Artificial intelligence is a vast body of knowledge, spanning areas such as supply chain management, retail, and the automotive industry. Systematic research is currently ongoing in the field of computational intelligence and the neural network is being applied on a large scale in the automobile sector. The field of AI and robotics are also growing together and the benefits are numerous.
There are several ways in which we apply AI to control motors in robots, drones, and smart vehicles. This article is an attempt to explain some of the possible applications. But first, let’s explore the history of the control of motors.
The History of Motor Control
It all began in the 1970s after the performance of DC machines was already established superior to that of AC machines. The advent of transistors and solid-state switches facilitated the vector control of AC machines and techniques such as constant V/F control and field-oriented control. These techniques, at the time, served to give these machines a very high level of maturity and placed AC motors on the same pedestal as DC machines.
Shortly after, the emergence of microprocessors and advanced digital signal processors engendered the rapid development of intelligent control of these motors. These advanced techniques offered much simpler alternatives to the control of motors.
Now, let’s move on to advanced artificial intelligence techniques for the control of motors.
The Fuzzy Logic Controller
The concept of fuzzy logic births the formulation of precise actions from mathematically imprecise ideas. this is governed by a set of fuzzy rules stemming from observations. There are three stages of the fuzzy rule; the fuzzification, inference, and defuzzification.
The fuzzy controller consists of two different input variables. These are the error variable and the error variation variable. In the control of the motor, the real speed of the rotor is compared with that of the reference model and the difference gives us the error variable. The system then regulates based on predefined fuzzy rules and functions to aptly control the motor.
Artificial Neural Network
The aim of neural modeling is to first bypass the parametric variations of the mathematical model of the motor being controlled. Initially, the elements of the network are taken after the application and it is trained, serving as the fixed network. Then, an online training is done so as to have the adaptation of the neural network in each moment. As a result, the behavior of a dynamic nonlinear system can be described accurately by the network even if the parameters aren’t known.
Usually, the network used for this command has a multilayer structure and it has a layer that is supported by the sigmoid function. The output layer, on the other hand, is enabled by a linear function. In general, the minimization function of a quadratic-modeled equation underpins the differences between the output of the neural network and the reference model.
Practical Application of Artificial Intelligence in the Control of Motors
Two companies recently signed a Memorandum of Understanding (MoU) to commercialize a technology that is used to detect and compensate for faults in electric motors. The two Canadian-based companies; Alizem and Opal-RT Technologies, came together to commercialize the AI technology that also allows motors to generate almost-optimal torque in the midst of faults. The beauty of this technology is that it can be used for permanent magnet synchronous motors, as well as brushless DC motors.
How does it Work?
The way in which this is achieved is that a data-based model of the motor in question is compared continuously with the model of a normal motor and certain voltage commands are applied to compensate for any discrepancy in the system. The algorithm comes in handy in certain critical situations, reducing the impact from catastrophic to manageable.
This technology is available in form of a software algorithm under patent. It is designed to be integrated into the firmware of motor drives and a simulation is carried out on the Opal-RT’s platform before it is implemented in order to ascertain its dynamic behavior.
We can go on and on about how impactful AI technology has been in the field of motor control. Just like every potential application of artificial intelligence, the future is bright. There’s no better way to say that now is the best time to get into the market.