having slept on this subject and giving a bit more thought to it today i thought i would post a couple of observations
when you look at automated commercial generators (the big boys) you find they generally do not control all facets of operation
from one controller, usually there is an autostart module, an electronic govenor, frequency sync if there are more than one installed
etc. Each having their own dedicated sensors and controlling specific actuators, relays, contactors etc.
one wonders why they do it that way? certainly they have the wherewithall to integrate into one controller?
likely it is done this way as a redundant failsafe method, should one part of the system fail other parts can control the engine and shut
it down if necessary.
so then the question becomes, why not follow the leaders in this reqard?
why not split the functions onto seperate controllers, beit microcontrollers( pic, stamps, plc, etc) that way there is redundancy of safety
built in. the down side is a few more parts,, but they are dirt cheap to start with in the case of pic's and stamps.
another thought would be to use some form of master controller, whose function is to simply enable a start command and send it to a
slave controller whose function is to do the start up sequence and report back when the engine is running
the master could then enable another slave controller whose function is switching in a generator and then apply loads etc.
the master would have ample time to do nothing but monitor for out of parameter conditions like overspeed, over temp, low oil etc.
and initiate an emergency shutdown of its own accord.
each slave could also have its own set of parameters that would trip the emergency shutdown sequence.
with a master, and the two aforementioned slaves all having access to the emergency shutdown of the engine, the
likelyhood of all controllers loosing control and failing to shutdown the engine would be dramatically reduced.
sort of follows pc archetecture in a way, it ressembles the ibm ps2 microchannel system with its buss master controllers
where the main processor released control to a buss master video card or a buss master scsi drive controller.
in my opinion the resultant coding is steamlined which in the end makes for less debugging and changes to one controller have
no effect on the function of another
going back about 25 year or so, i worked for an oil field fracturing company with detroits from hell
these were powered by the lowly 12v71 all the way to 12v149's and many 16v92's between
all were hotrodded to around 2700rpm, custom injectors, no exhaust etc. most were developing well over twice the
design hp that detroit had intended.
they all were plc controlled, with seperate controllers doing specific duty, engine startup, electronic governor , etc.
they used linear actuators as well and also were fitted with the blower air shutdown systems (which were deleted on the turbo
engines on oem engines)
faults from loss of govenor control would would trigger shutdown, faults from other plc controller would trigger shutdown
and on top of that were the mechanical murphy switch guages as another layer of emergency shutdown.
so where am i going with all this?
first of all i would recommend placing all the shutdown elements such as fuel rack, decompressor, fuel shutoff and most especially an air intake
shutoff on a single buss
second i would allow access (to this buss) from at least one master controller "AND" a slave controller (electronic govenor) "AND" a slave controller (engine startup)
and also from any other mechanical overrides, switches, murphy guages etc.
all shutdown elements would default to shutdown if there is a loss of power to any controller "and" to any shutdown element, this is very important
and is sound engineering practice.
just my thinking, yours may vary
bob g
