California SP runaway

[gabe]

I remember seeing a story about a Southern Pacific runaway in California. I wish I could be more specific, but a heavy unit train was coming down a very long straight hill, lost it, and then predictably derailed when it hit the curve at the end of the hill. To make matters worse, the clean up effort punctured a major gasoline line and caused an incredible inferno in the residential neighborhood about a week later.

Anyway, there were three units on the front end and, pursuant to SP's practices, the two pushers who helped the train up the hill stayed with the train to allow their dynamic braking to help ease the train down the hill (unknown to the poor engineer, 3 of the 5 units did not have their dynamic braking functioning and every car in the unit train was 40% heavier than he was told it was).

My question: the entire train derailed including the two trailing units (the story goes the engineer in the last unit putting the train in emergency was the event that lost the train for good).

Were the two trailing units forced to ride the train down to its inevitable demise? Could they have uncoupled from the train and just come to a stop when it became clear it was hopeless for the rest of the train?

Gabe

[Anonymous]

i actually had the tape of that on video, it was called incident at duffy street. Kind of sad, but you could see all the damage had been done, the train wad going pretty fast down that line, and there is a pretty tight curve there. they were saying that people who lived there were worried about the fact that the trains were there, and that something like this might happen. If I remember correctly, they either smoothed out the curve, or realligned the line to make sure that it didn't happen again.
Brad

[richardy]

I was shooting video in Barstow when the accident happened preparing to move to Summit. The radio was alive with communications related to the rescue work. We arrived a few hours later, it was the worst derailment I have ever personally observed.

Richard

[Anonymous]

If I remember it correctly, wasn't there some interfacing feature of the dynamic brakes that was preventing them from making an emergency air application? Some kind of cut-out or override? If the helper crew had cut off as mentioned above, and they left the rear car's angle **** open, would that have been enough to trigger an emergency application, or was their air already pissed away by then?

[mudchicken]

From one who was there at the initial derailment (acting duty Roadmaster for the railroad next door, asked by the DS to find where the train derailed over the radio)...

(1) Duffy Street was probably the worst derailment I ever was around (and I saw a few)...The videos do NOT convey the scale of the thing...Cannot ever forget how bright cobalt blue and orange the locomotive wheelsets were after the derrailment. I was fortunately in the bay area when the pipeline blew. (Many attribute that to heavy equipment nicking the pipeline exterior toring derailment clean-up .... The pipeline company failed to adequately test the line with a smart pig prior to placing it back into service....like in rail, a tiny defect was given a chance to grow with the resulting catastrophic failure)....

Instant Urban Renewal....

(2) Curve story does not fly, the relatively new piece of railroad on the Palmdale Cut-Off still predated the housing project that was Duffy Street. The current curve would not have fared any better with the speed and momentum of the train. The heavy train rolled-over the high rail in the curve before it left the top of the fill.

[gabe]

The pipe where the rupture occurred fit the teeth on back hoe perfectly.

[vsmith]

Wasnt one of the causes of this accident determined to be that the gravel/ore cars were each overloaded 10 to 20%, they were filled literally to the brim? The engineer thinks his train weighs X tons when in reality it weighed Z tons? Somewhere I remember this being a primary cause of this accident. The train was in serious trouble from the moment it began downgrade.

[dldance]

One analysis of the wheels showed that melting had occured. Thus, there was sufficient air pressure to set the brakes, but the resulting friction melted the surface of at least some of the wheels which lubricated the brakes eliminating braking effects for those wheels. That melting could cause the coloration of the locomotive wheelsets that Mudchicken mentioned.

dd

[jeffhergert]

QUOTE: Originally posted by Clemente

If I remember it correctly, wasn't there some interfacing feature of the dynamic brakes that was preventing them from making an emergency air application? Some kind of cut-out or override? If the helper crew had cut off as mentioned above, and they left the rear car's angle **** open, would that have been enough to trigger an emergency application, or was their air already pissed away by then?

I think it works the other way. On some locomotives, an emergency application would override the dynamic brakes. There is a note in my UP Student Engineer's manual about this saying: UP locomotives equipped with dynamic brakes have been modified to maintain dynamic braking capapbilities during an emergency or penalty brake application.
There is also a Dynamic Brake Interlock that is supposed to keep locomotive brake cylinder pressure from developing during automatic brake applications while using dynamic braking.
Jeff

[mudchicken]

QUOTE: Originally posted by M.W. Hemphill

Mud: To quote the NTSB: "The NTSB determined that the probable cause of the pipeline rupture on May 25, 1989, was the inadequate testing and inspection of the pipeline following the derailment that failed to detect damage to the pipe by earth-moving equipment."

