Automation software/hardware

[JohnW2]

Precisely, and I believe we've shown that compared to the manual system it no worse. ... What if the operator falls asleep ? What if he goes for a "call of nature" ? What if he's just watching something else and misses a visual signal ? As you should be aware, a manual system is very error prone - which is why manual systems are reduced or eliminated wherever practical.

That does, indeed, seem to be the part of the argument which is being overlooked. I obviously don't know for certain that the approaches being discussed would be 'no worse than the manual system' - but, for reasons you have described, I strongly suspect that such is probably the case.

I suspect that one of the problems is that, whilst a mechanical/electrical system can be tested and analysed, resulting in various 'quantitative' estimates of reliability, to do the same (and hence 'produce numbers') with a human-involved manual system can be difficult or impossible. That can perhaps be taken, erroneously, to suggest that the non-manual system is 'more reliable', even when it isn't.

Kind Regards, John

 

[bernardgreen]

If the system has to be "really safe" then the following may satisfy the extremely fussy.

When the pump has started and the sprayer is spraying a command is sent to the pump that the pressure at the pump is OK. This pressure is then noted as an acceptable pressure. From then on the pressure at the pump is monitored and if it then deviates too much from the acceptable pressure then the pump shuts down.

The alternative is to not pump at positive pressure but instead to suck with negative pressure created by a pump on the spraying end of the pipe.

OH NO that might end up with someone being sucked into the pipe.

 

[JohnW2]

If the system has to be "really safe" then the following may satisfy the extremely fussy. ... When the pump has started and the sprayer is spraying a command is sent to the pump that the pressure at the pump is OK. This pressure is then noted as an acceptable pressure. From then on the pressure at the pump is monitored and if it then deviates too much from the acceptable pressure then the pump shuts down.

... provided that the process for doing the 'shutting down' does not fail! It is clearly safer to have a (usually) 'fail safe' system in which the default state is for the pump to be shut down, and for it only to operate whilst it is receiving a regular stream of 'do not shut down' commands - but even that mechanism could stick/fail.

As discussed, it comes down to what you mean by "really safe" and how "really safe" one is prepared to accept. To get anywhere near that "really" would probably require some redundancy - e.g. at least two, independent and independently controlled, methods of shutting down the pump. I haven't a clue as to how low a probability of malfunction/failure would be considered necessary for the application being discussed, but I suspect that some of the discussion/argument may, in context, be a little 'OTT'.

Kind Regards, John

 

[bernardgreen]

... Failsafe does not mean “safe from failure.” It means that the situation resulting from a failure is a safe one.

I can't find the article but it detailed how a critical valve in the Apollo spacecraft became 13 valves with actuators and feed back systems.

If it jammed shut a second valve in parallel was needed. But if one of these jammed open then valves 3 and 4 were needed, One in series with each of the first two..

........[1]........X........[3].........

........[2]........Y........[4].........

That took care of all single faults..

But if [2] and [3] jammed shut then a valve [5] between X and Y would allow flow [1] X [5] Y [4]

But valve [5] need to be reliable so had to have [6] [7] and [8] added.

The rest of the complex development has faded from memory.

The result was that the design of critical valves was improved to reduce the risk of valve failure low enough that the 4 valve configuration would suffice.

 

[SimonH2]

For this application, sucking is not an option.

Rambles off ... A few years ago I was at an event involving lots of Land Rovers. The portable toilets were being "serviced" with a setup on the bed of a Series IIB Forward Control - and of course the operators jot lots of jokes that they've probably heard many times before. Anyway, the portable "servicing" kit used vacuum and a long hose to empty the tank. After I'd quipped about someone having to do it, the guy came back (so quickly it's obvious it wasn't the first time) with "it's not so bad, but getting the syphon going can be a bit unpleasant" while gesturing in a manner that implies manually sucking on a pipe to start the syphon. How we laughed

But I digress.
The fluid in question is largely water, with various dissolved compounds (including plenty of urea). It also has a significant quantity of suspended solids - mostly organic matter, semi-digested grasses and grains. There may also be sand and small stones in suspension as well - and for good measure, occasionally a larger stone (it's been known to find a house brick in a tanker that's been sucked up by the hose but then settled in the bottom of the tanker) though these shouldn't get through the pump. The whole lot will be in a semi-fermented state - it has a very "distinctive" smell when disturbed , and it can take a bit of washing out if you get covered in it Don't go into farming if you object to a bit of muck !

So while sucking is an option when filling tankers (you suck the air out, the vacuum sucks the slurry up the hose), it's not an option to use a positive displacement pump which would be needed to continuously suck the slurry up a pipe. Also, the pipe for this application is considerably smaller, and can be quite long - so needs more than the one bar differential pressure achievable through suck alone.

So the setup will be a centrifugal pump of sufficiently coarse design as to cope with the suspended matter, and submerged in the slurry so as to have a positive head at the inlet. There probably isn't an inlet screen as this would get blocked - instead it will (ideally) be positioned so the house bricks will sink out of reach of the inlet flow (that's if they haven't just sunk to the bottom of the store and stayed there).

For filling "buggies" (open top tankers), the pump outlet just goes up a pipe like a giant kitchen sink tap spout that the buggy can be parked under. For the application here, it goes out a long hose across the fields to a tractor mounted unit. Some of these have a number of small plough like heads that put the slurry directly under the surface of the ground. Others just have a spray head that spreads it over a spread width of anything up to 15-30 feet - usually fairly narrow. This isn't a high pressure spray, it's more like if you put a finger half over the end of a hosepipe - not to create back pressure, but just to spread the flow into a fan shape.

Some systems drag the hose around, some pay it out from a big reel as the tractor drives along - if it's flat hose, then it has to be fully unreeled before it can be used. Either way, the hose can get to a fair length once you get a few fields away.
No, you aren't just going to tape a cable to a 4" hose that's full of water/suspension, drag it around a field, and expect it to last even 5 minutes.


So risks ?
Well the obvious one is that if a hose bursts or becomes disconnected, and the pump keeps running, then you dump a lot of slurry on one point. This isn't good for the ground (but it will recover), but there is the risk of run-off into a watercourse depending on the ground and proximity.
Personal damage from a burst or disconnection is unlikely - this isn't a high pressure, high flow rate system that will have the hose flailing around like some giant demented snake, or create high-pressure high-velocity spray. It'll sit on the ground with sh1t pouring out the end. Of course, personal contamination is a risk - but you don't need a burst for that, a change in wind direction can do it. As an aside, if you aren't familiar with it, have a listen to Blaster Bates's "Shower of sh1t over Shropshire" - it's just hilarious.

And, err, that's probably about it !

 

[stillp]

As an aside, if you aren't familiar with it, have a listen to Blaster Bates's "Shower of s**t over Shropshire" - it's just hilarious.

Having heard him perform that live, I second that!