طراحی کلنی مورچه ی کنترل کننده مبتنی بر بازخورد برای نرم افزار آغازگر بانک فدرال درایو موتور القایی

The objective of this work is to design and implement a closed loop system for induction motor starting at rated current. Thyristorized AC voltage regulator is used as the starting equipment and motor current regulation is carried out using an optimally tuned Proportional-Integral (PI) controller. Since, AC voltage controller fed starting of induction motor is a non-linear process, identification of optimal values of PI controller constants is performed using a novel ant colony based optimization technique. The complete drive system including AC voltage controller fed induction motor in conjunction with optimal PI controller is first simulated in MATLAB and subsequently verified experimentally. The successful implementation with a low cost microcontroller illustrates the feasibility of the new approach.

The methods of starting of three-phase induction motors are generally classified into four basic categories [1]: direct online starting, electromechanical reduced voltage starting, solid-state reduced voltage starting and variable-frequency-drive (VFD) starting. The cheapest way is direct online starting, but the major disadvantage is large initial current surge leading to voltage dips in the supply system. Electromechanical reduced-voltage starting comprises auto-transformer starting, star/delta starting and resistor/reactor starting. Among these, star/delta starting is more popular since it is cheap, compact and causes no power loss. However this method can be applied only for normally delta-connected stator provided with six leads. Conventional starting elements require some type of mechanical switch or contact and have several drawbacks [2] and [3], such as need for frequent inspection and maintenance, non-simultaneous switching of motor phases to the supply, failure in the moving parts due to large amount of switching etc.,. Availability of high-power devices has led to the development of solid state starters replacing conventional starters. An earlier work in this direction is seen in [4] and [5]. The solid state starter consisted of a pair of back-to-back connected SCRs and the firing angle of SCRs was varied to control rate of change of KVA with respect to time. Such a scheme is examined for pulp and paper industry in [6]. A solid state voltage contactor employing anti-parallel SCRs is given in [7]. The concept of use of SCR voltage regulator with firing angle control for smooth induction motor starting is labeled as “soft starter” and a closed loop soft starter with fuzzy logic controller is available in [8]. A closed loop optimal soft starting ac voltage controller fed induction motor drive based on voltage across the thyristor is seen in [9]. While earlier works were confined to current limit starting, performance enhancement of starting torque is profile was later addressed [10] and [11]. Numerical solution method is attempted in [10] for improved starting current and torque profile whereas the method suggested in [11] consists of two parts: a linear initial firing angle scheme which eliminates starting torque ripples and a current control strategy which consists of successive co-sinusoidal and constant function segments of triggering angle of SCRs. A soft starter with voltage and power factor angle as feedback signals is given in [12], whereas starting torque optimization is carried out in [13]. The work in [14] explains the use of Artificial Neural Network (ANN) for the generation of SCR firing angles for soft starting. Soft starters are also employed for energy saving too [15] and [16]. While the schemes in [4], [5], [6] and [7] are of open loop type, the soft starters in [8], [9], [10], [11], [12], [13] and [14] work in closed loop mode. The closed loop control strategies are obtained by repeatedly simulating the motor model several times and employing trial and error based approach. While the soft-starter has been increasingly employed in the industries now-a-days, the design procedure of feedback controller for rated current starting is not available in literature. This is important in particular because, AC voltage controller fed induction motor drive system is described by a fifth order differential equation and is highly non-linear. This demands a detailed design and implementation of a feedback controller for soft-starter fed induction motor drive. In this paper, the feedback controller design is formulated as an optimization problem and the solution is achieved to obtain good dynamic response. A close examination of the available research work indicates that systematic design and implementation of a soft-starter with PI controller is not available in the literature. This paper proposes an optimization model for PI controller based motor current regulation during motor starting. The optimization model makes use of d–q axis model for evaluating objective function and controller constants are then iterated through ant colony optimization (ACO) [17]. We have employed a modified ACO for the design of a PID controller to be used in the closed loop of a boost converter for output voltage regulation. The concept of a group of ants as a colony of cooperating agents was first proposed by Dorigo et al. [17], [18], [19], [20], [21] and [22]. It was observed that real ants, which are almost blind and have very simple individual capabilities, are capable of finding shortest path between their home colony and food source. When an ant moves, it deposits a chemical substance called pheromone on its path which can be detected by other ants and their movements are biased by pheromone trails left on the ground by the preceding ants. This pheromone laying mechanism thus acts a channel of communication between the ants. While each ant starts essentially at random, the one that finds food source first returns to its nest first and this ant therefore, makes many more traverses between the food and nest. Thus, the pheromone accumulation in this path becomes stronger and all ants eventually use this shortest path. The convergence of all ants to a single, shortest possible path can be attributed to the collective behavior of ants in a colony; ants coordinate their activities exploiting indirect communication among them through pheromone laying mechanism. Algorithms based on ACO had been developed and tested successfully for a variety of combinatorial optimization problems such as power electronic circuit design [19], self-structuring antenna [20], fuzzy controller design [21] etc. Various swarm intelligence techniques are available in [22], [23], [24] and [25]. While ACO is employed for a variety of applications, such methods are suitable for discrete optimization; in discrete optimization, distinct points already exist as possible solutions and the task is to identify one of the many combinations of array of these points which are best suited to the problem. Thus when initially ACO concept was applied to discrete optimization tasks, the ant moves from one distinct point to the next. In this paper, the feedback controller design for a soft-starter fed Induction motor drive is formulated as an optimization problem and the existing ACO is tailored suitably for continuous optimization of the PI controller parameters namely Kp and KiKp and Ki which are not generally discrete. With continuous optimization, only bounded solution space is available where the solution has to be continuously explored. Extensive simulation and experimental works were carried out with different types of disturbances using the feedback controller identified through ACO. The results show that the proposed design methodology gives a robust controller which is far superior to traditionally designed controller. Extensive simulation results are provided to validate the proposed method. In addition, a prototype soft- starter with microcontroller based PI controller is fabricated in the laboratory and experimental results are recorded to validate the simulation findings

This work has focused on the design and implementation of a closed loop scheme for soft-starter fed induction motor drive with PI controller. The objective of the closed loop operation was to start the motor at rated current. Towards this goal, soft-starter fed induction motor drive with closed loop operation is first modeled in MATLAB. The feedback controller constants are identified for enhanced starting dynamics using ACO. A modified version of ACO with a new variable termed as step movement of ant is newly introduced. The complete drive system is simulated in MATLAB first and subsequently implemented in the laboratory. A low cost PIC16F876A microcontroller is used as the feedback controller and firing pulse generator for AC voltage controller. The computed and measured results corroborate each other and validate the proposed method. As a direction for future work, the step length of ant movement can be dynamically changed than keeping it constant at a fixed value. This can enable faster convergence of the algorithm.