Skip to content

ElecSolver

Overview

ElecSolver formalizes electric systems as linear problems, suitable for both temporal and frequency-domain studies. It focuses on constructing the linear system representation, leaving the actual numerical resolution to the user.

This repository is not a general-purpose electrical system solver. Instead, it acts as a bridge between:

  • The graph-based description of an electric network
  • The corresponding sparse linear system to solve

In a very simple way, ElecSolver takes as an input the Resistances \(R\), Capacitances \(C\), Inductances \(L\), Mutuals \(M\), Current sources \(I\) and Voltage sources \(V\) along with the connectivity graph \(G\) of the system and outputs the linear system: matrix \(S\) and vector \(b\) to solve in order to get the solution of the electric problem.

\[f_{\text{ElecSolver}}(R,C,L,M,I,V,G) = (S,b)\]

It is then up to the user to choose the solver they want for solving the system \(Sx=b\) and to manage the time iterations if needed for temporal problems.

The main goal of ElecSolver is to provide a friendly Python interface for simulating and optimizing analog electric systems. While suitable for small circuit simulations, its strength lies in its scalability: it is able to build linear systems with millions of nodes and components.

Tip

ElecSolver has been designed with the following specifications in mind:

  • The time needed for building the linear system must be negligible compared to the time needed for solving it.
  • Handle natively inductive mutuals and resistive mutuals.
  • Handle as many coupled electric systems as needed.
  • Deal with lonely nodes and lonely edges in the electric graph when the problem is still well posed.
  • Allow backpropagation of gradients through the system for optimization purposes.

Warning

Non-linear components are not supported. You must manage event detection and system updates yourself.

How to install

Tip

Although not having any solver dependencies, any user will want to tackle the linear system it outputs. Therefore we recommend installing and using python-mumps along when it is available on your platform.

The recommended way of installing ElecSolver is spack since spack provides a MUMPS solver (and its python bindings python-mumps) optimized for your platform.

spack install py-elecsolver py-python-mumps

Using conda

If you do not have access to a compiler or want a faster installation with MUMPS solver we also provide a conda package since python-mumps is available in conda. This installation may come at the cost of performance. You can install it with:

conda install elecsolver python-mumps

Using pip (if you must)

For convenience, ElecSolver is distributed on PyPI and can be installed with pip. In this case we recommend using scipy's sparse solvers since MUMPS is not available via PyPI. No need to make any special installation for this, since scipy is a dependency of ElecSolver.

pip install ElecSolver

Components

The documentation presents the two main components of ElecSolver:

Then presents extra features: