EV Fleet Charging Management and V2G Planning (MATLAB, IEEE-33 Bus)

A complete, ready-to-run MATLAB project that optimizes EV fleet smart charging and Vehicle-to-Grid (V2G) dispatch on the IEEE-33 Bus radial distribution network. The scheduler uses linear programming (linprog) with LinDistFlow voltage constraints built from BIBC/BCBV matrices and validates results using a backward/forward sweep (BFS) load flow. No genetic algorithms are used.

Highlights

• Network-aware EV scheduling with voltage limits (0.95–1.05 pu)
• V2G-enabled: buy low/sell high under TOU pricing
• Soft voltage slacks to avoid infeasibility, with penalties in the objective
• Full feeder validation via BFS load flow (voltages and losses)
• Clear plots and command-window tables for quick interpretation
• 
  

  

  
  

What’s included

• EV_V2G_IEEE33_main.m
  • Main driver script
  • Builds network and load profiles, defines EV fleet and availability (home/work), TOU price
  • Calls the LP optimizer, runs BFS validation, prints summaries, and generates figures
  • Outputs: energy bought/sold, net energy cost, EV SoC compliance, per-bus energy, worst-bus voltage/time, losses

• build_ieee33_data.m
  • Provides IEEE-33 bus (Baran–Wu) feeder data
  • Line parameters in ohms, static P/Q loads per bus, bases (Vbase=12.66 kV, Sbase=100 MVA)
  • Returns clean, ready-to-use arrays for the rest of the scripts
  • make_BIBC_BCBV.m
    • Constructs BIBC (branch current mapping) and BCBV (bus voltage from branch currents) matrices for radial feeders
    • Computes per-unit branch impedances from ohmic data and base values
    • Handles parent/child relationships along the feeder to assemble path matrices
  • bfs_loadflow.m
    • Backward/Forward Sweep load flow for radial distribution systems
    • Computes complex bus voltages and total real power loss
    • Used hourly to validate the optimization results (voltages and losses)
  • schedule_ev_lp.m
    • Builds and solves the linear program with linprog (dual-simplex)
    • Decision variables: EV charge pch(i,t), discharge pdch(i,t), and optional voltage slack variables
    • Enforces:
      • Availability and power limits per EV (e.g., 7.2 kW L2, symmetric V2G)
      • SoC dynamics with efficiencies, capacity bounds, and target at last-available time
      • LinDistFlow voltage constraints via BIBC/BCBV sensitivities (soft bounds with penalties)
    • Returns schedules, voltage slack usage, and energy cost
    • System requirements
      • MATLAB R2020a or newer
      • Optimization Toolbox (linprog)
      • Windows/macOS/Linux, 8 GB RAM recommended
      How to run
      1. Place all .m files in one folder
      2. Open EV_V2G_IEEE33_main.m
      3. Press Run (reproducible via rng(42))
      4. View results in the command window and figures pop-up
      Key features and modeling choices
      • Network: IEEE-33 Bus radial feeder, per-unit conversion from 12.66 kV, 100 MVA base
      • Voltage model: LinDistFlow using S_P and S_Q sensitivities from BIBC/BCBV
      • EV fleet: Bus mapping to selected PQ buses; availability patterns (70% home nights, 30% workdays)
      • Power/energy: 7.2 kW charge/discharge per EV; ηch=ηdch=0.92; Ecap 50–70 kWh; target SoC = 80%
      • Pricing: 24-hour TOU tariff; easily editable
      • V2G: Discharging earns revenue (negative cost in objective)
      • Validation: BFS load flow per hour for voltages and feeder losses
      • No genetic algorithms: deterministic convex optimization with linprog
      Customization points
      • N_EV (number of EVs) and candidate_buses (connection buses)
      • Charging power limits, efficiencies, capacities, initial SoC ranges
      • Availability windows (home/work or custom patterns)
      • Voltage limits (Vmin/Vmax), soft slack penalty (lambda_v) and cap (vslack_max)
      • TOU price vector, time step dt, horizon T, and load shape loadMult
      • Replace IEEE-33 with your own feeder by editing build_ieee33_data.m
      Outputs and visuals
      • Command-window tables:
        • Energy bought/sold, net energy, total cost
        • EV SoC target compliance
        • Per-bus EV charge/discharge energy
        • Worst bus voltage and hour; losses summary
      • Figures:
        • TOU price vs. net EV power (charge/V2G)
        • Voltage profile at worst hour with Vmin line
        • Minimum bus voltage over time
        • Stacked bar of EV energy by bus
      Who is this for?
      • Researchers and students studying smart charging, V2G, and distribution networks
      • Utilities and aggregators exploring feeder-aware EV scheduling
      • Educators looking for a clean LP-based case study with power-flow validation
      Keywords
      MATLAB, EV charging, V2G, Vehicle-to-Grid, IEEE-33 bus, Baran–Wu, radial distribution network, LinDistFlow, BIBC, BCBV, linear programming, linprog, Optimization Toolbox, BFS load flow, TOU pricing, SoC management, distribution planning, smart grid, power systems, feeder constraints