A neural-network model of the input/output characteristics of a high-power backward wave oscillator

Abstract

This paper discusses an approach to model the inputloutput characteristics of the Sinus-6 electron beam accelerator-driven backward wave oscillator. Since the Sinus-6 is extremely fast to warrant the inclusion of dynamical effects, and since the sampling interval in the experiment is not fixed, a static continuous neural network model is used to fit the experimental data. Simulation results show that such a simple nonlinear model is sufficient to accurately describe the inputloutput behavior of the Sinus-6-driven backward wave oscillator (BWO) and that the fitted output waveforms are basically noiseless. This model will be used to control the BWO in order to maximize the radiated power and the efficiency. This paper is also intended to introduce high-power microwave researchers to control concepts that may enhance the outputs of a wide spectrum of sources.

Toward "smart tubes" using iterative learning control

“…From a controls perspective, it turns out that the fast dynamics and changes in the operating characteristics of the BWO render traditional automatic-control methods ineffective. In fact, our results in [7] for the Sinus-6-driven BWO have shown that a static affine model is an accurate representation of its input/output characteristics. A static model is too fast to be controlled in real time while maintaining the sameorder dynamics (and thus, the same bandwidth and speed of response) of the open-loop system.…”

“…Note that in the figure, the controller actually contains two parts, one which is based on the known model and is thus fixed and produces , and another denoted by and obtained as the output of the learning controller. In our simulations, the known model is obtained from the identification we performed in [7], which is the neural network fitted to the actual experimental data.…”

“…This agility was obtained through an axial displacement of the SWS with respect to the "cutoff neck" inlet to the electrodynamic system. Furthermore, a companion study to this paper provided a static affine model of the input/output characteristics of the Sinus-6driven BWO, and described how this information can be used as part of an algorithm to meet specific control objectives, such as maximizing the radiated frequency bandwidth for a fixedpeak radiated-microwave-power level [7]. The next step in the research is then to automate the control algorithms in order to build a "smart-tube" HPM source.…”

“…On the other hand, the control problem for static systems is relatively undeveloped. Control issues for static systems have recently arisen in many areas such as rapid thermal processing [8] or in pulsed power systems [7]. Static systems may, in fact, have dynamics, but their input-output response is so fast that they defy the usual description with differential/difference equations.…”

Advances in the control of a "smart tube" high power backward wave oscillator

“…It was this periodic dependence on the output characteristics of our HPM source that we exploited to achieve a "smart tube." A neural network model was developed to describe the input/output characteristics of our HPM source [3]. Furthermore, an ILC algorithm was developed to achieve the control objectives using a minimal number of iterations [4].…”