Are there any plans to re-open Buckeye Yard in Columbus with the pickup in business? It seems there is a general slowdown in the NS system these days, or at least thru Elkhart. Is this due to weather from a couple of weeks ago out east, traffic levels increase, laid off crews or what?
What is involved in shutting down a yard? Obviously new blocking patterns must evolve. How much planning must go into this and then re-opening?
Ed
I haven’t heard anything, but I usually don’t until it happens! I don’t know the nuts and bolts of shutting down a hump, such as anything you need to do to the retarders, hump computer, etc. But, the blocking patterns and train schedules all get done ahead of time. NS has a traffic flow model and blocking model that easily accommodate changes to the network. It is used to figure out the trains to operate and their blocking. Once the best scenario is worked out and agreed upon, the blocks are added to the blocking system and the train schedules are added/changed in a staging database. On “go”, the staging database is migrated to production and the new plan is up and running. An effort on the order of opening or closing a hump is done in a matter of a couple weeks. After implementation, the block and yard volumes are monitored pretty closely to make sure they are close to the model’s prediction and the plan is tweaked from there.
Thanks Don. That would be an interesting project. No doubt the historic data will support such a blocking scheduling change. NS has quite a system of routes, yards, and operations.
With traffic coming back (is it?) what are the determining factors in looking at changes? When revenue/traffic is falling, obviously one looks at reducing costs. Would a factor in a rising economy be increased dwell times? Late schedules? Recrewed trains? etc?
Ed
Who in the railroad works this out, and who has to agree with it, for it to go forward?
There are staff people whose job it is to monitor how well the plan is working as well as design changes to it. They are also responsible for making sure the revisions don’t violate customer commitments. Once they work out what the changes are, they send out the changes for approval of the operating folk who will have to operate to the revised plan. This includes the crew and locomotive control folk. The plan has to be approved by all before it is implemented.
That is a pretty major undertaking.
New business would be a challenge, to see how it fits into the matrix with levels of service.
Which do you think would be more of a challenge, ramping up the carloading levels, or planning for lower levels? My guess would be the lower levels…lots of assets to park, crews to furloigh, blocking patterns to change, etc.
Is that all done with simulation models based on historic or anticipated volumes to end point yards/destinations? Does it get specific with customers? (shipper ABC with certain volumes?)
Ed
Last I saw when driving over Buckeye Yard, they were still using it for storage of cars and locomotives. Before reusing the yard, I think they would have to either return that equipment to service or find another location for storage.
Kevin
I would agree. When volume goes up, you start running extras. When volume goes down, you have to find combinations to make. Much harder to do while keeping the plan efficient. The planning tools had a lot to do with being able to keep the plan decent while volumes were crashing.
Both. Marketing has a model they use to forecast loads. It can feed the traffic model. The traffic model also will take a historic sample of waybills as a source of traffic.
From the NS website* -
“Operating Plan Developer (OPD) provides flexibility to make adjustments to the network operating plan when required. NS uses OPD to generate what-if simulations on yard, train and blocking changes to its merchandise transportation network. OPD allows planners to quickly evaluate the impact of infrastructure and traffic changes on TOP to ensure the rail network is effectively utilized. By using traffic volume projections, transportation planners can ensure an operating plan is designed to meet the demand for higher business volumes. The goal of this enhancement is to reduce train starts and improve network fluidity.”
(There’s that word again - ‘‘fluidity’’ - we had a heckuva debate about that term here about a year ago . . . [:-^] )
This whole topic would make a good article for Trains, in my opinion. I believe there are only 2 or 3 firms** that specialize in writing the ‘master’ software, and each railroad has pretty much used a custom-fit and proprietary version of it. Some
Interesting article? I want to read the book.
Paul, your ability to find stuff on the web is right up there with Dale. A tip of my Chicago Cubs baseball cap to both of you for your efforts over the years.
If you look long enough on the web, one can find old routing/blocking guides and freight schedules, particularly for PC or Conrail.
Amazing how everything is based or built on the “correct volume”. Operating plans, freight schedules, power planning, crews, tonnage, length, yard capacity, local operations, etc. are all based on the right volume. All of which flows down to the bottom line.
Flexibility is key to fluidity.
Now, I dont know what I will read tonight…all those links you posted Paul, or the Coal special issue.
Guess it will come down to either unit trains or mixed trains.
Ed
Don: are there at least preliminary plans built around any major infrastructure failure closing one or more yards?
Yes.
I forgot all that good stuff was there! The only thing I’d add is about ABC. NS’s system is different from the other roads. Most roads use static blocking tables that determine the classifications made at different locations. They define the path the car takes across the network. You are going from A to B? Then, you go via C and D. NS’s system considers multiple paths and then picks the cheapest one. It looks at the network like it’s a system on resistors, with each handling point assigned an impedance based on the overall cost to classify a car at that location. Then, it picks the lowest impedance route.
Buckeye Yard is still technically “open” with local jobs based out of it, plus to interchange with the Ohio Central. As for the hump yard I have heard that Buckeye is finished for good as a classification yard. The CSX Intermodal yard is also still open. I wouldn’t be surprised to see parts of the yard be used by the proposed 3-C corridor passenger service in the future as this is on the proposed line connecting Cleveland, Columbus and Cincinnati. This would be the middle of the line and the car shop and engine service areas are already in place. That is just a guess though.
Don:
Does it consider real time costs and conditions or historic. For example, every indication now is that Elkhart is plugged up. I hear trains being tied down in sidings and overall congestion.
Does the system take into account that there are real costs to these conditions and look for alternative routings or do adjustments to the system need to be made?
Ed
That’s interesting - and recognizing that is undoubtedly a greatly simplified explanation for the masses here - so it appears to correlate the costs based on the nodes, and not the ‘links’ ?
In the civil engineering business we have a couple of similar tools. One is a numerical methods procedure generally known as the Hardy Cross* Pipe Network Analysis method, which is commonly used for analyzing city water distribution networks, for example. In contrast, it focuses on the links, not the nodes - which are usually considered to be frictionless, or nearly so - just the usual head loss/ pipe friction for any fitting, such as elbows or tees, etc. Further, a ‘link’ or pipe segment need not be constant or consistently the same along its entire length - the pipe size, grade, and roughness coefficient, etc. can change any number of times along each segment. Then once set-up with the network parameters and the characteristics of each pipe segment, any set of flows into and out of the system - as from reservoirs and large users, etc. - can be analyzed/ calculated for the steady-state flow of the system in equilibrium, with the least possible ‘head loss’ due to friction on any route, as well as t
Ed / MP173, I don’t think there’s a book - yet. But below is something that might come close . . .
Thanks much for the compliment, but the sad truth is, I’ve been to most of those sites before - several times, even - so that this wasn’t any kind of a ‘start from scratch’ effort for me. But I think Dale and Mike/ wanswheel do much better at finding the true ‘unknowns’. This was more like just recalling the names or acronyms, and then letting Google do the work . . . [:-^] I’m glad you find it interesting, but that would figure . . . there’s a few of us here, I suppose.
Anyway, here’s a link to what used to be ‘‘RASIG’’ = the Rail Applications [Operations] Special Interest Group, which is now the RAS = Rail Applicatoins Section of INFORMS - http://www.informs-ras.org/index.htm I encourage anyone interested in this kind of thing to explore the newsletters at the on-line presentations from the various conferences, such as at the INFORMS 2009 tab or link.
For instance, pertinent to portions of this thread is the Oct. 11, 2009 presentation by Dan Plonk of Norfolk Southern on the Next Generation Car Routing System at Norfolk Southern at - http://www.informs-ras.org/Presentations/2009/SC1.pdf [20 pages, approx. 1.10 MB in size for this ‘PDF’ format version]. On page 8 of 20 it addresses the question posed by blue streak 1 above as some of the characteristics of the ‘ABC’ = Algorithmic Blocking and Classification as follows -
Facilitates easy, quickly reversible, changes to the operating plan
Yes. Nodes now, links are coming. There’s a bigger difference between the nodes than the links. A hump yard is a cheaper place to classify cars than a small, flat switch yard, for example, but 100 miles on a train isn’t much different than 110 miles.
[quote user=“Paul_D_North_Jr”]
‘ABC’ = Algorithmic Blocking and Classification as follows -
Facilitates easy, quickly reversible, changes to the operating plan
Normal plan tweaks
Derailments
Disasters
OK - I can tell there’s an engineer in charge down there . . . [swg]
From the NS webpage that I referenced yesterday -
http://www.nscorp.com/nscportal/nscorp/Investors/Financial_Reports/Investor%20Book/technology.html
Operating Plan Developer (OPD) provides flexibility to make adjustments to the network operating plan when required. NS uses OPD to generate what-if simulations on yard, train and blocking changes to its merchandise transportation network. OPD allows planners to quickly evaluate the impact of infrastructure and traffic changes on TOP to ensure the rail network is effectively utilized. By using traffic volume projections, transportation planners can ensure an operating plan is designed to meet the demand for higher business volumes. The goal of this enhancement is to reduce train starts and improve network fluidity.
‘‘TOP’’ is the ‘‘Thoroughbred Operating Plan’’. Although it’s not described on the above webpage, it is described in more detail on slides 4 - 6 of 20 of Dan Plonk’s RAS presentation that I linked above. Pertinent to and expanding a little bit on Don’s post above are these excerpts from slides 4 and 5 [emphasis added - PDN] -
Thoroughbred Operating Plan
Planning