Elsevier

Electric Power Systems Research

Volume 116, November 2014, Pages 29-35
Electric Power Systems Research

Design optimization of PM couplings using hybrid Particle Swarm Optimization-Simplex Method (PSO-SM) Algorithm

https://doi.org/10.1016/j.epsr.2014.05.003Get rights and content

Highlights

  • A PSO-SM optimization technique has been proposed to optimize the design of the PM coupling drive.

  • The coupling performance is predicted using a layer model approach that compared with FEA for validation.

  • The proposed PSO-SM technique is compared to other published techniques like PSO, and GA-SM.

  • The simulation showed that the proposed algorithm (PSO-SM) is the most efficient one.

Abstract

The aim of this paper is to explore the use of the proposed hybrid Particle Swarm Optimization-Simplex Method (PSO-SM) algorithm to optimize the design of PM couplings subject to several key design constraints. The proposed hybrid optimization algorithm is constructed based on combining two well-known optimization techniques: Particle Swarm Optimization (PSO) and Simplex Method (SM). The PSO has obvious capabilities in global search while the SM has exceptional advantages in local search. As a hybrid algorithm, the PSO-SM has the outstanding feature of combining the ability of global searching and local canvassing. On the other hand, Permanent Magnet (PM) drive couplings are used in power transmission in a wide range of industrial applications. A standard coupling design is used as a good starting point for the conventional Simplex Method and to define the performance constraints for the proposed hybrid optimization algorithm. New coupling designs are developed and optimized to demonstrate the superior capabilities of PSO-SM algorithm as a global optimization technique.

Introduction

The design of electric machines has long been the field of experience, extrapolation, trial-and-error and a hit of good fortune! Modelling and simulation techniques on the other hand have improved progressively over the last years and have somewhat helped to reduce some of the uncertainty and risk associated with machine design. Formal optimization techniques are readily available and could make a considerable contribution to machine design. The reason why their impact has been relatively minimal lies in the quality of the models that represent the heart of the optimization algorithm. A model of poor quality cannot be expected to give a truly optimal design. The conventional deterministic optimization techniques themselves have problems normally associated with local solutions rather than global solutions and the dependence on their starting point. In this respect, the newer stochastic methods and in particular the PSO algorithm approaches overcome these issues but usually at the cost of relatively increased search times.

The main drawbacks of the conventional deterministic techniques are that their efficiency depends on starting point, the search step length, and the accuracy of the descent direction evaluation. Also, the algorithm generally searches a point of local optimum unless the objective function is uni-modal. These disadvantages are clear in the results of simplex optimization method (SM) in references [1], [2], [3], [4] where convergence of the function to different points indicates the existence of local minima. On the other hand, although the PSO algorithms have been used for many types of applications especially for electric machine control [5], [6], [7], [8], [9], [10], they still have not been tuned to get rid of relatively convergence slowness especially for large search space optimization problems.

This paper aims to give further contribution in the “optimum” design of PM coupling drive using the proposed hybrid PSO-SM Algorithm. The results will be compared with the optimized results obtained by hybrid GA-SM of ref [3], and with that of using the PSO alone in ref [11] for exploring the merits of the proposed hybrid optimization technique.

Section snippets

Optimization Algorithms

In the design optimization of electrical machines, the following issues are worth bearing in mind [12]:

  • -

    There are usually many contradictory design objectives, which need to be optimized together and measured in different scales like volume and efficiency.

  • -

    Most design methodologies contain relaxable or soft constraints, as well as rigid or hard constraints.

  • -

    Sufficient information about the consumer requirements is usually available for the designer to determine his objectives and hence the

PM Coupling Drive

In the last two decades a range of permanent-magnet eddy current couplings was introduced into the industrial market offering the advantages of efficient operation, vibration/electrical isolation and jamming protection. These have found a wide scale application in blowers, compressors, conveyors and pumps [22], [23].

