FCS MPC for three phase inverter

Two-level three-phase inverter using Finite Control Set Model Predictive Control (FCS-MPC) with dead-time compensation.

This Simulink model demonstrates Finite Control Set Model Predictive Control (FCS-MPC) for a two-level three-phase Voltage Source Inverter (VSI) driving an LC filter and a Y-connected resistive load.

The controller operates in the stationary α–β reference frame and performs voltage reference tracking while accounting for dead-time effects in the switching devices.

The predictive algorithm evaluates all eight possible switching combinations of the inverter, forecasting the filter capacitor voltages and inductor currents at the next sampling instant. A cost function is then computed for each candidate state, considering voltage tracking accuracy, current smoothing, and switching minimization. The state with the lowest cost is applied to the inverter, ensuring fast dynamic response and low steady-state error.

An integrated dead-time compensation block models practical inverter switching behavior and corrects voltage distortion introduced by non-ideal transitions.

Key Features:

• Two-level three-phase inverter using predictive control
• LC output filter and Y-connected load for realistic power stage modeling
• Voltage control in the α–β (Clarke) stationary reference frame
• Dead-time insertion and compensation for accurate voltage reconstruction
• Discrete-time predictive model of the LC filter and load dynamics
• Cost function optimization combining voltage error, current ripple, and switching effort
• Fully parameterized model with adjustable sampling time, load values, and filter parameters
• Pure Simulink implementation — no toolboxes or external dependencies required

Applications:

• Predictive control research and algorithm benchmarking
• Inverter and converter voltage regulation
• Educational demonstrations of FCS-MPC and dead-time effects
• Development base for grid-connected or motor-drive systems

Included Files:

• FCS_MPC_3Phase_Inverter.slx — Main Simulink model with inverter, LC filter, and Y-load
• FCS_MPC_loop.m — Core predictive control algorithm (state prediction and cost evaluation)
• deadtime_insert.m — Dead-time insertion and compensation logic
• All system and control parameters are defined within the model for direct simulation

How It Works:

1. The controller predicts future values of the inductor currents and capacitor voltages for all possible switching vectors.
2. For each vector, a cost function evaluates the deviation between the predicted and reference voltages, current ripple, and switching transitions.
3. The optimal vector is selected and applied at the next control instant.
4. The dead-time module enforces non-overlapping switching in upper/lower IGBTs and compensates voltage distortion.
5. The result is accurate voltage tracking, reduced harmonics, and fast transient performance.