SimPowerSystems

Choosing an Integration Method

Three solution methods are available through the Powergui block. These are:

• Continuous solution method using Simulink variable-step solvers
• Discretization for solution at fixed time steps
• Phasor solution method using Simulink variable-step solvers

Continuous versus Discrete Solution

One important feature of SimPowerSystems is its ability to simulate electrical systems either with continuous variable-step integration algorithms or with a fixed-step using a discretized system. For small size systems, the continuous method is usually more accurate. Variable-step algorithms are also faster because the number of steps is fewer than with a fixed-step method giving comparable accuracy. When using line-commutated power electronics, the variable-step, event-sensitive algorithms detect the zero crossings of currents in diodes and thyristors with a high accuracy so that you do not observe any current chopping. However, for large systems (containing either a large number of states or nonlinear blocks), the drawback of the continuous method is that its extreme accuracy slows down the simulation. In such cases, it is advantageous to discretize your system. In the following two sections, we explain these two methods, their advantages, and their limitations.

What do we mean by "small size" and "large size"? Although the distinction is not sharp, you can consider small size a system that contains fewer than 30 electrical states and fewer than 6 electronic switches. Circuit breakers do not affect the speed much, because unlike power electronic switches, which are commutated at every cycle, these devices are operated only a couple of times during a test.

Phasor Solution Method

If you are interested only in the changes in magnitude and phase of all voltages and currents when switches are closed or opened, you don't need to solve all differential equations (state-space model) resulting from the interaction of R, L, C elements. You can instead solve a much simpler set of algebraic equations relating the voltage and current phasors. This is what the phasor solution method does. As its name implies, this method computes voltages and currents as phasors. The phasor solution method is particularly useful for studying transient stability of networks containing large generators and motors. In this type of problem, we are interested in electromechanical oscillations resulting from interactions of machine inertias and regulators. These oscillations produce a modulation of the magnitude and phase of fundamental voltages and currents at low frequencies (typically between 0.02 Hz and 2 Hz). Long simulation times are therefore required (several tens of seconds). The continuous or discrete solution methods are not appropriate for this type of problem.

In the phasor solution method, the fast modes are ignored by replacing the network differential equations by a set of algebraic equations. The state-space model of the network is replaced by a complex matrix evaluated at the fundamental frequency and relating inputs (currents injected by machines into the network) and outputs (voltages at machine terminals). As the phasor solution method uses a reduced state-space model consisting of slow states of machines, turbines and regulators, it dramatically reduces the required simulation time.

Continuous variable-step solvers are very efficient in solving this type of problem. Recommended solvers are ode15s or ode23tb with a maximum time step of one cycle of the fundamental frequency (1/60 s or 1/50 s).You must keep in mind however that this faster solution technique gives the solution only in the vicinity of the fundamental frequency.

How SimPowerSystems Works

Simulating with Continuous Integration Algorithms