Until now, how ACSES would function in service was mostly speculation. Although Amtrak has had it in service for 15 years now, there have been few reports (and no video) of the system from an Engineers point of view. This made me optimistic that any impact would be minor and mostly involve positive stop points at stations. Of course there is wide variation and little regulation regarding how ACSES (and all other PTC systems) are implemented from a human factors perspective. There is also a fair bit of wiggle room in how the speed and stop enforcement take place with a good deal of it being policy decisions by railroad management, not necessarily the FRA or other third party safety scolds.
What I discovered today hit on both of these points, and neither fr the better. Note that the observations below apply only to SEPTA andwhat they saw fit to require of their vendors and signaling department. I know for a fact that similar railroads are examining other options. Hopefully if anyone involved in other projects they will take away some things to avoid.
Captured below are two runs on the Fox Chase Line from NEWTOWN JCT to the end of the line in Fox Chase and back. ACSES is in service on the entire Fox Chase Line NEWTOWN JCT excluded. The second video has a better angle on the cab display unit. Station stops were omitted and a few ACES / CSS changes were lost.
As you can see, the SEPTA ACSES system communicates the braking curve to the engineer by means of an Authorized Speed countdown system. As a speed restriction approaches the ACSES speed will begin to drop. If the train's speed is suddenly above the ACSES speed, at some point an overspeed warning will light and it sounds like 5 seconds after that there will be some sort of penalty application if either the brake is in the suppression position or the overspeed no longer no longer exists. Two problems with this system are plainly evident.
First, the braking curve is ridiculously conservative. Every time the ACSES speed began to drop, the engineer was able to get well ahead of the dropping curve without anything close to "aggressive" braking. The capabilities of the equipment have been completely ignored and some leisurely braking rate has been chosen. If PTC "ideal" is to stay out of the engineers way except in case of a dangerous condition, the penalty braking curve should follow the full service braking rate. SEPTA paid both $ and weight for full dynamic, disc and tread brakes on the Silverliner V, which should allow for later braking into curves and stations. Are we now to believe that was just a waste?
Second, the human factors of the speed countdown entourage the engineer to just proceed at slower speeds to avoid potential overspeed warnings. You can watch the student here get dinged by ACSES a couple times, and later, at the instructor's urging, keeps the train at a slower than authorized rate of speed in anticipation of another speed countdown. This is exactly the sort of behavior I warned of and is not exactly what is happening. In addition to a realistic braking curve, the system should try to be invisible and not trigger compensating behaviors. A warning light at 3mph above curve followed by a penalty application at 6mph is all that is needed. This is the standard used on the British Rail ATP system back in the 80's.
UPDATE: The effects of the ACSES roll out were evident in SEPTA's January 2017 schedule. Travel times were uniformly increased on the Media and West Trenton lines between 4 and 6 minutes. Again this runs completely counter to the PTC propaganda that operations would not be affected.
No comments:
Post a Comment