Due to increasing intermittent renewable energy utilization in smart grids, smart management of energy balance is becoming major concern of research studies. In the case of highly volatile generation and demand conditions, it is very important to estimate instant energy balance state of electricity grid to preserve energy balance automatically and avoid from the cases of generation deficiency and overloading. The utility frequency deviation is commonly used to restore power imbalance in AC grids. This study gives a discussion on the utilization of frequency deviation of AC power system as an indicator of energy balance state in smart grids. We consider two approaches to estimate instant AC frequency deviation of power systems under serious harmonic and noise conditions. The first one is based on frequency to amplitude conversion technique, which use band-pass ramp filtering of AC signal. The other method is based on the sinusoidal to pulse waveform conversion, and it uses measurement of pulse periods of the resulting square waveforms. We compare these two methods and discuss their advantages and disadvantages to generate a balance error signal for smart grid applications. A brief discussion on application of these signals for smart grid energy management is given.