Three dimensional modelling and optimisation of multistage collectors

• Main Author: Ghosh, Tushar Kanti
• Other Authors: Carter, Richard
• Published: Lancaster University 2002
• Subjects: 621.381335
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.403731

Multistage collectors are commonly used in travelling wave tubes (TWTs), klystrons, gyrotrons and inductive output tubes (IOTs). A dc beam transfers some of its energy to the input RF wave during its travel through the interaction circuit. The purpose of a collector is to recover most of the remaining dc power from the spent beam and thereby increase both the collector and the overall efficiency. Secondary electrons play a detrimental role in collector performance. Use of the asymmetric collector geometries and application of a magnetic field in the collector region have proved to be effective in recapturing the secondaries. A fully three-dimensional simulator LKOBRA (MF) – the mainframe version of Lancaster KOBRA, has been developed at Lancaster University; this is capable of simulating multistage collectors including the effects of secondary electron emission and magnetic field. It is based on KOBRA3-INP1 but has been modified and improved together with the pre- and post-processors of the package. Efficiency is an important parameter in space applications of microwave tubes so it is always desirable to optimise the collector performance to maximise the overall tube efficiency. As a first step the potentials at the collector electrodes are optimised to achieve the maximum possible theoretical efficiency. A computer code based on the well-known enumerative technique has been developed for this purpose. In the next step the geometry of the collector electrodes is optimised using an automated design package that is based on a genetic algorithm. The genetic algorithm creates a new geometry through a search procedure that works from a population of possible geometries. A new set of geometries is generated using three basic operators namely reproduction, crossover and mutation. The collector efficiency is used as the fitness parameter in the genetic algorithm that produces a new population of geometries. This package has been used to optimise the efficiency of a 4-stage symmetric collector and a 2-stage asymmetric collector.