Re: PID vs. temp stability question

papenfussDIESPAM@xxxxxxxxxxxxxxxxxxxx wrote:
So... back to the original idea (which doesn't seem to be discussed much in the archives)... feed-FORWARD control.

As you've seen, feedforward and feed-forward come up a few times in Google search, but not often. There are at least two reasons why you aren't getting a lot of hits:
1. Implementing a robust FF scheme is expensive and/or a PITA for people who aren't EEs.
2. There are plenty of folks here who use manual feedforward schemes without even knowing the word; they turn on the steam switch for x seconds at the beginning of each shot. I tried this for a while, but it didn't seem to improve my shots.

> The introduction of cold water to the boiler is a disturbance, but the
> system it is being applied to is relatively constant. If
> one eliminates the variables of flow rate (puck tamping and grind), initial > boiler temperature (with feedback control such as a PID), and inlet water
> temperature (keep ambient room temperature constant), linear feedforward
> should be able to compensate for the disturbance on the (now time invariant) > plant.

FYI, here's another bit of information: Notwithstanding a consistent grind and tamp, it's an oversimplification to say that flow rate is non-variable. My machine feeds about 30ml of water into the boiler in the first 5 seconds. This wets the puck and fills the air space. At that flow rate, it would take somewhere around 1800 watts to maintain temperature.

Then, for the next 20-25 seconds, it feeds ~25ml of water to complete the extraction. That flow rate corresponds to about 300-400 watts.

I'm sure you will check these figures to see if my math is in the ballpark.

> Like I said, when I get a spare boiler (well, thermoblock in this case) to
> play with, I'll take some measurements and see what I can come up with.

I use a thermoblock upstream of the boiler that's under PID (feedback) control. It does a fair job. If you took the time, you could certainly add a flowmeter to the loop and manage the feedforward situation to account for the variable flow rate as outlined above.

A simpler solution for a non-professional like me would be to replace the thermoblock with a small preheat boiler. This would probably do 90% of what a fancy feedforward loop would accomplish. I haven't gotten around to it.

As I'm sure you aware, some folks with dual boilers preheat their feedwater by running it through a coil wrapped around the steam boiler. The Brewtus (dual boiler) runs the feedwater through a heat exchanger in the steam boiler.

Many of us here would be very interested in hearing how your experiments come out.
--


-Andy S.
 picture page:  http://tinyurl.com/eh0x
.

Relevant Pages

  • Re: Simple boiler control for off-grid?
    ... LED/Phototransistor pairs at each temperature setpoint, ... the original thermostat well. ... > Managed to install an olive kernel boiler last week and need to pick your ... > collective minds regarding temp control. ...
    (sci.engr.control)
  • Re: fashioning temperature probes for high pressure?
    ... it is the boiler temperature that you want to ... A probe in the brewhead is nice to monitor temperature, but to control ... keeps the brewhead at a stable temperasture separate from the brew ...
    (alt.coffee)
  • Re: Random Walk Up Shots with a PIDd Heat Exchanger
    ... You are not trying to control to ... inexpensive boiler machine. ... The group head it decoupled from the boiler. ... loop of brew water driving by a continuous pump. ...
    (alt.coffee)
  • Re: Interesting
    ... The boiler section is long proven nsafe and efficient, ... The burner and stoker and the control systems are ... and the MSM puts fourteen fact-cherckers ... you have your facts wrong. ...
    (rec.crafts.metalworking)
  • Re: PID controller
    ... needs improvement by using PID controller in tuning a boiler ... Are you asking if there are shortcomings to using a straight PID ... Need to learn how to apply control theory in your embedded system? ...
    (sci.engr.control)