I got a signal as attached. My doubt is can I break the signal in a triangle and a straight line, then calculating fourier series for each. In other words, should I calculate fourier series of triangle only to get equivalent signal.
Ajay, the signal is a triangle how you say. If the signal is not periodic, you can use the Fourier Transform insted of Fourier Series. Moreover, if you assume that the signal is periodic, you can use the Fourier Series and then you constrain the domain to the graphic shown.
Dear Gustavo Sir, I have shown a single cycle. Signal contains 1200 cycles. Each cycle has a triangle (1sec) followed by straight line of 10 sec. Now can I calculate Fourier series of triangle and ignore the straight line. One cycle calculation will be further used for entire signal.
Ajay, if your cycle really (exactly) is as you said, a triangle followed by a straight line, the signal could analitically be studied, as you could do for a simple sinusoidal cycle (and obtain a dirac on each positive and negative component). Fourier transform (discret) of this periodic cycle (continuous). Of course, the total current cycle must be used, as said Gustavo, otherwise the result of triangle alone (well known) and line alone (well known too) cannot be simply added (possible by introducing delay of one of them--> phase delay in spectral domai, to study...). If the mathematical solving is too difficult (integral problem) you could do a Fourier serie (discret) ot the periodic discret cycle (discret) by using the classical cooley-Tukey algorithm (or equivalent). On my side, i use fftw3.