The principle of operation is relatively simple: one shaft is connected to a disc containing a number of embedded alternating polarity magnets. These are magnetized in the axial

Simulation results

A full list of the basic coupling design parameters are summarized in Table 1.The steel relative permeability was fixed at μr = 500 in the optimization because the flux density levels in the original design were low enough to ignore magnetic saturation. The assumed operating temperature was 20 °C which would be unrealistic in practice; however a nominal fixed temperature was used simply to illustrate the success of the optimization in reducing the volume of the standard coupling design and to

Conclusion

A constrained multiobjective optimization of PM drive coupling has been presented using Simplex Method (SM), Particle Swarm Optimization (PSO), and hybrid PSO-SM techniques. The hybrid technique combines the advantages of the PSO algorithm for locating the interval of a global optimal solution in the parameter search space with the faster execution times of classical SM technique in converging to the global solution.

A design optimization of the PM coupling with the objective of minimizing the

References (26)

  • H.B. Ertan et al.

    A method for optimum design of switched reluctance machines

  • A. El-Wakeel et al.

    Optimal design of switched reluctance motors using genetic algorithms

  • A.C. Smith et al.

    Formal design optimization of PM drive couplings

  • A. El-Wakeel et al.

    Genetic design of a switched reluctance motor based on minimum outer volume

  • M.E.-S. El-Telbany

    Employing particle swarm optimizer and genetic algorithms for optimal tuning of PID controllers: a comparative study

    ICGST-ACSE J.

    (2007)
  • A.S. El-Wakeel et al.

    Optimum tuning of pid controller for a permanent magnet brushless DC motor

    Int. J. Electr. Eng. Technol. (IJEET)

    (2013)
  • Z-L. Gaing

    A particle swarm optimization approach for optimum design of PID controller in AVR system

    IEEE Trans. Energy Convers.

    (2004)
  • Q.H. Haibing Hu et al.

    Optimal PID controller design in PMSM servo system via particle swarm optimization

  • J. Kennedy et al.

    Particle swarm optimization

  • Y. Shi et al.

    A Modified Particle Swarm Optimizer

  • A. Wang et al.

    Structural optimization of the permanent magnet drive based on artificial neural network and particle swarm optimization

  • A.S. El-Wakeel

    Design optimization of electric machines

  • R. Eberhart et al.

    Comparison between genetic algorithms and particle swarm optimization

  • Cited by (36)

    • Optimization of IC engine mounts for BAJA All-Terrain Vehicle

      2022, Materials Today: Proceedings
      Citation Excerpt :

      The function is only highly reliant on the initial value provided to the function and it will proceed to find other local minima based on the tolerance and step size given to the function (the more accurate, the more computationally intensive). Utilizing Gradient Descent, SQP produces a relatively fast convergence rate, that is robust and efficient and has been implemented in various fields over the years [6–7]. The process works on the Active set algorithm that is in-built.

    • Design optimization of APMEC using chaos multi-objective particle swarm optimization algorithm

      2021, Energy Reports
      Citation Excerpt :

      However, only the influence of structure parameters on output torque is considered, and other performances are not optimized. In [6], particle swarm optimization (PSO) combined with simplex method was used to optimize the cost objective of permanent magnet eddy current coupler. In order to optimize the structure and performance of PMEC, obtain larger output torque and lower eddy current loss of axial permanent magnet eddy current coupler (APMEC), a MOPSO based on chaos algorithm is proposed.

    • A review on experimental design for pollutants removal in water treatment with the aid of artificial intelligence

      2018, Chemosphere
      Citation Excerpt :

      Another similar approach is simplex method, which is characterized by a derivative-free line search method. This is based on the idea of comparing the values of the objective functions at the N+1 vertices of a polytope in N-dimensional space and moving the vertex towards the minimum point as an optimization is in progress (El-Wakeel, 2014). This limitation of ANN has motivated researchers to generate ideas of merging or hybridizing ANN with other heuristic approaches in the search for a better performance.

    View all citing articles on Scopus
    View full text