doing favour to ourselves and not to others. Even, production cost can be brought down, if we do PSM sincerely and follow it. If we do not care to spillages/leakages of chemicals, air, steam; do not care to study in detail before going for a new process / chemical / modification; do not make checklists for operations, then it means that enough attention is not paid for plant operations and it adds to cost of production by way of wastage, inefficient operation, accidents, and so on.
We should not feel like Alice in the wonder land when the unidentified hazards manifest into incidents/ accidents. One may feel stressed to follow the safety principles all the while. But, the stress will be more and can also lead to penalties and punishments when there is loss of life/damage to environment/property.
Many times when we land into simple accidents outside factory, we vow not to repeat such mistakes. But, when we commit such mistakes in a factory, the consequences will be very high and at that time we can move the time back to correct our mistakes. There is no time machine yet available to correct ourselves otherthan following safety principles in toto.