Technical Article

Conceptual Modeling of Power Electronics

June 23, 2021 by Anushree Ramanath

The second article in this series will focus on the process of modeling and simulating power electronics from the problem definition phase to the experimental phase.

In recent years, simulation models are becoming large and complicated. When every simulation model is designed from scratch, a considerable lack of reusability develops and leads to the process being both time-consuming and expensive. This demands the need to consider the reusability and fidelity aspects of modeling from the initial stages of development. Power electronics engineers rely on several modeling and simulation tools to develop effective power electronic systems. Read on to learn more about the process of modeling and simulation, along with its characteristics.

Figure 1. Power Electronics Modeling and Simulation Process. Image courtesy of National Instruments
Figure 1. Power Electronics Modeling and Simulation Process. Image courtesy of National Instruments


An Overview

Conceptual modeling is recognized to be a crucial step in the process of simulation. It is a top-down approach that involves partitioning the system into logical sub-systems and defining relevant relationships. However, while developing the simulation-based model, a bottom-up composition approach is adopted. A wide variety of methods are employed for representing models and conceptual simulation aspects. It includes graphs, activity diagrams, process flow diagrams, and so on. But there is no single representation or unanimously accepted conceptual modeling technique for simulation endeavors [1]. Large interconnected systems are simulated using component connection models wherein, the specifications of component characteristics and its interconnections can be separated for a circuit description. This helps create efficient simulation methodologies as the derived equations from such models are generally sparse in nature.

Simulation of the entire system, starting from the schematic stage to the production stage, is the ultimate goal. This involves extensive modeling efforts, long simulation times, and modular architecture. Reusing simulation modeling is preferable and involves the use of pre-existing modeling and code elements [2]. Improving the reuse of existing modeling methodologies requires development in state-of-the-art conceptual modeling, efforts, and simulation. 

The modeling and simulation lifecycle typically comprises a source system derived based on observation from reality followed by conceptual modeling that leads to the conceptual model and then the simulation model. Analysis of the system and simulation of the implemented model facilitates experimentation results. However, it is evident that the analysis, design, and practical implementation of power electronic systems without computer-based simulation is extremely laborious, time-consuming, and therefore an expensive endeavor.



There are several processes involved in the process of modeling and simulation. Based on the modeling and simulation lifecycle discussed, a component-based hierarchical integrated modeling and simulation framework is illustrated, as shown in Figure 2 [4]. It is typically composed of five layers and demonstrates the process of going from problem definition to the final evaluation stage as we flow through the integrated solution chain. Roles and processes are integrated with the flow along with possible supplementary processes in other workflows.

Figure 2: Sample framework for hierarchical modeling and simulation [4]
Figure 2: Sample framework for hierarchical modeling and simulation [3]

The problem definition phase involves the owner who defines the purpose of the modeling and simulation endeavor or problem statement and requirements based on specifications. This helps choose the appropriate boundary conditions and value factors or key performance indicators (KPIs) based on which the performance can be evaluated.

In conceptual modeling, a model designer creates the high-level abstraction of the real system based on the studied characteristics and prepares the conceptual model. In general, the details at this stage include brief text-based information along with graphical representations to facilitate documentation, knowledge transfer, and support model transformations. This stage helps bridge the owners and simulation experts.

The specification stage comprises the definition of specifications for the model that is independent of the platform. This facilitates the modeling of intended system functionality without relying on any specific platform requirements.

In the implementation phase, an executable platform-specific simulation model is developed. This facilitates the realization of the conceptual model while incorporating all the defined specifications in the previous phase. Typically, specification and implementation layers can be supported with component libraries such that each component used can be matched to its specific implementation. This can help cater to different platforms during the upcoming phases.

The experimentation phase involves running the implemented executable model such that it replicates real-world scenarios, and the data thus obtained can be utilized for analysis and comparison purposes. This can help understand the actual experimental results, verify the correctness of the model and validate the chosen platform for execution. In addition to these factors, it can serve as a good test for model fidelity while helping in enhancing the model’s credibility. The validation experiments can also provide good insight into the input or output behavior of the simulation model and map it with the intended goals that need to be achieved as part of the simulation study being carried out.


Key references:

  1. Dragan Maksimovic, Modeling, and Simulation of Power Electronic Converters, 2001. 
  2. Pidd, Reusing simulation components: simulation software and model reuse: a polemic, 2002.
  3. Deniz, Applying a model-driven approach to component-based modeling and simulation, 2010.