PPT_Monday_EqualPoly_Western_Ryerson.pdf

Poly

1. 7
th
May 2018, 3.00PM Presentation for Experts
Optimal Feedback Control for
Enhanced Oil Recovery – Lab Scale
Debaprasad, Master Student, Chemcial Engineering, RU
Dr.Indiran Thirunavukkarasu – Visiting Professor, Manipal Inst. Of
Tech, India,
Dr.Simant Upreti – Professor & Chair, Chemical Engg,
Ryerson University, Canada
Transport Modelling Lab, KHE-133

2. 7
th
May 2018, 3.00PM Presentation for Experts
Heavy Oil – Definition
●As defined by the U.S. Geological Survey (USGS), heavy oil is a type of crude oil
characterized by an asphaltic, dense, viscous nature (similar to molasses)
● It also contains impurities such as waxes and carbon residue that must be removed
before being refined. Although variously defined, the upper limit for heavy oil is 22°
API gravity with a viscosity of 100 cp (centipoise).
How it forms
●originated with plant life millions of years ago. When the plants and small organisms
(plankton) that fed on them died off, the sediments containing their remains were
buried at the bottom of inland seas. In a highly simplified explanation, over time, the
heat and pressure converted the carbohydrates into hydrocarbons.

3. 7
th
May 2018, 3.00PM Presentation for Experts
Key Facts
1. Canada is the fourth largest producer and third largest exporter of
oil in the world
2. 97% of Canada’s proven oil reserves are located in the oil sands
3. 99% of Canada’s oil exports go to the U.S.
4. GHG emissions per barrel of oil produced in the oil sands have
fallen over 35% since 1990

4. 7
th
May 2018, 3.00PM Presentation for Experts
Interesting Facts
1. Canada in world production – 5%
2. Canada in world exports – 8%
3. Canada in world proved reserves – 10%

5. 7
th
May 2018, 3.00PM Presentation for Experts
Experimental Setup

6. 7
th
May 2018, 3.00PM Presentation for Experts
Objectives
1. To identify the mathematical model the lab scale enhanced oil
recovery system as MIMO process.
2. To simulate various control algorithms for the identified model.
3. To implement the optimal control with the lab scale model.
4. To maximize the recovery of heavy oil under optimal conditions.

7. 7
th
May 2018, 3.00PM Presentation for Experts
Journey of experimental setup
Before we start Heavy oil at barrelMeasuring vessel weightFilling heavy oil in vessel
Measuring mesh weight
Measuring oil weight
Glass beads filled in meshPreparing feed
Heating heavy oil 75 Dec CelFixing physical modelTrial run to get signals
Setup ready for open loop
modelling

8. 7
th
May 2018, 3.00PM Presentation for Experts
Open loop modelling
We aimed to model the system as Multivariate process
with 2 Inputs and 2 outputs
S. No Inputs Outputs
1 Heater
supply
(Watts)
MV1
Temp. Of
physical
model
(Deg. Cel)
PV1
2 Coolant
flow
rate(LPG)
MV2
Viscosity
of live oil
(CP)
PV2
[
T
ProcessModel
μ
LiveOil]
=[
K
p11
∗e
−t
d11
s
(τ
11+1)
K
p12
∗e
−t
d12
s
(τ
12+1)
K
p21∗e
−t
d21s
(τ
21+1)
K
p22∗e
−t
d22s
(τ
22+1)]
∗[
Q
Heater
F
CoolingWater]

9. 7
th
May 2018, 3.00PM Presentation for Experts
Closed loop control
Once after the modelling part, the closed loop control needs to be
implemented in simulation followed by the hardware in loop.
1. Classical PI/PID Control.
2. Necessary and sufficient conditions for optimal control.
3. Robust PID Control.