Slashdot Mirror
Chevron Got North Sea Contract Despite IT Safety Crashes
DMandPenfold writes "The UK government gave Chevron the go-ahead in September to drill in the North Sea off the coast of Scotland, in spite of the US oil giant's admission that its contractor's spill prediction software constantly crashed and was not a reliable predictor of how far oil could travel if an accident took place. The news comes in a week that US investigations into BP's disastrous Deepwater Horizon oil spill hit the buffers, after an IT contractor firm refused to hand over access to its software."
Oil and government... do mix.
Oil deposits in the North sea have been propping up the British economy for decades now. While the UK is facing a big deficit, and getting stung by other economies in the Eurozone looking wobbly, they'd probably overlook a few things to get their hands on the money. -So much for a 'greener' coalition!
When you understand the unique chemicals and items we can synthesize out of this ancient black goop, you realize how stupid we are for simply burning it.
That's a lot of potential plastic...
People have been drilling in deep water in the North Sea for decades, with admittedly a couple of nasty accidents, but so far things have gone pretty well. And do you know what? No-one had oil spill prediction software when they started. They relied on the skill and experience of the people operating the rigs.
Bear in mind that this is the UK, where we have far, far tighter safety regulations than the US for the oil industry. We know what we're doing. Oil companies in the US clearly don't, or don't care to do it properly.
http://en.wikipedia.org/wiki/Piper_Alpha
"An explosion and resulting fire destroyed it on July 6, 1988, killing 167 men, with only 59 survivors. The death toll includes two crewmen of a rescue vessel. Total insured loss was about £1.7 billion (US$ 3.4 billion). At the time of the disaster the platform accounted for approximately ten percent of North Sea oil and gas production, and was the worst offshore oil disaster in terms of lives lost and industry impact."
"People were still getting off the platform several hours after the initial fires and explosions. The main problem was that most of the personnel who had the authority to order evacuation had been killed when the first explosion destroyed the control room. This was a consequence of the platform design, including the absence of blast walls. Another contributing factor was that the nearby connected platforms Tartan and Claymore continued to pump gas and oil to Piper Alpha until its pipeline ruptured in the heat in the second explosion. Their operations crews did not believe they had authority to shut off production, even though they could see that Piper Alpha was burning."
Why is Snark Required?
Shit is blowing up and burning around you...and you do nothing because you don't believe you have the authority to do anything (at least, anything significant)?
One hell of a corporate indoctrination seminar that had to be. I'm sorry, but if it were me, shit were blowing up, flaming out, people getting killed, it really wouldn't matter to me who had told me I couldn't do X, Y, and/or Z, if X, Y, and/or Z could at least prevent things from going from worse to OMFG (and provided I had the ability to do those things--else I'd be all 'see ya later bye'). When it's life and limb, only the guy with a uniform, badge, and gun is in charge.
I don't post AC. I like my -1, Flamebaits. Trump/Sheen 2012 on the Batshit Insane ticket!
You do know that Guy Fawkes was fighting to replace one monarchy with another? That his intention was not to get rid of Parliament but rather to kill Protestant Christians and replace them with Catholic Christians and maintain the political status quo (a group of rich, landed members of Parliament serving a monarch), just change England from Protestant to Catholic rule?
I think a lot of people cheer Guy Fawkes because they think he was standing up for democracy and overthrowing tyrannical states whereas I am pretty sure the situation was that he just wanted to overthrow one religious authority with another. He wanted the same political system (monarch served by ineffectual unrepresentative MPs) just a different religious flavour.
His goal was to replace King James the first of England (and sixth of Scotland) with his nine year old daughter princess Elizabeth.
I'm not a modelling guru but I've delved a little into current and weather modelling and it's not simple stuff. The inputs would vary depending on the weather conditions and weather predictions change significantly over that time. The best you could hope for would by a dynamic model that updated daily and became less reliable the further away from now that you went. Having a model that predicted where the spill was headed after the spill occured so you could direct clean up to the most effective places on a daily or hourly basis would make a lot more sense to me than a model that took a theoretical set of conditions and set in stone the response. I'd suspect that the complexity of this sort of thing probably means the guys who write the software sell the service.
As one prof of mine said:
When you understand the unique chemicals and items we can synthesize out of this ancient black goop, you realize how stupid we are for simply burning it.
That's a lot of potential plastic...
Why do you mention it as if it were a good thing?
If your software crashes when run for long periods, the root cause is almost always one of the following:
None of these should be present in the "standard industry systems" of multi-billion (trillion?) dollar industries, especially if they pertain the safety systems. Memory leaks and boundary overflow/underflow are trivial to avoid by any programmer who takes the time to code defensively. Race conditions can be a bit harder to detect and avoid, but they are a less common issue, and handling them is well within the expectations of even a newly-graduated programmer.
There are also more esoteric error conditions. For example, in locations which are higher above sea-level, the "average neutron flux density" (i.e. number of cosmic rays hitting your chips) is higher, and thus the incidence of random transient faults in electronics is higher. It is not unheard of in large computer clusters to have the occasional bit-flip error in RAM due to a random cosmic ray. At the same time, these systems have built-in checkpointing, and when for whatever reason the running software develops a fault, the entire system can roll back to the last known good checkpoint, and restart.
Serious customers would never accept a system or simulator which exhibited the sorts of problems these "industry standard systems" seem to be plagued with. More to the point, I have worked indirectly with some petroleum companies on simulation software, and know for a fact the acceptable "unrecoverable error rate" was written into the contract in a very forceful way. Then again, the simulation software was being used to locate oil, so I guess that says something about the industries priorities. (And it was designed to run on a "real" cluster)
There were emergency evacuation plans in place. These were put in place because people had sat down in a calm setting and worked out the best ways to get everyone safely off the platform. There are temporary refuges on (most) platforms. These are designed to keep you safe from the fires/explosions for at least an hour. When it looks like the fires are serious enough that you have to get right off the platform there are more plans that will get everyone off safely - everyone jumping off the side usually leads to most people breaking their necks when they hit the water, and the rest dying from the cold.
Jumping into a tiny escape boat is not something people want to do in most situations, so they were waiting in the refuge to see what was going to happen. 'Authority' is probably not the best word to use here, it refers to the people with the experience to know what's going to happen, and full knowledge of the situation at hand.
The only people who survived Piper Alpha were the ones who ignored the plan. A lot of these people were seriously injured when they hit the water, and a lot of them did not survive. This makes it fairly clear that the plan wasn't good enough.
The lessons we've learned from that event have changed the way we write our evacuation plans. There are still people in charge with the authority to direct the evacuation, but everyone now knows the procedures. If no-one has told you to get off the platform but you've been sitting in the refuge for an hour you know should know how to get off the platform safely.
Yes, I am a safety engineer.
Is 1563649 a prime number?