/_TINY_ACMSIMC

AC Induction Motor Simulation in Python

Primary LanguageCGNU General Public License v3.0GPL-3.0

AC Induction Motor Simulation in Python

Overview

❯ python .\main.py
Validate: 0.02486440087288949 = 0.024864400872889467 ?

Kp_omg=30, Ki_omg=0.08943749999999999

Kp_cur=20, Ki_cur=0.005

Simulation Run Time=15.0 s
Simulation Costs Time=4.840036630630493 s

Screenshot

Introduction

This project aims to simulate AC induction motor in Python.

  • Forked from https://github.com/horychen/ACMSIMC_TUT

  • Thanks to Jiahao Chen | 陈嘉豪

  • GPL-3.0 licensed

  • Code structure

    • Main Loop
      • Main : Simulation Main Loop
      • ACM : AC Induction Motor Class
      • CTRL : Controller Class
      • Observer : Observer Class
      • Utils : All Utils imports and constants
      • FileIO : File Input and Output class
      • Macros : All lambda functions
      • ACMPlot : Draw Treand plot
      • ACMPlot2 : Draw Trend plot in Plotly
    • Addition Files
      • B-H_Cruve : Draw B-H curve from 33.txt
      • LZ4-example : Test LZ4 and Zstandard performance
      • ResultPrint : Print result table in tabulate
      • rk4 : Runge-Kutta 4th ODE solver in C shared library

  • 6th-Derivative-Of-Position (for motion control)

  • Prerequisites

  • HOWTO

    • Open your powershell in windows

    • Change work path to a directory you like

    • Run code below line-by-line

      mkdir ACMSIM
cd ACMSIM
git clone https://github.com/tinyko/_TINY_ACMSIMC.git .
./make.ps1
python main.py

    • Wait for seconds until algorithm.dat and info.dat file are generated

    • Then matplotlib.pyplot is shown figures

Development plans

  • Visualized Configuration of AC Induction Motor
    • Stator and rotor Laminations Design
    • Coil Winding Calculations
    • Skewed Rotor bar
    • Materials Difference
    • Losses
      • Mechanical Losses
      • Iron Losses
      • Copper Losses
      • Wind friction Losses
    • Deep bar effects , Skin effect and the Eddy currents
  • GUI in PyQT5 with Trend Viewer , Plots Evaluating ,.etc
  • Full No Load Test and More Detailed Motor Model
  • Control Schemes
    • V/F Scalar Control
    • Vector Control
      • DTC(Direct Torque Control)
        • DSC(Direct Self-control)
        • SVM(Space Vector Modulation)
      • FOC(Field Oriented Cowntrol)
        • Direct FOC
        • Indirect FOC
    • Sensorless Control
    • Magnetic-field Weakening
    • (Un)Coiler tension control
    • Two-mass drives issue
    • Ramp
      • Linear Reference
      • S-curve
      • 6th-Derivative-of-Position
  • Stable And Transient Motor Model Analysis
  • AC-AC(DC-AC) converter Topology
    • Converter with DC-link
      • (regenerative) voltage-source
      • current-source
    • Matrix Converter
      • the Direct Matrix
      • the indirect matrix
    • Hybrid Matrix

Develop Notes (Tiny 2020.02.16)

TO DO List

  1. Separate Programm Functions into different python modules
  2. GUI in PyQT5 (Auto Scale GUI, QT GUI layout) / PySimpleGui
  3. Async communications between Real-time and GUI module
  4. Design a python to c source translator(like simulink) (further plan)
  5. motor topology
  6. Add json config file to manage parameters
  7. create a Trend Viewer to evaluate trends data , which could have some functions to calculate,zoom in/out,.etc(like IBA©analyzer)
  8. Save all critical data in trends
  9. Create a speical HDF5 data file structure to contain data informations(timebase,timestamp,.etc) and values(bool,string,int,float,.etc)
  10. Remove all redundant code and cross-reference
  11. Separate ODE-solver to a single module , Current Method is Runge-Kutta 4th order Method
    • Methods
      • Implicit(Explicit) Runge-Kutta Methods
      • Adams–Bashforth methods
      • Adams–Moulton methods
      • Gear's Backward differentiation formulas (BDF)
      • Euler Forawrd/Backward Method
      • Kaps-Rentrop Methods / Rosenbrock methods
      • Gauss-Jackson Method
      • Predictor-Corrector Methods
      • Relaxation Methods
    • Issues
      • Stiff problems
      • Stability and convergence
  12. Decoupling sub-modules
  13. cProfile found out some slow code in main loop in cProfile.ps1 file:
   cprofilev -f .\log.txt

Done List

  1. Converted C program into Python Single File "Alter"
  2. Modified ACMPlot as a library to Main Python Script
  3. Compared LZ4 and Zstandard (Winner is LZ4! In Realtime scenanio)
  4. numpy ndarray could be converted into bytes , vice versa
  5. Add a 6th-order speed ref curve
  6. Add _slot_ to Class object

Benchmarks

2020-02-16 Python 3.8.1 TotalTime:22.796 seconds

Screenshot

2020-02-21 Python 3.8.1 TotalTime:18.026 seconds

Screenshot

2020-03-05 Python 3.8.1 TotalTime:6.308 seconds

Great Performance Achived Screenshot