Ten90 Solutions LLC

Railway Age has published on its website an interesting article  "Push the envelope with autonomous freight trains?" by Robert L. Peterson of SMART-TD locals 303 and 951.

The issue in contention is, as the title says, the prospect for the operation of "autonomous" (without on-board crew members) freight trains in the United States.  The issue is made more acute by the initiation of limited autonomous train operation in Australia. 

Mr. Peterson identifies the significant improvement in railroad safety made over the years, reporting that since 2000 total train accidents have declined 44 percent, with equipment caused accidents falling 38 percent, track causes down 55 percent, and derailments down some 42 percent. 

Mr. Peterson concludes  that: 

The data implies an improved safety trend is due in part to current operating rules that utilize two-person train crews and technological safety advances in train operations.   

Uhh...pardon the interruption, but no, the data implies no such thing, since there is no measure of crew size; no, since the declines in equipment and track causes of accidents have nothing to do with operating train crew size.  The decline in equipment and track causes has everything to do with better practices, better management and better technology that allows for railroads to target reliability, which precipitates action before failure, rather than repair, which precipitates action after failure.

Strangely enough, Mr. Peterson leaves out the decline in human-factor causes for train accidents, which in the same period declined, in absolute numbers, some 48 percent.  A good record, indeed, but again it tells us nothing about the safety of two person train crews.  A significant number of the decline in human factor accidents is certainly due to improvements in train control systems, roadway worker protections, training and certification requirements, which improvements can be coded directly into autonomous train operations.

In fact, Peterson has almost nothing to say about human factor causes for train accidents, only referring to such causes obliquely, writing

  ...according to the FRA data, 30% to 40% of railway accidents are "track-related,"...second only to human-caused. 

Now you might think that if it's the safety performance of train crews of any size, then the largest category of accident causes-- the human factor causes--would trigger some further analysis.  You might, but it didn't in Mr. Peterson's piece. 

So let me point out a few things:  first between 2000 and 2018, the frequency of track failures as a cause of rail accidents declined from a rate of 34.7 percent to 29.0 percent.  This becomes a stastically significant improvement when railroads can maintain the decline as a trend.  

Meanwhile, the frequency of human factor causes for accidents declined from 39.3 percent in 2000 to 37.4 percent in 2018, except the trend from 2009 through 2015 shows an increase in frequency of such causes before the 2018 decline.  The fact is that human-factor causes of train accidents occur at a rate that has stubbornly resisted reduction over the years, even while the total numbers decline. Just for what it's worth, in 1979, human factor causes accounted for 29.1 percent of train accidents. 

Now Mr. Peterson says: 

To the average reader, bold, matter-of-fact-sounding statement claiming no data exists indication two-person crews are safer than one-person crews can be misleading .  Such statements make no mention of data suggesting that one person crews are safer than two person crews, or that automation is safer than no train crew at all--because such data does not exist.

Say what? If there is no data showing two-person crews are safer than one-person crews, then there can be no data showing that one-person crews are safer than two-person crews.  There is no data.  It's just that simple.  No data is no data. 

Railroads are not arguing that one-person crews are safer than two-person crews.  The railroads are responding to the claim that two-person crews are safer. Those making the claim, the assertion, the allegation that two-person crews are safer have to provide the evidence to support the claim. 

Railroads are claiming that, with technological enhancements, and mobile support personnel  in case of breakdowns, one-person train crews do not present any additional risk to safe train operations.  That is quite different from claiming such operations are more safe. 

I am not an advocate of one-person train crews in general.  My position is, and has been, until you can operate trains safely with no crew members, it is short-sighted and inefficient to operate trains in main line road service with a single crew member.  Short-sighted in that valuable skills will be lost; inefficient in that over sections with limited capacity, over single track, or multiple tracks where headways are 30 minutes or less, the risk of breakdown and the increased time it will take to identify, assess, and correct the breakdown will cost more in lost "slots," lost track capacity, than will be "saved" in reducing the crew cost. 

Safety is not the issue.  Collective bargaining is the issue.  We should not confuse the one with the other.

David Schanoes

February 18, 2019

Your ad here. Actual size