The Story of The Cabin Creek Explosion | Plainly Difficult

Working in confined spaces can be a terrifying thing. The closest I have to this is when servicing the old family chariot. Crawling around under a 15-year-old Honda isn't the most pleasant experience. But it's a holiday compared to today's subject. That is working in a tunnel that was not designed or intended for human occupation. It is of the drained penstock of a hydroelectric plant. Our story today is about a fire in a pure horrific situation in which the victim's final moments must have been the stuff of nightmares. It has it all. A tunnel deep underground, chemicals, a single point of exit, and of course a fire. Now, you'd think that such a disaster was maybe from 100 years ago. Well, no. It was from the mid-2000s. Today, we're looking at the tragic Cabin Creek Hydroelectric Plant disaster. My name is John, and welcome to Plainly Difficult. This video wouldn't have been possible if it wasn't for my YouTube, Kofi, and Patreon members. If you want to get early access to channel's videos, then you can from just £1 per month. And as always, the links for this will be in the pinned comment below. In addition to this, the link for the video's references as well. Background. So, our story will begin and ultimately end here at the Cabin Creek Pumped Hydroelectric Plant. It is here in the Rocky Mountains near Georgetown, Colorado. It's in a remote area, making use of the gravity of the mountains for the generation of electricity, estimated at a full power output of 324 MW. The site is at an elevation of greater than 10,000 ft above sea level, and distributes the electricity it generates via transmission lines. It opened in 1967, and worked off a pretty simple concept. That is water loves to flow downhill. The plant was made up of four main parts, an upper reservoir of water, a lower reservoir, a power plant that housed the turbines, and a penstock that linked all these parts together. The penstock is long at 4,163 ft or 1,269 m long. This was between the upper reservoir's intake, which looked like a little guardhouse known as the mushroom, to a section of the penstock that splits into two parts near the powerhouse to power the two turbines. 3,123 ft of this or 952 m of the tunnel is big enough to be traversed by foot, of course sans water. The final 1,040 ft of the penstock, however, was rather steep requiring the assistance of ropes and ladders. The lining material of the tunnel was made up of welded steel with other places using concrete. In order to stop the steel from rusting and distorting under the constant flow of water, it was coated with an epoxy. The plant was only really intended for peak demand use as eventually the upper reservoir would run out of water. This is where the pumped in the pump storage comes into play. When not required for energy generation, the penstock is used to pump water in the other direction from the lower reservoir back up to the upper one ready for another session of electricity generation. Now the lining would in the year 2000 be discovered to have deteriorated enough to allow parts of the penstock steelwork to begin rusting. Although when discovered, it wasn't enough rust to cause structural issues, but it still had to be addressed and repair works were needed. But it would take 7 years to get to begin as the plant's operator put in request to extend the deadline for the repair works, but eventually the work had to be done and Xcel Energy hired a contractor called RPI Coating Incorporated after a tendering process. We will get into this company a bit later on, but works to remove and reapply epoxy coating will begin in September 2007, which leads us on to our next section of the video. The disaster. So works for the penstock relining would begin in September. First by shutting down the power station and draining the penstock. In order to gain access to the tunnel, an access panel was cut using flame torches. This would be the sole entry and egress point for the project. This entry point was just 4 by 6 ft or 1.2 by 1.8 m. I'm a smidge over 6 ft and that seems too small for getting into a tunnel in my books. A wooden ladder was provided to help with workers getting in and out. Once the opening was cut, Xcel and RPI staff then began removing the standing water, dead fish, mud, and debris from the penstock. It was noted by some of the workers that the build-up of moss on the surfaces made it a very slippery task with one worker actually dislocating their shoulder after slipping over. To remove the epoxy that was beginning to fail, RPI was going to use sandblasting and to contain the sand from going further up and down the penstock, wooden bulkheads were constructed named the east and west. In addition to this, ventilation holes were cut in the steel lining. These were roughly 2 ft by 2 ft. In order to keep some semblance of air flow, two dehumidifiers were employed. These forced air into the working area and a dust extractor sucked air out to collect the sand media. This gave the workspace an estimated air replacement rate of 4.4 per hour. The sandblasting and prep work would take the best part of September to complete. And hand sanding and grinding was finished on the morning of the 2nd of October 2007. The next stage of the project would now be able to continue. This was the application of the two-part epoxy. A special spraying machine would be used and had two compartments for the two parts of the epoxy, i.e. the base and hardener. For the respective parts, they needed to be heated in order to ensure good flow through the pumps. You see, both parts are mixed together during the spraying in a mixing block, and a good flow helps the application as it has a relatively short working time before the epoxy hardens. I have some experience of working with epoxy from a model making and installing strengthening rods in guitar necks, and the stuff can get very messy very quickly. To help with this, solvents were employed to help clear out the lines. The solvent chosen for the job was the highly volatile methyl ethyl ketone, or MEK. This stuff can be flammable at room temperature, so you need to be very careful with it. The sprayer was set up on a wheeled scaffold, which allowed positioning up and down the penstock when needed. The hoses were laid out with the sprayer in between the workers and the egress point facing towards the west. By late morning on the 2nd of October, workers began staging the MEK and epoxy for the beginning of the application. Is it just me or does this all seem very rushed? They had only just finished removing the old stuff and needed to begin the new epoxy application straight away. Anyways, the sprayer was set to be flushed out with the MEK. This was done with 10 US gallons of the