Determination of LED Lamp Model Parameters by Using Analytical - Optimization Approach

Light-emitting diode (LED) lamps are widely used for electric lighting due to their low energy consumption, high luminous efficiency and long lifetime. There are, however, several important drawbacks of LED lighting technology such as their single-phase character and non-linear voltage-current characteristic resulting in a distorted electric current. To predict the harmonic current injection from LED lamps into the grids, this paper investigates suitable models that can accurately predict these injection. Since the structure of the typical LED lamp models is known, the parameters of these models need to be determined. This paper presents a method for estimating the unknown parameters of typical, widely used LED lamp models. The proposed method is based on: (a) a general approach to analytically determine the waveform of the input LED lamp current and (b) the definition of the objective function for minimizing the deviations between the calculated and the measured LED lamp input current. The Nelder-Mead search optimization method was used to determine the optimum of the objective function. The described LED lamp modes and the parameters determined with the proposed estimation method were verified on the example of two different, widely used LED lamps. Excellent agreement was found between the estimated and measured LED lamp model parameters, as well as between the measured and simulated input LED lamp current waveforms. Furthermore, the influence of the defined initial vectors of the optimization method and the influence of the harmonic order of the source voltage on the results of the parameter estimation of the analyzed LED lamp models was investigated.