1,721,007 research outputs found
Systems integration using evolutionary algorithms
No full textThis paper describes an approach to the systems integration problem using multiobjective genetic algorithms. An architecture for evolutionary systems integration is presented and the component parts discussed. An example of an aircraft gas turbine engine control system design problem is shown demonstrating aspects of the proposed architecture that allow many design objectives from different disciplines to be considered in parallel. Potential closed-loop control configurations are evaluated and compared against one another within an optimization framework. As a result of this analysis, it is shown how informed decisions may be made regarding the nature of the control employed, acceptable performance margins and elements of the engine design
Parallel computing in CACSD
No full textComputer-aided design of control systems via multiobjective optimisation is a computationally demanding task that may benefit from parallel processing techniques. In this paper, we report on a new parallel processing gateway that supports the use of parallel processing within the framework of an existing computer-aided control system design software package. In many control system design exercises which employ optimisation, the bulk of the computational effort is devoted to the evaluation of the objectives of the optimisation at each iteration. This paper demonstrates, with an example, how, using the gateway, parallel processing can be used within the framework of existing computer-aided control system design tools to compute these objective values
Simplifying particle swarm optimization
Full text linkThe general purpose optimization method known as Particle Swarm Optimization (PSO) has received much attention in past years, with many attempts to find the variant that performs best on a wide variety of optimization problems. The focus of past research has been with making the PSO method more complex as this is frequently believed to increase its adaptability to other optimization problems. This study takes the opposite approach and simplifies the PSO method. To compare the efficacy of the original PSO and the simplified variant here, an easy technique is presented for efficiently tuning their behavioural parameters. The technique works by employing an overlaid meta-optimizer, which is capable of simultaneously tuning parameters with regard to multiple optimization problems, wheras previous approaches to meta-optimization have buned behavioural parameters to work well on just a single optimization problem. It is then found that not only the PSO method and its simplified variant have comparable performance for optimization a number of Artificial Neural Network problems, but also the simplified variant appears to offer a small improvement in some case
Hybrid control and evolutionary decision support within a sustainable environment
No full textDue to the increased global demand for energy, and the potential dangersof relying too heavily on our fossil fuel reserves, more and more research is being directed towards alternative, and preferably reusable or sustainable forms of energy supply. Many of these real world systems have operating regions that exhibit varying degrees of non-linearity. An example of this are the significant variations in the dynamic characteristics of a distributed collector field within a solar power plant. Here a control scheme employs a fuzzy PI controller, with feedforward, for the highly nonlinear part of the operating regime and gain scheduled controller for the more linear part of the operating envelope. In order to satisfy performance characteristics for the plant at different points in the operating regime, a multiobjective genetic algorithm with enhanced decision support system, is used to design the parameters of the fuzzy controller
Tuning differential evolution for artificial neural networks
No full textThe efficacy of an optimization method often depends on the choosing of a num-ber of behavioural parameters. Research within this area has been focused on devising schemes for adapting the behavioural parameters during optimization, so as to alleviate the need for a practitioner to select the parameters manually. But these schemes usually introduce new behavioural parameters that must be tuned. This study takes a different approach in which finding behavioural parameters that yield good performance is con-sidered an optimization problem in its own right and can therefore be attempted solved by an overlaid optimization method. In this work, variants of the general purpose op-timization method known as Differential Evolution have their behavioural parameters tuned so as to work well in the optimization of an Artificial Neural Network. The re-sults show that DE variants using so-called adaptive parameters do not have a general performance advantage as previously believe
Reconfigurable flight control strategies using model predictive control
No full textTracking control for large amplitude manoeuvres in the presence of damages to the airframe and control surface is addressed. The control reconfiguration algorithm is based on model predictive control, a constrained reseding horizon optimization is solved under the constraints of hard limits of actuator position and rate saturation and critical aircraft state limits imposed by allowable structural loads. Changed stability and control derivatives of the damaged aircraft are identified online and used buy the receding horizon controller as internal model for prediction. An emphasis is given to incorporate handling quality specification according to MIL-STD-1797A. The results demonstrate the ability of the proposed scheme to maintain flight after a failure; trach the pilot commands despite loss of actuator effectiveness, and to coordinate the use of the remaining active control surfaces to provide the decoupling between the rotational axes. Finally the issue of the online (onboard) implementation of the constrained optimization is examined
Evolutionary design of gas turbine aero-engine controllers
No full textThis paper describes a novel approach to the design of a control system for an aircraft gas turbine engine. A multi-level multiobjective genetic algorithm is employed to design controllers at both individual operating points using system linearisations and small signal response characteristics, and over the full-flight envelope using a nonlinear model and large signal responses. The proposed approach should allow the selection of smoother controller parameters over the flight envelope and ensure that more consistent control demands are made at off-design operating point
PARSIM: a parallel optimization tool
No full textCurrent computer-aided control system design environments seldom support optimization methods for controller design in a truly interactive manner. A prototype tool, called PARSIM, supporting parallel processing, optimization, and a graphical user interface is presented, addressing many of the problems inherent in current approaches to multiobjective optimization-based design methods. An XWindows interface is used to simplify problem formulation and control the optimization processes. Using a previously developed interface, it is shown how the computational burden may be alleviated by parallel processin
Highly nonlinear control of a solar thermal power plant using soft computing fuzzy tuning techniques
No full textSociety is experiencing massive growth of global industrialised populations, which is putting
increasing pressure on western governments to pursue more persuasive means to maintain and increase
their share of the world’s diminishing fossil fuel reserves. To combat this, there is a growing body of
enlightened researchers who are directing their abilities towards the development of alternative and
preferably renewable energy types of supply systems. Many of these real world systems exhibit varying
degrees of non-linearity. An example of this is the significant variations in the dynamic characteristics of
a distributed collector field within a solar thermal power plant. Here a Sugeno-type fuzzy incremental
controller was tuned using an ANFIS (Adaptive Neural Fuzzy Inference System) to optimise the fuzzy
controller’s pre-clustered input membership functions, while a multiobjective genetic algorithm with an
enhanced decision support system was used to fine tune the parameters of its first order output
membership functions. The resulting solution choice produced an incremental fuzzy controller which was
used to successfully control the plant exclusively in its high nonlinear regions, i.e., where the oil flow fell
below 5 litres per second. This allowed the plant to function in environments where local solar radiation
conditions have always been regarded as marginal. A feedforward term was also used to control plant
disturbances caused by solar irradiation, mirror reflectivity etc
- …
