Design Optimization For Power-Supplying The Nodes of WSNs - Master Thesis Presentation - Ioannis Mandourarakis

Ioannis Mandourarakis
Master Thesis Presentation
entitled
“Design Optimization of an Electric Energy Production

System for Power-Supplying
the Nodes of Wireless Sensor Networks”
1
Outline
• Problem Statement & Solution Approach
• System & Circuit Modeling
• Circuit Simulation & Results
• Hardware Implementation
(Experimental Tests, Prototypes, Calibration, Real-Time algorithms)
• Robustness Tests
• Wind Generator & MPPT
• System Simulation & Results
(Experimental Tests, Circuit Comparison, Overview)
• Conclusion, Submitted Work & Installation-site Photos
2
Problem Statement (Generalized Block Diagram of Novelty : No study has been carried out on stand alone RES system
focused on low power consumption in WSNs
the RES-based energy management system of WSNs)
by considering annual time-series profiles
Eb (t) = Eb (t - 1)+ nb  Pb (t)  Δt

3
Problem Statement (Sizing Optimization on System)
Eb (t) = Eb (t - 1)+ nb  Pb (t)  Δt

4
Problem Statement (Sizing Optimization on Circuit)
Eb (t) = Eb (t - 1)+ nb  Pb (t)  Δt

5
Solution Approach
Optimization Using Genetic Algorithms for both the system and circuit level design
System Circuit
Design Design
Variables Variables
1. Number and 1. Inductor
Type of PVs 2. Output Filter
or W/G Capacitor Contraints
2. PV angle or Contraints 3. # MOSFETs
W/G height 4. Switching 1. Enter DCM
4. Number and
1. Minimum Frequency 2. Maximum
Type of Batts
5. Number and SoC Current
Type of (given a DoD) Ratings
SuperCaps 2. Batt Energy 3. I Ripple
> 50 % 4. V Ripple
Outline
7
System Modeling Stochastic time-series as part of the data fed into the GA
8
System Modeling
Spreadsheet containing the operational data for the system GA feed
9
Circuit Modeling (Equivalent Circuit & Power Loss Equations)
10
Outline
11
Simulations Total DC-DC boost converter power loss for
various values of Inductance and operating frequency
12
Simulations
Parametric iso-L curves for the cases of 1 to 4 parallel MOSFETs
13
Simulations
Analysis of the power losses on the electronic parts
Simulations
(b) A zoomed-in part of plot
in (a), demonstrating the
fluctuation of the overall
power loss and its local
minimum at 31kHz
(a) Power loss distribution

among the dominant
(in terms of power-loss)
components for various
switching frequency values
15
Simulations
Distribution of power losses in individual
components of the Boost-type DC-DC converter
power MOSFET Inductor
16
Simulations Flags
Constraints (due to component fail tolerances and load regulation specifications)
(Incident Sensing)
1 The inductor makes the circuit get into DCM
2 MOSFET reached half of its maximum current ratings
3 Output ripple current too high
4 Output ripple voltage too high
17
Circuit P Po The optimal values of the design variables.
n= o =
Simulation PPV Po + PL
Results 8760
=
EL ,tot ∑P
i =1
L ,i ⋅ ∆t
Cm Cm
= =
LCOE 8760
∑ (P
Etot
pv ,i − PL ,i ) ⋅ ∆t
i =1
where
PPV is the power produced by the PV source
under MPPT conditions The DC-DC converter performance for the alternative objective functions,
which have been employed in the proposed optimization process.
Po is the output power
PL recursively by Ppv = PL + Vo ⋅ I o
PL ,i is the total power loss of converter at hour i
Ppv ,i is the output power of the PV source at hour i
.
Cm is the total manufacturing cost of the converter
Etot is the total energy transferred to the battery
bank and electric load during the year
Outline
20
Towards Hardware Implementation A schematic diagram of the power sub-circuit.
21
Towards Hardware Implementation
Block diagram illustrating the connection of the microcontroller,
the sensors and the converter’s output control circuit.
22
1st Experimental Tests
25
During
Calibration
26
Real Time
Algorithms
27
Real Time Algorithms MPPT performance in relation to duty cycle stepping
Real Time Algorithms Screenshot of the interface developed for monitoring and debugging the DC-DC converter’s behavior.
The maximum power efficiency achieved was over 96 %
meeting the performance directive of the project’s specifications.
29
Outline
30
Robustness Test
ICL7667 chip power diode
Thermal
Imaging
the wiring
the electric load
connectors
Robustness Test 4-days Continuous Operation
Voltage variation of the batteries for four days under moderate cloudiness.
Outline
33
Wind Generator MPPT Experiments
Wind Generator MPPT Experiments
The MPP is successfully

