Active Control of Thermoacoustic Instability in a Model Combustor with Neuromorphic Evolvable Hardware

Abstract

Abstract. Continuous Time Recurrent Neural Networks (CTRNNs) have previously been proposed as an enabling paradigm for evolving analog electrical circuits to serve as controllers for physical devices [6]. Currently underway is the design of a CTRNN-EH VLSI chips that combines an evolutionary algorithm and a reconfigurable analog CTRNN into a single hardware device capable of learning control laws of physical devices. One potential application of this proposed device is the control and suppression of potentially damaging thermoacoustic instability in gas turbine engines. In this paper, we will present experimental evidence demonstrating the feasibility of CTRNN-EH chips for this application. We will compare our controller efficacy with that of a more traditional Linear Quadratic Regulator (LQR), showing that our evolved controllers consistently perform better and possess better generalization abilities. We will conclude with a discussion of the implications of our findings and plans for future work.

Evolving analog controllers for correcting thermoacoustic instability in real hardware

“…In a previous research, we had demonstrated that EH techniques are better than traditional control techniques for suppressing TA instability in real time [6]. It was shown with evidence that controllers evolved using EH techniques were robust and superior in quality than the traditional controllers [8,9] for a simulated combustion chamber.…”

“…In the referred work by Gallagher and Vigraham [6], a five neuron CTRNN had been used for suppressing TA instability in a simulated combustion chamber. Our early experiments suggested that a 5 neuron CTRNN was not sufficient for suppressing TA instability in this analog circuit.…”

“…Our early experiments suggested that a 5 neuron CTRNN was not sufficient for suppressing TA instability in this analog circuit. While the model employed in [6] seems to have more complexity as indicated by the state equations on which the combustion chamber was simulated, a five neuron CTRNN was sufficient to suppress TA instability. In this work, we observed that a five neuron CTRNN was not sufficient to control TA instability on real analog hardware.…”

“…In our previous work [6], we applied an Evolvable Hardware (EH) control technique for suppressing TA instability in a simulated combustion chamber. In those studies, our evolved controllers outperformed traditional controllers reported in the literature and were able to do so without introducing additional resonance modes.…”

A Case for Using Minipop as the Evolutionary Engine in a CTRNN-EH Control Device: An Analysis of Area Requirements and Search Efficacy

“…They have been previously demonstrated as effective control devices for 1. Suppressing Thermoacoustic (TA) instability in simulated jet engines [1]. 2.…”