Sspesmann Project

SENSORLESS POSITION ESTIMATION FOR
STEPPER MOTOR USING ARTIFICIAL

NEURAL NETWORK (ANN) CONTROL
Abstract-This paper presents a novel approach for sensorless position
estimation and direct torque control (DTC) of stepper motors using
artificial neural network (ANN) control. The proposed method eliminates
the need for position sensors, reducing cost and increasing reliability. The
ANN is trained offline using data from the stepper motor and then used to
estimate the motor's position and speed in real-time. The estimated
position and speed are used to control the motor's torque and flux using
DTC. With this method the computing time is reduced. Improved
Diagnostics can thus be achieved through The use of Sensor less
algorithms instead of via additional sensors with their associated higher
costs and lower reliability. Keywords: ANN(Artificial neural
network),Stepper Motor, DTC(Direct Torque Control)
Ⅰ. INTRODUCTION
The stepper motor [1] is used in several industrial applications where torque at low speed,
positioning accuracy, and high-speed dynamic are determining factors. It can, thus, be found
in numerous applications. For example in medical applications for, accurate dosage with
peristaltic pumps or pipettes and motion control in dialysis equipment. Also, in automotive
applications, combustion engines contain several systems to control emissions and reduce
NOx level. These systems always combine a stepper motor that drives efficiently integrated
valves. In telecommunications, the stepper motor is used to actuate antennas and
combiners. All these applications require a robust actuator that can withstand vibrations and
must respond quickly and precisely to a signal position, while at the same time guaranteeing
that it can overcome a dynamic torque load. The stepper motor has the advantage over
other kinds of motors because it works well in open loop since it carries out motion step by
step when voltage impulses are applied to its phases. But, due to its inertia, the rotor
oscillates around the final position before stabilizing. In addition, the motor can lose steps if
the variation of the load torque is fast. The use of new programmable architecture like field-
programmable gate arrays (FPGA) circuit [2] allows the creation of microstep movements.
That smoothes the movement but leaves the open-loop control problem unsolved. To solve
it, it is necessary to introduce a closed loop to improve the performance and the robustness
of the control. A first solution was the introduction of mechanical sensors. This solution
increases the size and the cost of the system. Furthermore, analysis of past experience from
automotive applications, for example, shows that mechanical measures such as position or
speed suffer in high-temperature- and high-vibration environment. Due to these limitations,
sensorless control emerges, with today’s low cost and high-performance DSP’s, as an
alternative and a very interesting solution since it saves the expensive and bulky mechanical
sensor. Several sensorless methods including standstill were proposed for various motors
and can be applied to the stepper motor. The knowledge of the initial rotor position
guarantees a starting of the motor in the desired direction [3]. The standstill methods [4], [5]
are based on the inductance variation according to the rotor position. Voltage signals are
injected to the windings and the initial rotor position is obtained by monitoring the phase
currents. For low speed, this principle is used in [6] to estimate the rotor position without
voltage injection. The phase incremental inductance is estimated in real time and compared
to an analytical model to estimate the rotor position. The highfrequency signal injection
method [7]–[10] can detect the initial rotor position and extend the estimation to low
speeds. The back electromotive force method is widely used because of its easy
implementation [11][16]. For middle and high speeds, the observers are more suitable. The
sliding mode observer is characterized by its robustness against the disturbance and motor
parameters variation as shown in [17] and [18]. A new nonlinear observer has been recently
proposed in [19] to estimate the rotor position without information about the motor speed.
The extended Kalman filter (EKF) [20] is often used in the speed sensorless control because it
has low-pass filter characteristics. To reduce the computing time of the EKF algorithm, [21]
assumed an infinite inertia of the motor. Unscented Kalman filter [22] can improve the
estimate and reduce the computing time. In [23], use of the Kalman filter in steady-state
case for hybrid stepper motor with a speed control is proposed, but only simulation results
were shown. The position control of the stepper motor is the subject of many pieces of
research [24] that integrates a mechanical sensor. This study proposes to investigate a
simplified version of the EKF algorithm to estimate the mechanical state variables of the
stepper motor by only measuring the line voltages and currents. Then, the feedback control
is achieved from these estimated variables instead of the measured ones. Furthermore, by
applying the principle of field orientation, the dynamic performance of the stepper motor is
considerably improved 1 and it becomes a high-dynamic ac drive.
Ⅱ. PRINCIPLE OF OPERATION OF STEPPER

MOTOR
Stepper motors work on the principle of electromagnetism which helps this motor to
achieve precise control over its position. A stepper motor consists of a rotor and a stator. A
rotor is a permanent magnet toothed with an iron core and a stator consists of multiple
windings arranged in a circular pattern. The working principle of the stepper motor is based
on sequentially energising the stator winding which generates a rotating magnetic field and
causes the rotor to rotate in discrete steps.
When an electrical current flows through the stator winding, it creates a magnetic field. The
direction and magnitude of this magnetic field depend upon the current’s direction. By
sequentially energising the stator winding in a controlled manner, the magnetic field rotates
around the stator and causes the rotor to rotate in discrete steps.
The sequence of energising the stator winding is controlled by an external driver or
controller. The energising sequence determines the rotor’s movement. There are several
methods to energise the windings.
Full step method energises the two phases of windings simultaneously causing the rotor to
rotate at full step (1.8° step) and providing high torque. Half-step mode alternately energises
the phases of windings and causes the rotor to rotate at half step (0.9° step) and offers a
good combination of speed and torque. Micro stepping method partially energises the
multiple phases of windings simultaneously which provides smooth movement with high
speed and low torque.
The movement and speed of the stepper motor are controlled with the help of controllers.
The driver receives the signal from microcontrollers or computers. These signals dictate the
timing and sequence of energising the windings, which determines the motor's speed and
direction of rotation.
The rotor aligns itself with the initial energised winding. The controller sends a pulse to
energise the next coil in the sequence which causes the rotor to align itself with a new
magnetic field and rotates in discrete steps.
Each subsequent pulse from the controller energises the next winding and causes the rotor
to rotate to the next position. This process continues with the rotor moving in discrete steps
in response to the pulse of the controller. This process provides precise control over the
position, speed and torque of the motor.
Conclusion- A stepper motor works by converting the electrical pulses into precise
mechanical movements. As stator windings are energised in a particular sequence, which
causes the rotor to move in discrete steps. This operation provides precise control over the
speed and position of the stepper motor and makes them indispensable components in
applications that require precise control movements.
Ⅲ. ANN BASED POSITION ESTIMATION

The sensor less control of stepper motor can be achieved by the ANN control to determine
the rotor position and speed so the ANN position used to determine position of stepper
motor based on real time control signal and making them suitable for dynamic environments
by using ANN we can reduce the need of expensive sensors by relaying on predictive
modelling.
DTC- In order to control the torque instantaneously and improve the performance of stepper
motor, direct torque control strategy is used according to the motor characteristics and
depending upon relation between flux and current.

Sspesmann Project

Uploaded by

Copyright:

Available Formats

Sspesmann Project

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sspesmann Project

Uploaded by

Copyright:

Available Formats

SENSORLESS POSITION ESTIMATION FOR

STEPPER MOTOR USING ARTIFICIAL

Ⅱ. PRINCIPLE OF OPERATION OF STEPPER

Ⅲ. ANN BASED POSITION ESTIMATION

You might also like