Team:KULeuven/Design/Integrated Model

From 2009.igem.org

(Difference between revisions)
(Robustness)
(Proportional design (P controller))
 
(44 intermediate revisions not shown)
Line 1: Line 1:
{{Team:KULeuven/Common2/BeginHeader}}
{{Team:KULeuven/Common2/BeginHeader}}
-
{{Team:KULeuven/Common/SubMenu_Project}}
+
{{Team:KULeuven/Common/SubMenu_Intergated_Model}}
{{Team:KULeuven/Common2/EndHeader}}
{{Team:KULeuven/Common2/EndHeader}}
__NOTOC__
__NOTOC__
-
=Full model=
+
=Controller design=
-
The complete model of our vanillin producing bacteria is shown in the next figure. The boxes around some species have now biological meaning
+
Because we want to optimize the design of the feedback loop in our system, we developed a more abstract block scheme of the bacterium. It shows each component as a block performing a specific task. The diagram is used to develop some theories about the performance of the feedback loop.
-
they merely serve to distinguish between the different subcomponents of our system.
+
-
[[Image:Fullmodel.jpg|750px|center|thumb|Biological model of our system]]
+
==Proportional design (P controller)==
-
Because we want to optimize the design of the feedback loop in our system, we developed a block scheme of the  
+
Because the controller has to be implemented in 'biological technology', we choose the simplest possible design of  
-
bacteria. Which can be used to develop some theories about its performance.
+
controller, the proportional controller. The output signal in this type of controller is directly proportional to the error signal. The error signal is the substraction of the input and the control signal. however this type of controller has one importent flaw. When the input is a step function, as in most cases, there will be a steady state error. To make this steady state error as small as possible, the gain in the feedback loop must be as large as possible.
 +
The gain in the feedback loop can be adjusted by using low/high copy plasmids for the genes in the loop.
-
[[Image:Blockmodel.jpg|750px|center|thumb|Block model of the system]]
+
[[Image:Proportional10.JPG|750px|center|thumb|Block model of the system with proportional controller (Simulink)]]
-
=Control theory=
+
==Proportional and Integral design (PI controller)==
-
Control theory is an interdisciplinary branch of engineering and mathematics, that deals with the behavior of dynamical systems. The purpose is to design a controller who controls the system so that it behaves as wanted. There exists several criteria to measure the performance of the controller.
+
Because the steady state error on a step input signal will always be nonzero without an integrator, we considered to add some integral action, in the form of an proportional and integral controller. The input of the controlled system is, in this type of controller, a weighted sum of the error and the integral of that error signal. This way the staedy state error is avoided. However because of the limited amount of time and resources we stayed with the P controller design, which is more straightforward to implement in a biological systems.<br>
-
 
+
As an extention the integrator could be implemented in the cell by producing a species with a rate proportional to the amount of mRNA key.
-
==Stability==
+
[[Image:PI.JPG|750px|center|thumb|Block model of the system with proportional controller (Simulink)]]
-
 
+
-
Stability means that no matter what the input signal (the blue light) is, the output (vanillin concentration) will remain finite after an infinite amount of time.
+
-
 
+
-
==Tracking problem==
+
-
 
+
-
This criterium is an indication of how well the output well follow the wanted reference system, we want the difference between the output and the wanted reference signal as small as possible.
+
-
 
+
-
==Disturbance rejection==
+
-
 
+
-
Most controlled system are disturbed by other systems in their neighbour hood, in our case imagine someone adding a extra amount vanillin to the aqueous medium. We do not want to see these disturbances in our output of vanillin.
+
-
The disturbance rejection criteria indicates the ability of the system to reject those inputs.
+
-
 
+
-
==Robustness==
+
-
 
+
-
As with every model, our model of the bacteria is not perfect. Robustness is an property of a property, we say that the stabilizes the system in a robust way if also stabilizes all systems that are similar to the modelled system. It's then assumed that the controller will also stabilize the real system as it is also assumed to be similar to the modelled system.
+
-
 
+
-
=Biological implications=
+
{{Team:KULeuven/Common2/PageFooter}}
{{Team:KULeuven/Common2/PageFooter}}

Latest revision as of 19:06, 13 October 2009


Controller design

Because we want to optimize the design of the feedback loop in our system, we developed a more abstract block scheme of the bacterium. It shows each component as a block performing a specific task. The diagram is used to develop some theories about the performance of the feedback loop.

Proportional design (P controller)

Because the controller has to be implemented in 'biological technology', we choose the simplest possible design of controller, the proportional controller. The output signal in this type of controller is directly proportional to the error signal. The error signal is the substraction of the input and the control signal. however this type of controller has one importent flaw. When the input is a step function, as in most cases, there will be a steady state error. To make this steady state error as small as possible, the gain in the feedback loop must be as large as possible. The gain in the feedback loop can be adjusted by using low/high copy plasmids for the genes in the loop.

Block model of the system with proportional controller (Simulink)

Proportional and Integral design (PI controller)

Because the steady state error on a step input signal will always be nonzero without an integrator, we considered to add some integral action, in the form of an proportional and integral controller. The input of the controlled system is, in this type of controller, a weighted sum of the error and the integral of that error signal. This way the staedy state error is avoided. However because of the limited amount of time and resources we stayed with the P controller design, which is more straightforward to implement in a biological systems.
As an extention the integrator could be implemented in the cell by producing a species with a rate proportional to the amount of mRNA key.

Block model of the system with proportional controller (Simulink)