stuff brought in in two 5-gallon plastic bottles. The remaining and used MEK after flushing the system was left in open buckets in case it was needed again. There were around eight buckets of epoxy hardener and MEK placed around the sprayer on the scaffold. We have another 95 placed along the penstock for when the coating operations pass by. The green light was given to start spraying around 1:00 p.m. And a team of 12, 11 of whom were inside the penstock, began working. The foreman and the manager around this time left for lunch. The work would get bogged down almost immediately with the sprayer getting clogged up on multiple occasions requiring flushing. After four flushes, the foreman decided it was best to abandon for the day and to prepare to remove the sprayer. They had only done around 10 ft of spraying. To do this, the MEK had to be thoroughly circulated around the whole spraying machine, hoses, wands, and the mixer block. The two workers who were manning the wands made their way back to the sprayer with their equipment and buckets of MEK and waste epoxy. Another member of staff brought more MEK into the penstock for the cleaning process. Meaning there was an estimated of a total of 12 gallons of pure MEK, roughly around 45 L, and another 12 gallons of MEK epoxy waste. Basically, 90 L of highly flammable material all in one tiny space. At around 1:55 p.m., MEK was still being circulated around the sprayer. A flash fire ignited from the base of the sprayer. Burning solvents erupted from the sprayer, shooting flaming liquid out into the workers. Some workers who had left the sprayer to find more MEK reported seeing flames rushing out along the penstock. Five workers were on the other side of the sprayer and found themselves trapped between the bulkhead and the inferno. They shouted for assistance and fire extinguishers, but none have been staged or in around the sprayer. They were trapped with no help to hand. Workers in the other side ran down the penstock to get the extinguishers, which had been placed outside the entrance cut into the steel. After retrieving the extinguishers, two workers ran back down the penstock to try and put out the fire. But more MEK had caught a light. This mix with burning epoxy created a thick black smoke, which prevented an attempt at fight at fighting the fire. A worker had run to the cabin Creek powerhouse to call 911. This was at around 2:00 p.m. With the first responders arriving shortly after. However, the confined nature of the fire was not communicated. Upon arriving and finding out that the fire was underground in essentially a tunnel, they requested mutual aid from other better equipped units. But this was at closest over an hour travel time away. Initially four firefighters attempted to enter the area of the penstock, but again the smoke and flames resisted their attempts. During this time, the trapped workers were still in communications via radio. They were told to go to the bulkhead as far away from the fire as possible. The comms would continue for roughly 45 minutes after the initial fire, but eventually silence would awash the radio waves. Rescue workers would continue however, as everyone still clung on to the hope of survival. Extra rescue workers would arrive around 3:45 p.m. The ventilation fans would be reversed to try and draw some of the smoke away. Eventually the fire would burn out, allowing mine rescue workers to enter at around 5:45 p.m. And sadly they would find five bodies, apparently having succumbed to asphyxiation. This would later be confirmed at autopsy. They reckon that it was around the same time that the radios went silent. The disaster would understandably result in a lot of questions as to how on earth did five people die in such horrific circumstances? We will come to that part shortly. But the plant will be pressed back into service and would undergo a massive overhaul finishing fairly recently. The investigation. So the big question of why and how would come down to the CSB. The investigation would be hindered at almost every turn with management from Excel and RPI refusing to provide evidence and testimony. With some managers pleading the fifth when interviewed. Later on Excel would even try and block the release of the CSB report. The CSB would find some massive issues with the project. First of all, the contractor RPI. It had a safety rating of zero and amassed over $100,000 in fines from OSHA. This meant that a company shouldn't have been anywhere near the job site, let alone bidding on it. Both Excel and RPI knew of the chemicals involved and the confined nature of the job. This would have required them to classify it as a permit required confined space as required by OSHA. If they had gained the permit, the two companies would have had to make sure that the following was followed. Having a written program, entry permits, continuous monitoring, and trained attendants. The setup used in the fateful day would have most certainly not met these requirements and thus the extra requirements would have cost time and money. It was found that during the planning stage of the project that the dangers of the single egress and entry point were known to both Excel and RPI, but nothing was done to improve the situation for the workers. Training was also found to be inadequate with no proper emergency services action plan being set out before the work began. What was the actual cause of the fire then? Well, it was pretty clear that the MEK was the flash point for the fire. Interestingly, the CSB also found that there were other less flammable alternatives open to Excel and RPI, but for whatever reason, both companies opted not to use them, instead using a highly volatile substance in a very confined space. As I said before, MEK can self-ignite even in ambient temperatures, which means it was a pretty terrible decision to use it. Both RPI and Excel would be charged criminally for the negligence of the fire. Move on the 1st of June, 2011, Excel Energy going to court. However, on the 28th of June, the jury found Excel Energy not guilty. RPI on the 19th of December, 2011, pleaded guilty to workplace safety violations and paid $1.55 million in a cash settlement, taking full responsibility for the deaths of the five workers. So, it's scale time. It's going to be a three. I mean, it's still what I've got from my root cause analysis card. Do you agree? Let me know in the comments below. This is a Plenty of Foot Production. All videos in the channel creative commons attribution share alike licensed. Plenty of videos produced by me, John, in a currently in the middle of July hot and sultry corner of southern London. And all I have to say is thank you very much for watching, and Mr. Music, can you do me a favor and play us out, please?