Thermal Work - Matlab -- Write up
So I worked with Jeff on the matlab thermal work, since I had missed a fair amount during thesis time. WE managed to finish up everything. Umm ok here goes:
So I worked up the part 1 of the thermal challenge in class with Robbie. The cooling and heating of coffee was really easy. But hey, I've done tons and TONS of matlab coding this semester (aka > 200 pages of matlab codes). Any who, what happens with bang bang is lots of oscillation for the control, aka the temperature of the coffee oscillates around the desired temperature. When you code in proportional control, you have trouble reaching the desired temperature. Instead the coffee tends to be right below the desired temperature and never quite reaches the desired temp. I guess Starbucks can't use proportional control then!
It is quite fun to add in a time delay for the proportional control and bang-bang control. With bang-bang you see a fair amount of overshoot. You see some overshoot with proportional control too. and you tend to have it finally reach the desired temperature for a bit of time.
Part II was fun. I wired up the car with the thermistor and ran the test thermal script. My thermistor got much hotter than most - aka 400. So Jeff and I had fun calculating the variables. My rth = 24, my final temp was 400, my thermal capacity was 4.0625 and my initial temp was 304. Ok I did all of the simulations first. They went pretty well, bang bang was really standard. It oscillates around the final temperature a bit. I have a .fig file in the drop folder of all simulations and reality ones that I did.
Basically I used the same while loop in every single simulation and just changed the way power was set based on bang-bang, proportional control, proportional control + delay, and integral and proportional control. The outputs were as expected -- aka proportional control does not reach the desired temperature and with delay you get more of an oscillation. Also - integral control allows for the desired temperature to be reached easily.
I then ran the tests using the pretty much the same for loop that just changed the set power function depending on which control was used. They all worked...VERY SLOWLY (real world is annoyingly slow). But if I let my car alone and worked on my firepop (our name for the fire-robot cause pop == soda and we use diet coke soda to put out the fire). So all done -- all of my code is submitted with figures of each simulation and test.
So I worked up the part 1 of the thermal challenge in class with Robbie. The cooling and heating of coffee was really easy. But hey, I've done tons and TONS of matlab coding this semester (aka > 200 pages of matlab codes). Any who, what happens with bang bang is lots of oscillation for the control, aka the temperature of the coffee oscillates around the desired temperature. When you code in proportional control, you have trouble reaching the desired temperature. Instead the coffee tends to be right below the desired temperature and never quite reaches the desired temp. I guess Starbucks can't use proportional control then!
It is quite fun to add in a time delay for the proportional control and bang-bang control. With bang-bang you see a fair amount of overshoot. You see some overshoot with proportional control too. and you tend to have it finally reach the desired temperature for a bit of time.
Part II was fun. I wired up the car with the thermistor and ran the test thermal script. My thermistor got much hotter than most - aka 400. So Jeff and I had fun calculating the variables. My rth = 24, my final temp was 400, my thermal capacity was 4.0625 and my initial temp was 304. Ok I did all of the simulations first. They went pretty well, bang bang was really standard. It oscillates around the final temperature a bit. I have a .fig file in the drop folder of all simulations and reality ones that I did.
Basically I used the same while loop in every single simulation and just changed the way power was set based on bang-bang, proportional control, proportional control + delay, and integral and proportional control. The outputs were as expected -- aka proportional control does not reach the desired temperature and with delay you get more of an oscillation. Also - integral control allows for the desired temperature to be reached easily.
I then ran the tests using the pretty much the same for loop that just changed the set power function depending on which control was used. They all worked...VERY SLOWLY (real world is annoyingly slow). But if I let my car alone and worked on my firepop (our name for the fire-robot cause pop == soda and we use diet coke soda to put out the fire). So all done -- all of my code is submitted with figures of each simulation and test.
Comments
Post a Comment