With brain to fingers on the keyboard disconnect aside, sounds like it. Do not remember feeling the heat from the wheelsets even though I was right up against them. Do remember hacking & sneezing up gray dust for days afterward along with the general bedlam at the site.

Put the fear of God into me that day, for sure.
[:0][:0][:0]

[vsmith]

QUOTE: Originally posted by M.W. Hemphill


7. Because the mine and the railroad estimated the lading weights, rather than scaling the cars at the mine, the train was waybilled at 6,151 tons or 89 tons per operative brake but was actually 8900 tons and 128 tons per operative brake. By SP rules, a train exceeding 125 TPOB and 18 axles of operative dynamic braking would not have been permitted to descend the hill eastward from Hiland to West Colton.

8. Computer simulations found with 6,151 trailing tons and 24 axles of dynamics -- the amount the train crew believed they had -- the train was controllable and the descent was made at 30 mph. At 8,900 tons and 18 axles of dynamic braking, and including the efficiency losses of the brake shoes as they heated, the train exceeded 105 mph.

Summary: Had the engineer gone straight to a full-service brake application at the time he increased the reduction to 13 lbs., while the train was still in the curves at the top of the hill, he would have been able to stop the train. One the train had reached MP 469 and gone to 21 lb. reduction, the train was no longer controllable.

Thanks, I knew that a primary cause was overloading, I didnt know about the Sims. That load is an almost 45% greater total train weight than what was estimated. Thats unforgivable, but would an engineer today given the same information and the same conditions do things any differently? He was cursed with faulty information on his train. By the time he realized what might be the cause, it was too late and it was a runaway. [V]

[vsmith]

Thanks for the info. Not being an Engineer, wouldt the train being almost 50% overloaded have been laboring up to the summit? I know that there is a grade from Victorville towards the top of the Cajon summit . Should that have been a red flag?

[Anonymous]

This train came to Cajon Pass on SP's Palmdale Cutoff. It had to come down a short hill from Ansel, but after that it's mostly high desert running with slight ups and downs until starting down the 2.2% grade below Hiland. No sustained uphill grades like on the Santa Fe out of Victorville.

[Anonymous]

QUOTE: Originally posted by M.W. Hemphill

Vsmith: I think an experienced, smart, and attentive engineer would have sensed that something was wrong the moment the train didn't sit down as expected with a 10-12 lb. reduction, and dumped the air right then. It's been my fortune to ride with some very good engineers, and I think they would have figured out the train was way overweight long before that point from the way it pulled.

One of our first true tests as Engineer Trainees was taking a loaded coal train (100+ cars) down a certain secondary track with a five mile grade between 1.8% and 2.1%. Before that time you have run a local carrying rock salt with about 40 loads over the same hill. One oldtimer I worked with in my training liked to refer to it as "Breaking in your ***". It was very true. It really helps you to develop that "feel" in the engineer's most important sensory organ for train operation, his backside, in "The Seat". I can't really describe the difference in words between the two trains, but the best I can do is tell you that the coal train handles like a huge heavy boulder you are directing down an incline by walking in front of it. You know it is never really in your control you can merely keep it on track and steer it, you can stop it if you must, but it must be done very slowly and deliberately or it will run right over you. The salt can be very challenging too, but not nearly as hard as the coal. I can remember almost stalling with 42 salt loads and 4 SD50s pulling flat out in 8 with sand on the way back up that hill. I can only imagine what a train overloaded by almost 3,000 tons would be like...

LC

[Anonymous]

Gabe-

As to whether those helpers could uncouple, I don't know how they were set up. I'd guess they could uncouple on the fly as SP coast line helpers had that ability. Having been out on locomotive catwalks at 30 or 40 mph trying to restart a balky engine I can tell you it isn't any fun. At the speeds that train was achieving I'd be willing to bet they were going so fast by the time they thought of pulling the pin it was already too late.

LC

[ericsp]

Too bad the NTSB does not have documents that old on its website.
http://www.ntsb.gov/publictn/r%5Facc.htm

Does anyone know why the crosses for the two railroaders killed were place at the Tehachapi Loop instead of Cajon Pass?

Mark, have you ever seen the Discovery Channel show about this derailment?

[heavyd]

I remember seeing a picture somewhere of the derailment site years later and there are still NO HOUSES. The street is empty. Even as a railfan, I don't think I would want to live there either!!!

[peterjenkinson1956]

last week on pay tv there was a program called seconds from disaster...this program tells the story of different disasters around the world..it showed this derailment in great detail it interviewed people and showed evidence...the wagons were overloaded and it showed the diference this would make and it showed the investigation about the fire and how the fuel company kept pumping fuel...keep a look out for this program..the amtrack train cra***hat happened when the barge hit the bridge is another episode...peter

[adrianspeeder]

I saw part of the discovery channel show. Didnt some lawyer come in and screw everybody too?

