There is an Accident pumping down that Pipeline

Pipes have risks whether they are simple or complex.

In nearly every pipe trouble can lurk.  Think of even the seemingly simple domestic water supply.  A problem resulting in an incident can mean the ruination of your home.  The issues with domestic gas pipes are understandably treated with more caution, even so unfortunate accidents are all too frequent.

Industrial piping is many times more complex as this picture shows.

 Refinery Piping

Piping design, although seen as less glamorous than design of pressure vessels or high-speed machinery, governs the layout of fluid processing facilities.  Many serious accidents that have entered the mindsets of process designers and engineers have had their root cause in inadequate piping design, installation, or operation. 

If one looks at those accidents, which have become household names to process professionals, inadequate piping was often the starting point of the catastrophe.  Ultimately the result was often the demise of the company involved.

Major process accidents have many causes.

Early in my own site experience I was the piping member of the team trying to get three poorly designed Ammonia plants up to a satisfactory standard of reliability.  The extraordinary piping feature of these plants was that one major plant section was designed to the highest quality using for example forged tees.  The adjacent section was designed to the same API standard, but to the lowest interpretation of that standard using tees made of set on branches and saddle plates where pressure dictated.  Of course, this involved welders of variable standards of skills and varying operational results.  Often, usually in the middle of the night, steam issued from leaking welds and my call-out resulted in the inevitable decision to shut down as in the cold night air, because of the condensing vapour, on most occasions, you could not even see the area of the leak.  By far the most serious incident was a mal-operation where the distillation column lost its normal fluid level and the back pressure swept through to the parts of the plant not designed for that pressure resulting in a large bore pipe fracture and a major leak of ammonia, which fortunately occurring during heavy rain and a favourable wind, quickly dispersed away from a populated area.  Had those weather conditions not been the case, then hospitalisation or even fatalities would have been the consequence.  The resulting design of 20-inch throat, 5-inch-high lift relief valves for 5 UK plants of similar design was my interesting project.  It involved using Bellville washers as helical springs could not achieve the specification.  The design intent was to limit the amount of emission of any such incident if it were repeated.  Similar valves were fitted to 5 ICI Ammonia plants, but not one lifted through mal-operation.

The accident at Flixborough in June 1974 killed 28 and was caused by the inadequate design of a temporary replacement of a large diameter pipe.  There were no professionally qualified engineers on the site and a circular section pipe was replaced on a temporary basis by a lobster back fabrication.  Regrettably, the repair proved to be more temporary than was intended!  The explosion wiped out the control room staff and thus no actions of mitigation were taken on the site.

At Seveso in July 1976 even with poor operational control the piping systems worked OK until finally the relief valve popped releasing toxic material with disastrous consequences for the health of the inhabitants of that part of Lombardy.

The Three Mile Island accident of March 1979 had its source in a failure in the cooling system and this caused a heat and pressure build up resulting in a relief valve blowing and a release of nuclear material.  No deaths resulted and this is a testimony to the brave team of operators in the control room.

Bhopal in December 1984 killed between 3,000 & 11,000 people.  The reasons were many, but a cost cutting management and poor maintenance of the piping systems certainly made the impact greater.  Whether the root cause was operational error or deliberate sabotage has not been determined with any certainty.

Piper Alpha in the North Sea in July 1988 resulted in 167 deaths.  Maintenance work on pipes, removed a safety valve, which was regrettably not replaced prior for start-up.  There were many other issues.  In fact, 106 recommendations were made in the official report and all were accepted.

Poor maintenance is not the sole province of developing countries, Texas City, USA, in 2005 displayed the same basic problem of a lack of a safety culture and no financial provision for good maintenance.

Buncefield here in the UK also in 2005, was the same - neglect of simple maintenance needs and the UK was extremely lucky that at 6am on a Sunday, few people were around and only one unlucky man was blown of his bicycle and no deaths resulted.

The Buncefield Fire after just 10 minutes

In preparing this piece as I looked in detail at Seveso, Bhopal, and Texas City.  Each of these were operational errors and despite poor maintenance the piping systems held up well before succumbing finally to the excess pressure and heat.

In the main, in all the incidents the piping standards were not at fault, although the actual design may not have been of the best and indeed safety equipment performed beyond the specification.  I am not aware of the story behind the next graphic, but the force of the explosion whether because of over pressure or inadequate wall thickness or both can be imagined.

