Test Suite Problem 2.5 - Description

Power Converter Problem


Problem Size

Problem Definition

Solution Strategy

Source Code Availability


Power Converter Page

MDO Test Suite Home Page


The power converter problem can be considered a multidisciplinary problem comprising the coupling between an electrical subsystem and a loss subsystem. An optimal power stage design is essential to the development of quality power converter. The power stage design dominates the overall efficiency, size, and weight of the power converter. A schematic of the power stage of the power converter is shown in the following figure. Power Stage Schematic The geometry of the transformer core is shown in the following figure. Transformer Core Representation

The objective of the problem is to minimize the weight and yet maintain a high efficiency for the power stage subject to several constraints.

Problem Size: (see definition of terms)

Problem Definition:

The objective of the power converter problem is to minimize the weight. The problem consists of six design variables and twelve state variables of which four define constraints. The design variables and their lower bounds are: Description of variables

The state variables are described by the following equations: Definition of variables
Definition of constants

There are several constants which complete the description of the power converter.
EIInput Voltage (nominal)3.25 e2
EIMINInput Voltage (minimum)2.25 e2
EIMAXInput Voltage (maximum)4.25 e2
EOOutput Voltage5.0
POOutput Power5.0 e2
POMINOutput Power (minimum)0.5 e2
VROutput Ripple Spec.5.0 e-2
K1Aspect Ratio, center leg depth/width1.0
K2Aspect Ratio, window height/width2.0
XNTransformer Turns Ratio16
PXFRTransformer Related Losses0.0
FRSwitching Ripple Frequency0.1 e6
FCWinding Pitch Factor1.9
FWWindow Fill Factor0.4
WBOBBobbin Thickness2.0 e-3
BSPMaximum Flux Density0.3
DICore Density0.78 e4
DCCopper Density0.89 e4
DK5Capacitor Density25.0
KHHeat Sink Density88.0
ROCopper Resistivity1.724 e-8
RCKESR Time Constant0.3
CKESR Time Constant0.1 e-3
VDDiode Conduction Drop0.65
TNDDiode Turn-on Time1.0 e-7
TFDDiode Turn-off Time1.0 e-7
TREDiode Reverse Recovery Time0.5 e-7
VSTTransistor Saturation Drop0.
VBETransistor Base-Emitter Drop0.
GAINTransistor Current Gain0.
TSRTransistor Turn-on Rise Time1.0 e-7
TSFTransistor Turn-off Fall Time1.0 e-7
RDSMOSFET On Resistance0.5
CGSMOSFET Gate-Source Capacitance8.0 e-9
COSSMOSFET Output Capacitance4.0 e-10
VGSMOSFET Gate-Source Voltage10.

Solution Strategies:

Solution using CONMIN with Quasi-Analytic Gradients
The problem was solved using the method of feasible directions and a line search method as implemented by the CONMIN code. Gradient information was obtained by automatic differentiation of the system analysis code using ADIFOR.

Send us your method or comments about this one.
Describe your strategy and computational experience (in HTML form preferably). It will be included here. Send code if you'd like.

Source Code Availability:


Power Converter Page