That much weight moving at 100+, I dont even want to imagine what that crew thought right at the base of the hill. Not even enough time to say "Oh sh"

Adrianspeeder

[kenneo]

At the time of this accident, and for all of the years prior from about 1965, when an empty was upgraded to a load in the computer, an estimated net weight was to be applied because the car would be weighed and the record upgraded before the car was to operate byond the first division terminal. I would never upgraqde the standard (50 tons or "050") but would use the loadlimit as the capacity. As a result, trains operating on the "standard" entry would always seem to need to double any real hill, and when you were pulling a train up 2.86 and down 3.3, that was a real important piece of information to mess around with.

The train, when the cars were released and pulled in the computer would have been upgraded (and, I am told, were) with the standard "070" which would account for the 3,000 ton difference from what they should have been upgraded to - "131".

I also seem to recall that the train crew did not have too much experience. If you have a problem as they did that morning, on a 2% grade, you need to remember that your train will increase 5 MPH per second of uncontrolled movement. 5 seconds = 25 MPH INCREASE. On a 3% grade, from a standand start, in 2.3 miles, you will be exceeding 100 MPH. Try going around a 12 degree curve at that speed. This I can tell you, the outside rail won't roll over because the cars will walk over the railhead and become short range airplanes - maximum range about 150 feet. Flying parts for a few hundred yards more. One more thing to remember - it is not the runaway that does the damage, it is the sudden stop at the end that's the killer.

I have had the power fail on 3.3%, and the resulting change in forward motion is instantanious - puts you in your seat like accelerating from 0-60 in 3 seconds. You need to get a full service application - independant bailed - faster than fast - you really have about only 1.5 seconds to get the application started back. You have to wait to **** your pants until you have cleaned the clock.

After this incident, the computer was programmed to make a car weigh at its capacity plus tare (both numbers were in the UMLER file) and updatable only with a scale ticket entry.

[Sterling1]

QUOTE: Originally posted by M.W. Hemphill

Eric: I saw the Discovery Channel show. The video was interesting and useful. The narration was pretty breathless and silly.

Yeah I remember seeing that show some time ago, and now that I reflect on the incident, I find it quite sad[:(]. The narration said that the DB cut out, and I had never knew about that. I mean, do locos still have that feature? Seems quite sickening that the DB would cut out when the Emergency brake would be applied[:0].

[BigJim]

I saw this show on TV a few years ago, so please correct me if I am wrong about the following.
If I remember correctly, the show stated that an inexperianced clerk messed up the tonnage figures. Not knowing what was in the cars he looked and found that the cars were only loaded to maybe 3/4 capacity. Thinking this mineral weighted the same as coal and that since it wasn't loaded to the brim, the car must weigh less than a full load of coal, so the lower tonnage figures for the entire train.

[gabe]

I don't understand how snow, or worse ice, on the rails doesn't hurt a train's grip on the rail both for purposes of braking and for the purposes of tractive effort. Is the weight of the wheels so heavy that it creates enough kinetic energy for it not to be a problem?

I have seen what snow/ice can do to a rubber tire (which has better adhesion than metal), I am just at a loss as to how mountain railroads deal with this condition.

Gabe

[Sterling1]

QUOTE: Originally posted by M.W. Hemphill

Most locomotives no longer have a PC switch, which disable the dynamic brake and return the prime mover to idle when an emergency brake application ismade. Control systems have been redesigned to use vastly more sophisticated methods to avoid the wheel-slip condition that the PC switch prevented. (PC stands for Pneumatic Control, by the way.)

There's nothing necessarily "wrong" or "unsafe" about a lead locomotive equipped with a PC switch so long as the operating rules for mountain-grade operation are written in such a way that dynamic braking is not being counted on for a safe descent. SP's rules were so written. (There were some other problems with the rules, but they were not major factors in this runaway.)

These were the pertinent questions in this runaway:

1. Were the operating rules of the Southern Pacific written in such a manner that a train that cannot be controlled on a descent is prohibited from initiating that descent?
2. Were the conditions assumed by the rules in fact "true conditions" -- that is, was the crew properly trained, the equipment said to be in working order truly in working order, and was the weight of the train known accurately?

In this case, the rules could have been written better. The crew was inexperienced. Some of the equipment reported to be in working order was not in working order. Most important, the weight of the train was undereported by 45%!

