Energy System Control Framework Design
Budget: €8 – €30 EUR
This project proposes designing and implementing a Model Predictive Control (MPC) framework for an energy system comprising batteries, a fuel cell, an electrolyzer, and a hydrogen tank. The goal is to maximize profits by optimizing energy flows while meeting operational constraints and responding to external disturbances such as grid pricing and renewable energy variability. The project will be developed and simulated using MATLAB/Simulink.
System Overview
The energy system under consideration consists of:
• Grid Connection: Allows for buying and selling energy.
• Battery Storage: Acts as a short-term energy buffer.
• Fuel Cell: Converts hydrogen into electricity during demand peaks (Smooth behavior)
• Electrolyzer: Produces hydrogen using excess electricity.
• Hydrogen Tank: Stores hydrogen for long-term energy needs.
Model Predictive Control Framework
Controllers
• A centralized MPC controller will coordinate the operation of all subsystems.
• Decentralized controllers may be used for local management of individual components.
System Model
• The system model will include:
o Battery Dynamics: State of Charge (SoC), charging/discharging power limits, efficiency losses.
o Electrolyzer: Power consumption, hydrogen production efficiency.
o Fuel Cell: Hydrogen consumption, power generation efficiency.
o Hydrogen Tank: Storage dynamics, capacity constraints.
Process and Disturbances
• Process: Energy flow between the grid, batteries, fuel cells, electrolyzers, and hydrogen tank.
• Disturbances:
o Variability in renewable energy generation.
o Time-varying electricity prices.
o Load demand fluctuations.
Variables
• Controlled Variables (Outputs):
o Grid power exchange (buy/sell).
o State of Charge (SoC) of the battery.
o Hydrogen level in the tank.
• Manipulated Variables (Inputs):
o Battery charging/discharging power.
o Fuel cell power output.
o Electrolyzer power input.
o Grid power (buy/sell decision).
Reference
• Meet the energy demand while maximizing economic profits.
• Maintain SoC and hydrogen levels within operational limits.
Prediction Horizon (N): The prediction horizon will be chosen to capture relevant dynamics, typically 24-48 hours with a sampling time of 1 hour.
Control Horizon (Nu): The control horizon will be shorter than the prediction horizon, typically 6-12 hours, to balance computational complexity and performance.
Methodology
System Modeling:
• Develop dynamic models for all subsystems in MATLAB/Simulink.
• Incorporate constraints such as capacity limits, efficiency losses, and power ratings.
MPC Formulation:
• Define the cost function to maximize profits (e.g., revenue from selling power - costs of buying power and operation).
• Implement operational constraints in the optimization problem.
Simulation:
• Simulate various scenarios such as high/low renewable energy generation, peak/off-peak electricity prices, and different load profiles.
• Evaluate the performance of the MPC in terms of profit maximization and operational efficiency.
Validation:
• Compare MPC performance with traditional control strategies (e.g., rule-based control).
• Analyze system behavior under disturbances to ensure robustness.
Expected Outcomes
• A functioning MATLAB/Simulink model of the MPC-based energy storage system.
• A detailed analysis of the MPC performance, including:
o Economic profits under various scenarios.
o Efficiency of energy storage components.
o Robustness to disturbances.
• Other penalizations you consider
System Overview
The energy system under consideration consists of:
• Grid Connection: Allows for buying and selling energy.
• Battery Storage: Acts as a short-term energy buffer.
• Fuel Cell: Converts hydrogen into electricity during demand peaks (Smooth behavior)
• Electrolyzer: Produces hydrogen using excess electricity.
• Hydrogen Tank: Stores hydrogen for long-term energy needs.
Model Predictive Control Framework
Controllers
• A centralized MPC controller will coordinate the operation of all subsystems.
• Decentralized controllers may be used for local management of individual components.
System Model
• The system model will include:
o Battery Dynamics: State of Charge (SoC), charging/discharging power limits, efficiency losses.
o Electrolyzer: Power consumption, hydrogen production efficiency.
o Fuel Cell: Hydrogen consumption, power generation efficiency.
o Hydrogen Tank: Storage dynamics, capacity constraints.
Process and Disturbances
• Process: Energy flow between the grid, batteries, fuel cells, electrolyzers, and hydrogen tank.
• Disturbances:
o Variability in renewable energy generation.
o Time-varying electricity prices.
o Load demand fluctuations.
Variables
• Controlled Variables (Outputs):
o Grid power exchange (buy/sell).
o State of Charge (SoC) of the battery.
o Hydrogen level in the tank.
• Manipulated Variables (Inputs):
o Battery charging/discharging power.
o Fuel cell power output.
o Electrolyzer power input.
o Grid power (buy/sell decision).
Reference
• Meet the energy demand while maximizing economic profits.
• Maintain SoC and hydrogen levels within operational limits.
Prediction Horizon (N): The prediction horizon will be chosen to capture relevant dynamics, typically 24-48 hours with a sampling time of 1 hour.
Control Horizon (Nu): The control horizon will be shorter than the prediction horizon, typically 6-12 hours, to balance computational complexity and performance.
Methodology
System Modeling:
• Develop dynamic models for all subsystems in MATLAB/Simulink.
• Incorporate constraints such as capacity limits, efficiency losses, and power ratings.
MPC Formulation:
• Define the cost function to maximize profits (e.g., revenue from selling power - costs of buying power and operation).
• Implement operational constraints in the optimization problem.
Simulation:
• Simulate various scenarios such as high/low renewable energy generation, peak/off-peak electricity prices, and different load profiles.
• Evaluate the performance of the MPC in terms of profit maximization and operational efficiency.
Validation:
• Compare MPC performance with traditional control strategies (e.g., rule-based control).
• Analyze system behavior under disturbances to ensure robustness.
Expected Outcomes
• A functioning MATLAB/Simulink model of the MPC-based energy storage system.
• A detailed analysis of the MPC performance, including:
o Economic profits under various scenarios.
o Efficiency of energy storage components.
o Robustness to disturbances.
• Other penalizations you consider