AI-based Low Computational Power Actuator/ Sensor Fault Detection - - PowerPoint PPT Presentation

▶
ai based low computational power actuator sensor fault
SMART_READER_LITE
LIVE PREVIEW

AI-based Low Computational Power Actuator/ Sensor Fault Detection - - PowerPoint PPT Presentation

Apr 29, 2023 700 likes •952 views

AI-based Low Computational Power Actuator/ Sensor Fault Detection Applied on a MAGLEV Suspension K. Michail, K. Deliparaschos, A. Zolotas, S. G. Tzafestas 21 st Mediterranean Conference on Control and Automation Crete, Greece, 25-28 June 2013

slide-1
SLIDE 1

AI-based Low Computational Power Actuator/ Sensor Fault Detection Applied on a MAGLEV Suspension

  • K. Michail, K. Deliparaschos, A. Zolotas, S. G. Tzafestas

21st Mediterranean Conference on Control and Automation Crete, Greece, 25-28 June 2013

slide-2
SLIDE 2

Ø Introduction Ø The proposed sensor fault detection scheme Ø Training the fault detection unit Ø The case study: A maglev suspension Ø Simulations Ø Conclusions

CONTENTS

MED’13 Control Conference Crete, Greece

slide-3
SLIDE 3

System

Uncertainties

A Typical System

Faults Faults

Disturbances Non-linearities Inherently Unstable

propose FTC scheme for sensor fault detection with low computational cost

MED’13 Control Conference Crete, Greece

slide-4
SLIDE 4

Sensor/actuator signals estimation

iFD unit (NN-based) Actuators Sensors (Driving Signals) (Measurements) Typical bank of estimators for sensor FD

MED’13 Control Conference Crete, Greece

slide-5
SLIDE 5

Ø Introduction Ø The proposed sensor fault detection scheme Ø Training the fault detection unit Ø The case stud: A maglev suspension Ø Simulations Ø Conclusions

CONTENTS

MED’13 Control Conference Crete, Greece

slide-6
SLIDE 6

Proposed Fault Tolerant Scheme

BS

MED’13 Control Conference Crete, Greece

slide-7
SLIDE 7

Ø Introduction Ø The proposed sensor fault detection scheme Ø Training the fault detection unit Ø The case study: A maglev suspension Ø Simulations Ø Conclusions

CONTENTS

MED’13 Control Conference Crete, Greece

slide-8
SLIDE 8

Neural Network: Training data set of the iFD

Measured signals

Estimated signals

Sensors

MED’13 Control Conference Crete, Greece

slide-9
SLIDE 9

Neural Network architecture

Driving Signal Input Layers Hidden Layers Output Layers

MED’13 Control Conference Crete, Greece

slide-10
SLIDE 10

Ø Introduction Ø The proposed sensor fault detection scheme Ø Training the fault detection unit Ø The case study: A maglev suspension Ø Simulations Ø Conclusions

CONTENTS

MED’13 Control Conference Crete, Greece

slide-11
SLIDE 11

Suspended mass (m)

Mg F

Track

Flux circulation Airgap

The test case: Maglev System

Electromagnet Pole

  • Vert. Accleration
  • Vert. Velocity

Current Power Amplifier

Driving Signal

K

Controller

EMS serves two purposes: Ø Support the vehicle and passengers Ø Ensure proper ride quality

MED’13 Control Conference Crete, Greece

slide-12
SLIDE 12

Input disturbances and performance requirements

Deterministic Stochastic Performance requirements

Aim: Maintain the Control performance within limits under deterministic and stochastic disturbances while ensuring the control performance in the presence

  • f sensor faults.

MED’13 Control Conference Crete, Greece

slide-13
SLIDE 13

A FTC scheme for sensor fault detection of the EMS

Sensor Fault Accommodation Recovery

EMS

  • 1. Single or Multiple Sensor Occurs
  • 2. Fault is Detected
  • 3. Faulty sensor/s Isolation
  • 4. Controller Reconfiguration

MED’13 Control Conference Crete, Greece

slide-14
SLIDE 14

x5

Multiplicative Additive

Sensor Fault Modelling

MED’13 Control Conference Crete, Greece

slide-15
SLIDE 15

Sensor Fault Scenaria

Ø Abrupt faults; additive faults the normal sensor’s output value superimposed with a low frequency random signal and multiplicative faults result to sensors’ output 5 times larger than normal. Ø 4 sensors in the maglev sensor set; assuming only 3 can fail i.e current i, vertical velocity and the vertical acceleration. Ø Both the abrupt/multiplicative and abrupt/additive sensor fault profiles are used for each sensor, Ø Subsequent faults can happen with a time difference of about 1 second.

MED’13 Control Conference Crete, Greece

slide-16
SLIDE 16

Ø Introduction Ø The proposed sensor fault detection scheme Ø Training the fault detection unit Ø The case study: A maglev suspension Ø Simulations Ø Conclusions

CONTENTS

MED’13 Control Conference Crete, Greece

slide-17
SLIDE 17

Simulation results

  • Mu. – Multiplicative faults
  • Ad. – Additive faults

FA – False alarms

  • St. – Stochastic response
  • Dt. – Deterministic response

MED’13 Control Conference Crete, Greece

slide-18
SLIDE 18

Fault free

  • Accel. Fault
  • vel. Fault
  • vel. Fault

Sensor Fault Sequence

MED’13 Control Conference Crete, Greece

slide-19
SLIDE 19

Binary Switch state

BS Output change BS From Decision making unit

Switching point

MED’13 Control Conference Crete, Greece

slide-20
SLIDE 20

A comparison

A bank of Kalman estimators have been compared with the iFD unit. The results show that the iFD unit is about 10 times faster than the bank of Kalman filters.

Kalman Filters iFD

MED’13 Control Conference Crete, Greece

slide-21
SLIDE 21

Ø Introduction Ø The proposed sensor fault detection scheme Ø Training the fault detection unit Ø The case study: A maglev suspension Ø Simulations Ø Conclusions

CONTENTS

MED’13 Control Conference Crete, Greece

slide-22
SLIDE 22

Ø A fault detection mechanism is proposed aimed to minimise the computational cost. Ø The proposed iFD is applied to a MAGLEV system example. Ø The simulation results show promising results. Ø Potential in FPGA implementation

Conclusions & Discussion

MED’13 Control Conference Crete, Greece

slide-23
SLIDE 23

AI-based Low Computational Power Actuator/ Sensor Fault Detection Applied on a MAGLEV Suspension

  • K. Michail, K. Deliparaschos, A. Zolotas, S. G. Tzafestas
  • A. Zolotas acknowledges Univ of Sussex for

travel grant support

MED’13 Control Conference Crete, Greece