READ ME

========================
General information
========================

Author: J M Hale
Contact: jack.hale@ncl.ac.uk
DOI:
License:
Last updated:
Related article:

========================
Introductory information
========================
The files relate to an experiment to determine the effectiveness of a 
"hardware-in-the-loop" active vibration control (AVC) system.  There is no mathematical 
model of the structure; the physical structure is fitted with an accelerometer sensor 
and an electromechanical actuator to provide input and output to/from the controller.

The sytem comprised:
- a structure (a cantilever beam) fitted with with vibration sensor and actuator;
- an Arduino Due microcontroller used as a real time controller;
- a PC running a Matlab program to supervise the Arduino.

The controller implemented in the Arduino is a tapped delay line (TDL).  Each tap has a 
coefficient, provided by a supervising PC via a USB link.  The objective of this work 
was to optimise the values of these taps to reduce the vibration of the structure which 
was subject to forced vibration in two resonant modes simultaneously.

The optimisation was performed by a genetic algorithm (GA) running on the PC.  This 
generated sets of TDL coefficients to send to the Arduino and used the vibration 
amplitude returned to it to evaluate the "fitness" of each set.  From this it gradually 
"bred" increasingly effective sets of TDL coefficients.

The results obtained with this system applied to a cantilever beam are contained in the
Excel spreadsheet.


Files included:

1) GenAlgor4_1_5Arduino.m
A Matlab GA program running on a PC, used to optimise TDL coefficients for active 
vibration control.

2) DueMatlabControllerV6_2.ino
An Arduino sketch (program) for an Arduino Due used as a microcontroller for active 
vibration control.  The control function is a tapped delay line (TDL: a type of finite 
impulse response digital filter) where the tap coefficients are provided by the GA.  
Real-time control is obtained by the controller reading instantaneous vibration from the 
sensor (accelerometer), from which the TDL generates a drive voltage for the actuator.

3) run 3 data_250.xls
An Excel spreadsheet containing the raw data generated by the GA over 250 generations as 
the optimised TDL coefficients evolved.

========================== 
Methodological information
==========================
The spreadsheet contains raw data from a run of 250 generations of the GA optimising 
controller coefficients for active vibration control (AVC).  The run in this file was 
for a population of 50 individuals, each of 60 x 4 bit genes.  The "genes" form the TDL
coefficients: integers 0-15.
Breeding was by elite selection with double point crossover and 1% mutation.

The first six rows contain the GA set up:
  row 1: population size
  row 2: number of genes per chromosome (population member)
  row 3: number of bits per gene
  row 4: breeding selection type (1=elite, 2=roulette)
  row 5: breeding crossover type (1=single point, 2=double point or circular)
  row 6: mutation rate
  
Rows 7 onwards give population genes and AVC results in each generation, sorted in order 
of performance index (PI) = fitness - cost
1st generation and thereafter every 10th generation are given for the whole population. 
Other generations’ results are given for the best, middle and worst individuals only.
Each row gives the value of each gene in columns A-BH for 60 genes.  This is  followed 
by three columns containing the AVC results: 
  column BI: performance index = fitness - cost
  column BJ: vibration amplitude returned from Arduino
  column BK: cost
     fitness = amplitude reduction
     amplitude reduction = uncontrolled - AVC amplitude 
     cost = sum of delay line coefficients × a constant

Instruments, hardware and software used: 
 PC running Matlab for the GA program
 Arduino Due for the control program
 elactromechanical shakers: Ling Dynamics type LDS V101 for driving the structure
 accelerometer: Analog Devices type ADXL335 for feedback


Date of data collection: 
20th March 2019