A railroad's operating rules for mountain-grade operation recognize that trains come in all tonnages and lengths, with few locomotives or many, with all of the locomotives on the head end or interspersed in other locations in the train, with or without locomotives equipped with working dynamic brakes, and outside in a broad temperature range. Thus the rules present a flow-chart that enables a crewman to insert the train's trailing tonnage, its number of operative air brakes (one per car except for cars such as articulated double-stack cars), number of axles of effective dynamic braking, and in some cases the temperature. The train is inspected at its initial terminal and at required intervals en route to ensure that brake-cylinder piston travel is within limits, brakes shoes are present and within wear limits, and that the brake line is competent and that the brakes can be set and released. The train crew is given a trainlist that reports the tonnage. Before beginning the descent of the mountain grade, the train crew inserts their train data into the flow chart, and follows its decision path to determine (1) if their train is legal to descend that hill, (2) what speed is permitted for the descent, and (3) if retainers are required. If they do this correctly, the inspections have been done correctly, and the information reported to the train crew is correct, and the rules are correct, the descent will be made safely.

At the time, SP's rules (as well as most railroads' operating rules) contemplated that a train could have a complete dynamic brake failure after departing the top of the hill -- which is the condition, of course, that a train encounters with an emergency brake application when the PC switch cuts them out, and there are many reasons to initiate an emergency brake application (or have one initiated unexpectedly), not just to stop a runaway. Therefore, SP's rules prohibited a train from descending from Hiland to West Colton that could not be controlled by air brakes alone.

However, SP had gotten into the habit, as kenneo explains cogently, of underestimating weights. To quote the NTSB: "The accepted practice of estimating weights at the time cars were released, coupled with the belief that these weights would be changed at a later time, created a potentially hazardous situation in which yard clerks were merely satisfying a requirement of the Southern Pacific computer system."

An additional error arose in that SP operating rules stated that SP cars equipped with empty-load sensing devices had a braking capacity of 1.5 times normal, but in reality they had a braking capacity of 1.0 times normal.

An experienced and attentive crew, in my opinion, would have known something was up within moments after releasing the air at Mojave Yard and beginning their trip. A train so far over its reported tonnage will accelerate and decelerate noticeably differently than a train of that true tonnage. However, neither the road engineer or helper engineer observed that brake applications to control speeds on previous descents were excessively heavy, nor that the train was excessively sluggish under power. Nor did they seem to be aware that a heavy tonnage train coming over the top and beginning the descent of a mountain grade must "balance the grade" immediately, and if it cannot be balanced, an emergency brake application must be made at once. (Balancing the grade means that the train remains at a constant speed that is a safe speed at a constant braking effort without additional control input.) Thus the engineer let the train accelerate toward its 30 mph authorized speed (based on the wrong weights) and then looked to see if he could balance the grade, and when he couldn't, he did not dump the air immediately.

In making numerous trips over Soldier Summit, Utah, with experienced Utah Railway and Rio Grande engineers on 15,000-ton coal trains, they emphasized to me the crucial importance of balancing the grade immediately after leaving the summit. As one said to me, if he did not have the train balanced by a certain landmark after beginning the 2.0% descent (which I later determined was 4,500 feet beyond the summit), he would dump the air at once, and if the train did not promptly begin to decelerate, he and his conductor were leaving the train for the bushes while they still could.

Another trip, the same engineer and conductor later told me that one winter night they came off the summit in heavy snow. They were the first train down the westward main in about eight hours. They were plowing over a foot of fresh snow with the train, and apparently so much snow had built up on the trucks that the braking effort was very poor. He couldn't balance the hill, so he dumped the air. The train refused to decelerate so he and the conductor took their grips out on the front platform and started looking for a good place to jump off. While they dithered looking for a good snowbank, the brakes finally burned off enough snow buildup to take effect, and because there wasn't any wheel heating to speak of, they got a good solid brake application, and despite the speed having gone to 45 mph the train began to slow and stopped. Then the conductor tied handbrakes so the engineer could get a release and recharge, and their were able to come off the mountain safely. Since that time, he said, when there was a possibility of snow or ice buildup he made enough of a brake application at the summit to burn off anything that was there, then a full release and recharge using the power to hold the train on the grade while the helper cut out, and only then did he leave.

For how long on the road are engineers trained for? the Case that you stated above require many months or years working the same locations.

[BigJim]

"you'll see that's not the case."

Mark,
If you will READ my post, I was refering to what the program stated, which was that someone didn't know what the cars should weigh, looked inside and saw they weren't loaded to the brim and figured they had to weigh less than the gross weight stenciled on the car. That is the way that I remember the show.

