LOCATION, LOCATION, LOCATION
What have been the governing principles in the development of trains control systems in the United States? We can ask that question another way: Why did we create blocks, block signals, and block signal systems to begin with?
First principle: We operate trains at speeds where the distance required for the train to reach zero velocity [the braking distance] exceeds the sight distance of the locomotive operator.
Consequently we need to convey information concerning the condition of the "unseen" track to the locomotive operator. And the foremost condition we have to communicate is occupied or unoccupied?
Second principle: Within the defined section of track, we cannot determine precisely, or precisely enough, where the rear end of the train occupying the track is located.
We are so constrained by this second principle that we actually compromise our ability to maintain safe separation of trains. We "secretly" acknowledge the overlapping of authorities on tracks where automatic block signal systems are operational, provided the following train proceeds at a "restricted speed."
The unique characteristic of this speed that we use to buffer the possible violation of our vital "no overlapping authorities" rules is that it places a responsibility upon the crew that is totally at odds with the very first principle, the reason for developing the block signal system.
We organize block signal systems because we operate trains where the stopping distance exceeds the sight distance and yet we authorize trains to proceed by a signal indicating occupancy of the block based on the locomotive operator maintaining a sight distance twice as great as the stopping distance of the train.
We call that providing "adequate safety," and it is and does until it isn't and it doesn't.
"Adequacy" is, of course, a relative not an absolute, positive, function.
No matter how adequate this method for safe train separation is, it contradicts the very reason for our signal systems.
Now the first of those principles clearly retains its governing status in the deployment of positive train control systems. We always seek to operate trains at speeds where the stopping distance exceeds the sight distance.
But the second principle is not an immutable product of the simple physics of train movements. As a matter of fact, for PTC to regulate the speed of the train, whether or not block occupancy is the precipitating factor, the PTC system must be able to recognize the length of the train subject to its, PTC's, braking commands.
PTC employs a braking algorithm, a mathematical representation and accounting of all the factors that will impact a train's ability to decelerate. The algorithm provides a method, a calculus, for determining the braking effort that must be exerted in order to bring the train's speed into conformity with the approaching, or target, restriction within the distance to that target.
Train length is such a factor. When a locomotive operator initiates a brake application by moving the automatic brake handle on the locomotive, he or she is essentially opening a gate, allowing for movement of air in, between and among the brake pipe, brake cylinders, air reservoirs, brake valves, and atmosphere.
The movement of the air in the brake pipe, the change in pressure and volume, is not only an actuating mechanism, but also a signaling mechanism, propagating the operator's action throughout the brake system.
The signal has to travel from one end of the train to the other. It requires time to travel that distance internal to the train. That internal distance translates into time of delay, and that time of delay translates into the external distance the train travels before the braking effort is actuated throughout the train. This is an element of what we call "free run time."
Without knowing the distance internal to the train, we cannot calculate properly the external distance traversed during the time of propagation, and we cannot properly calculate the total distance required to reach the target velocity.
Now the nature of an algorithm is that it is a calculus, an estimation, derived from the analysis of an array of variables. The evaluation process will provide us with a most probable outcome, or profile, from which to establish, or in the case of PTC, establish and enforce, the braking curve for our train.
The greater the sample of data derived from different trains of various lengths, and various weights the more accurate the algorithm-- the closer the calculated braking requirement for any train will be to the actual braking requirement for that train. Quantitative analysis is too valuable a tool to be left in the hands of investment bankers.
It is even possible to "recycle" or feed the data from a train's previous braking effort back into the processing system in order to refine and improve the calculation of the braking curve for subsequent applications. This, of course, is exactly what the locomotive engineer does when braking the train, and why FRA quite rightly requires locomotive engineers to make a "running brake test" after taking charge of the train.
Automatic block signal systems, even in their most advanced configuration as cab signal/speed control block system do not calculate a braking curve for the train. In most instances, the speed control system requires the train to meet a target, not of speed as such, but of a steady-state rate of deceleration equal to the deceleration rate assumed in establishing the signal designed distances for a train to reach zero velocity. Speed is utilized only to measure compliance with that signal design rate of deceleration.
Certainly, such a system is "simpler" than PTC, but that simplicity is based on absorbing, embedding, that limitation of block signal systems-- the inability to determine, with any realistic precision, where the end of any train is located into the core of the train control.
PTC overcomes that limitation, in its basic design and functioning. However, instead of utilizing this technological advance to establish positive train control, we restrict it, bind it, smother it by making it subordinate to the limitations of the block signal systems.
We require PTC to display and enforce the indications of a wayside or cab signal system, and then we declare that doing so will be "close enough" to the Congress' mandate to prevent train-to-train collision, or movement through an improperly lined switch. We declare that the human operator operating at restricted speed prepared to stop is "close enough" to the mandate to positively impose a stop.
I am not arguing against block signal systems. I am not arguing that such systems should be abolished.
I am arguing that we not spend billions of dollars to establish train control that is not positive, that only enforces the weaknesses in the legacy train control systems.
I am arguing that for PTC to be positive train control, it cannot be established on the basis of minimal requirements that are "close" to its capability for maintaining safe separation of trains.
I am arguing that "close enough for government work" is not good enough for positive train control.
Copyright April 5 2012