On the Accuracy and Utility of Piecewise-Linear Models of Ferroresonance

In a previous paper1, one of us (JMF), presented a direct and conceptually simple non-linear analysis of the compensated transformer circuit that showed all of the normal and ferroresonant modes of behavior present in the real circuit. In that paper, we gave a long discussion of the two nodes of operation as well as some criteria for determining the thresholds for spontaneous mode switching. This will be only briefly reviewed below. The technique employed was reminiscent of magnetic amplifier analysis in which the T-I loop is replaced by a straight-sided approximation. We thought that the piecewise-linear nodel gave much greater physical insight into ferroresonant behavior than previous methods of analysis, but when we were asked from the floor whether the model was capable of accurately predicting the behavior of the actual circuit it purported to represent, we were unable to answer the question. Thinking of the model more as an explanation than a critical predictor, we had not attempted a point by point numerical comparison.