[ValleyX]

Mark, I'm sure that most locomotives have a PC switch, all locomotives I see have a PC switch in them and I'm not familiar, nor have I heard or read about any that do not. What has changed is that they now have a delaying feature that they didn't used to have to prevents the dynamic or power, as the case may be, from dropping out all at once. Now, both power and dynamic have a 20 to 30 second delay before the amperage drops. The trick is, for the engineer, is to keep the slack bunched with the independent or try and keep the independent bailed and the slack from running in and out in the event of a undesired emergency.

[mvlandsw]

On CSX the PC switch still cuts off the power after about a 20 second delay but the dynamics will continue to work indefinitely. I don't know why it was ever done any differently. All B&O and Chessie units had a dynamic brake interlock which prevented any independent brake cyclinder pressure from developing from an automatic brake application while in dynamic braking so wheel sliding was not an issue.

[kenneo]

Balancing the grade ---- in Oregon, at any rate, the standard practice when we still had cabeese was for the engineer to ask the conductor what the air guage in the caboose read just as the road power reached the top of the hill. It had to read above a specified limit individually set for each hill and was based on tons per operative brake and axels of operative dynamic brake.

To illustrate, at Cochran, the maximum TOP was 100 and AoDB was 30, and BP in the caboose was to be 90. Anything heaver, or fewer axels of dynamic or lower BP required a complete stop, retainers set up and speed not to exceed 15 mph. The number of retainers was calculated based on the three factors mentioned above and could actually mean every car had its retainer set up. If you had less than 100 TOB, 30 Axels of Operative Dynamic Brake and 90 lbs in the caboose, you could proceed without stopping.

The process, not specified in the rules or special instructions, to balance the hill, was for the engineer to make a 10 lb reduction prior to 1/8 of his train departing the top, dynamic brake under full application within the next 15 seconds, and then the service application was to be released. This allowed the dynamic to take effect, kept the train under control while power was reduced and then the changeover to dynamic was made, and permitted the train to bunch slowly as each car released to control the slack, and the air to recharge if a stop needed to be made. The train usually approached the top at about 6-8 MPH (full tonnage) and would proceed downgrade eastward at 20 mph and westward at 15. This is a bit oversimplified, but I wanted to illustrate what we are talking about for those that (it seems to me) were asking for a bit more explanation.

At Cascade Summit, eastward trains (northbound by compass) were required to make a service reduction of not less than 10 lbs far enough prior to the beginning of the descent that the air could be recharged and equalized to 90 lbs in the caboose and the balance of the rules above applied (not the illustration of balancing the hill at Cochran).

My understanding is that the UP required a complete stop and a standing set and release (what amounts to a terminal air test - no "blow and go").

My understanding of the air-dynamic-TOPB equation on the runaway is that that particular train should have had all cars "turned up" and wheel radiation stops made at intervals not to exceed 10 miles for 15 minutes (minimum) at each stop and speed not to exceed 15 MPH. And while it is stopped, the crew is required to make a wheel inspection for heat damage after they (wheels) have cooled prior to departing.

I can only speak for the Oregon Division prior to 1993.

That load-empty brake rule. The only thing that I can figure out is that "empty" = 1.0 and "load" = 1.5. Load-empty braking was supposed to make an empty act like a load when the brakes were used. Intent was to eliminate undesired slack action and eliminate flat spots. But load-empty brakes do not increase the amount of braking, but rather decrease the amount of braking on empties.

As any trainman who has ridden cabeese can tell you, empties "squat" when the brakes are applied and can really drag a train down quickly (speed wise) because the brakes are set to stop a load. Rules of motion state that any body going in a specified direction at a specified speed will continue to do so unless an appropriate force is applied to change things. More weight on the move means more force needed to stop it. With a freight train, when you get to 23 lbs per square inch of shoe pressure against the wheel, then that's all there is, folks. There ain't no more.

[MP173]

kenneo and others:

I am impressed with what has to be done to safely get a train down a grade. I feel the same way with trying to read music...I understand about 1/3 of what I am "reading", but it is still fascinating.

Now, a couple of q's:
1. Do mountain crews get paid more than other crews, based on the difficulty of the job?

2. By retainers....do you mean hand brakes?

ed

[PNWRMNM]

Further on retainers. They are designed to release very slowly. The idea is that you set your air, then when you release the retainers retain (for a time) the set in the train. With the engineer's brake valve in release, the prake pipe and air reservoirs on the cars are recharged. As the retainers release on a long grade (the only place you use them) speed increases and you have to make another set. Because the recharge is typically less than complete you typically have to made a bigger reduction each time so there is a limit to how many times you can do the process.

Dynamic braking and pressure maintaining have almost elimiated the use of retainers.

Mac