reached for various fixed
drill torque values. Τhis
experiment depicts 3 cases.
Outline
36
System Simulation PV & WG produced power and Stored Energy Levels in Battery
37
The operational characteristics of the RES-based power-supply system units.
System Simulation Results
in the physical system too)

(Implemented as such
The optimal values of the design variables.
38
Circuit Comparison Experimentally Measured Optimized and non-optimized
(Validation) DC-DC converters’ time-series under MPPT conditions.
39
Circuit Comparison
(Validation)
Comparison of the optimized and non-optimized DC-DC
converters in terms of the energy conversion efficiency.
40
Overview
The experimentally measured time-series of the power produced by the PV
source, as well as the voltage and input/output power of the battery bank.
41
Outline
42
Conclusion
Problem : Maintaining Robustness while finding the cheapest solution
for the implementation of an energy management system
customized for wireless network sensor nodes
Proposed Method : Sizing Optimization based on Genetic Algorithms

Based on Both a System and Circuit Level
Customized for certain areas & Load Demand Profiles
Novelty : No study has been carried out on stand alone RES system focused on
low power consumption in WSNs by considering annual time-series profiles
Results : RES based power supply cost reduced by 15.7 %

DC-DC boost converter’s efficiency increased by 3.6 %
Future Work : - Extended experiments focused on the system design

- Holistic Approach combining the system and circuit results
- Dynamic Programming (solar tracking incl. expenses)
- SMD / industrial design / customer-centric solutions
43
Submitted Work
• I. Mandourarakis and E. Koutroulis,

“Optimal Design of a Boost-type DC-DC Converter
for PV Power‐Supplied Wireless Sensor Networks”
Accepted for presentation in
IEEE International Conference on Industrial Technology,
Cevilla ICIT 2015
• I. Mandourarakis and E. Koutroulis

“Design Optimization of a RES-based Power-Supply
System for Wireless Sensor Networks”
Submitted for review in
IEEE International Conference on Environment and Electrical Engineering,
Rome EEEIC 2015
44
Photos
RevE Prototype
45
Photos (Comparison)
Photos (Comparison)
47
Photos (1st stand experiment)
48
Photos (Agios Ioannis – outside the tank)
Photos (Agios Ioannis – inside the tank)
Photos (Agios Mathaios – outside the tank)
51
Photos (Agios Mathaios – outside the tank)
52
Photos (Meeting in Cultural Center & Prototypes)
53
Thank you!!!
Any questions please ?

Design Optimization For Power-Supplying The Nodes of WSNs - Master Thesis Presentation - Ioannis Mandourarakis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Design Optimization For Power-Supplying The Nodes of WSNs - Master Thesis Presentation - Ioannis Mandourarakis

Uploaded by

Copyright:

Available Formats

Ioannis Mandourarakis

Master Thesis Presentation

“Design Optimization of an Electric Energy Production

Eb (t) = Eb (t - 1)+ nb  Pb (t)  Δt

Eb (t) = Eb (t - 1)+ nb  Pb (t)  Δt

Eb (t) = Eb (t - 1)+ nb  Pb (t)  Δt

(a) Power loss distribution

power MOSFET Inductor

ICL7667 chip power diode

The MPP is successfully

in the physical system too)

The optimal values of the design variables.

Proposed Method : Sizing Optimization based on Genetic Algorithms

Results : RES based power supply cost reduced by 15.7 %

Future Work : - Extended experiments focused on the system design

• I. Mandourarakis and E. Koutroulis,

• I. Mandourarakis and E. Koutroulis

You might also like