 

So Standards are sound when correctly interpreted and applied.

In-line fittings

In many of these instances in-line fittings, principally relief valves, were the common factor.  Now these are designed to open, but this operation can cause the release of noxious fluids.  Further problems arise if the valve does not reseat when the excess pressure is released.  Standards have been tightened up since 1999, but it is surprising to note that the mean time between inspections is still around 3 years.  After three years perched in the open, high up in a structure, exposed to all the elements, makes one wonder if any equipment can be expected to perform well with that treatment.  Much more can be done to ensure that the release is contained and is made away from the operators and the installation of 2 valves allowing one to be removed for more frequent maintenance and testing would be a major improvement and is already adopted in many cases.

What else can go wrong?

Leaks at gaskets are the most obvious.  This is usually caused by inadequate gaskets design or installation.  Gasket suppliers are well able to solve this for you. 

 

More hidden and therefore potentially even more dangerous is corrosion caused by trapped moisture under insulation.  Not only does the water corrode, but salts are formed potentially causing a highly destructive mix.

CAD has been part of the solution.

Much of piping design was governed by heuristic rules and was amongst the first process discipline to be computerised, a technology which contributes to safety. 

The progressive roll out of CAD over the last 50 years has provided a control to the safety of piping.  The first programs were geared toward accurate take-off of material, indeed in the late ‘60s some items such as stainless steel were often still in short supply in the UK.  Later 3D modelling ensured that no pipe occupied the same space as another pipe, or a structural steel member or equipment item.  Such clashes are expensive and inherently dangerous to correct on site. 

Both of these applications were based on an approved piping specification.  Only these pre-approved materials could be called up by the design.  In Germany, the DIN standard goes further than API and regards the pipe isometric as a controlled pressure vessel drawing.

The strides in pipe stressing computer software have been immense and thus good design meets the requirements of Standards and avoids failures in service, where operated within the specified plant pressure and temperature ranges and suitably maintained.

Training is available for Piping Designers.

A leading textbook on piping provides only 100 words in 440 pages to the safety issue and there the text emphasis the needs of good layout to facilitate the escape of the operators.  At Flixborough it was the control room and its staff who were wiped out first, clearly an impediment to any recovery or mitigation.  In fairness the whole book is dedicated to good and hence safe design.  But as this review of accidents shows, a serious incident often goes far beyond the plant boundary.  Texas City and Flixbourgh caused death only on the site of the facility.  Buncefield was a case where society was incredibly lucky, occurring as it did early on a Sunday morning and thus blowing only one man (a shift worker) off his bike..  Seveso, Chernobyl, Bhopal and Deepwater Horizon had impacts well outside the plant boundary.

Unlike vessel and machinery design, we question if the piping discipline gets its share of training investment.  And also, whether there is enough academic rigour behind it.  An established training environment has been development for five levels of competence with the result of enhanced safety. 

Five levels of training have been identified by SPED, the Houston based Society of Piping Engineers and Designers.

Prerequisite       The basic educational standards to enter the piping discipline, lie somewhere around the HNC in the UK.

Level I                 Knowledge of equipment and fittings necessary for the routing of pipe from the equipment nozzle to the pipe rack.

Level II                The candidate has performed these layout tasks on a project, but with close review and assistance from a supervisor or senior piper.

Level III               Candidate has performed these tasks as a regular part of his or her job and only in unique or unusual situations did the candidate require assistance or review by a supervisor or a senior employee.

                            In addition, layout within the process plant, equipment placement, spacing and nozzle orientation, pipe routing to key equipment nozzles considering operations, maintenance and safety issues.

Level IV              Other employees regularly consult candidate for his or her expertise and assistance in performing their tasks and the candidate has managed, trained, or supervised others so that they can perform good quality piping design.

Training is being developed for Operators.

The IMechE, IChemE & Cogent are cooperating in developing what is called a Gold Standard for operators of all levels.  The vision is to increase industry’s profitability and competitiveness through world-class skills.

The Gold Standard is a UK national framework for continuous professional development setting out the skills required for world class performance in key job roles in the process industries.

Summary

The objectives of both CAD and the InIPED & SPED training programmes are to produce a good first routing, laying the foundation for safe piping.  The fundamental is good practice at design time, in order that the engineering design is straightforward.