Implementation of predictive spectrum control of a LC filter using ADMM

Constraints on current spectrum are common in many applications of power electronics such as power systems or traction applications. Model predictive control theory for systems with spectrum constraints has been recently developed. However, the implied optimization problem is the quadratically and linearly constrained quadratic programming (QLCQP) which is too complex to be directly implemented on short sampling periods common in power electronics. We are investigating the use of the alternating direction method of multipliers (ADMM) for QLCQP. This method allows for parallel processing which would allow its efficient implementation of a FPGA. On the problem of spectrum control of a LC filter, we study sensitivity of the method to its parameters to select the best possible parametrization and check suitability of its for implementation in FPGA.