Linearized power flow for stochastic optimization

The behavior of network constraints in economic dispatch and unit commitment problem is examined in this paper. It analyses the constraints related to the power flow in a stochastic programming tool and tests different approaches for transmission loss computation. This work is part of an ongoing effort to develop tools for stochastic analysis of hybrid generation systems based up on the existing models but stressing on the network constraints and developing some proposals in linearizing AC power flow equations. The work takes a sample power network system in order to evaluate different approaches to model the transmission losses in a dispatch problem then compares the methods based on accuracy, computational burden and simplicity. The methods comprise iterative algorithm, piecewise linear loss function and a proposed linearizing approach on approximating AC power flow equations. The results show that the linearized AC method utilizes more computational resources and complicated formulation but with an improved accuracy. It can also integrate voltage magnitudes for further improvement in accuracy while the other two only employ a simplified formulation with a lowered accuracy.