This book describes the stochastic and predictive control modelling of electrical systems that can meet the challenge of forecasting energy requirements under volatile conditions.
The global electrical grid is expected to face significant energy and environmental challenges such as greenhouse emissions and rising energy consumption due to the electrification of heating and transport. Today, the distribution network includes energy sources with volatile demand behaviour, and intermittent renewable generation. This has made it increasingly important to understand low voltage demand behaviour and requirements for optimal energy management systems to increase energy savings, reduce peak loads, and reduce gas emissions.
Electrical load forecasting is a key tool for understanding and anticipating the highly stochastic behaviour of electricity demand, and for developing optimal energy management systems. Load forecasts, especially of the probabilistic variety, can support more informed planning and management decisions, which will be essential for future low carbon distribution networks. For storage devices, forecasts can optimise the appropriate state of control for the battery. There are limited books on load forecasts for low voltage distribution networks and even fewer demonstrations of how such forecasts can be integrated into the control of storage.
This book presents material in load forecasting, control algorithms, and energy saving and provides practical guidance for practitioners using two real life examples: residential networks and cranes at a port terminal.
Author(s): William Holderbaum, Feras Alasali, Ayush Sinha
Series: Lecture Notes in Energy, 85
Publisher: Springer
Year: 2023
Language: English
Pages: 217
City: Cham
Contents
List of Figures
List of Tables
1 Introduction
1.1 A Potential Solution
1.2 Energy Storage
1.3 What Type of Control Methods This Book Presents
1.4 Demand Forecasting for Low Voltage (LV) Systems
1.5 Why Do We Need Control Which Utilise Demand Forecasts?
1.6 What Is the Book All About and What the Book Is Not About
1.7 Why Do We Need This Book
1.8 How to Read This Book
1.9 Note for a Semester Delivery Course
1.10 Acknowledgments
References
2 Short Term Load Forecasting (STLF)
2.1 Role of Load Forecasting in Power System
2.2 Introduction to the Load Forecasting
2.2.1 Time Series Analysis and Forecasting
2.2.2 Point and Rolling Forecasts
2.3 Notation and Load Forecasting Model Evaluation Methods
2.4 Statistical and Traditional Forecast Methods
2.4.1 Benchmark Models
2.4.2 AR, MA, ARIMA, ARIMAX and ARWD Methods
2.4.3 Vector Autoregressive (VAR)
2.5 Machine Learning for STLF
2.5.1 Artificial Neural Network (ANN)
2.5.2 Recurrent Neural Network (RNN)
2.5.3 Long Short Term Memory (LSTM)
2.6 Probabilistic Forecast Methods
2.6.1 ProbCast: Probabilistic Forecast Model and Visualisation Tool
2.7 Choosing a STLF Method
2.8 Comparison of Load Forecasting Techniques
2.9 Summary
References
3 Case Study: Low Voltage Demand Forecasts
3.1 Introduction
3.2 RTG Crane Demand Forecasts
3.2.1 Initial Experimental Design
3.2.2 Data Analysis
3.2.3 Model Selection
3.2.4 Results, Analysis and Discussion
3.3 Summary
References
4 Introduction to Control Strategies
4.1 Introduction
4.1.1 Closed and Open Loop Controllers
4.1.2 Control Engineering Strategies
4.1.3 Terminology
4.2 PID Controller
4.2.1 Theory for PID Controller
4.2.2 Discrete PID
4.2.3 PID Control Examples
4.2.4 Simple PID Tuning
4.3 Optimal Control
4.3.1 Notation
4.3.2 Mathematical Formulation
4.3.3 Optimization Examples
4.3.4 Global and Local Optimisation
4.3.5 Constrained and Unconstrained Optimisation
4.3.6 Continuous and Discrete Optimisation
4.3.7 Convexity
4.3.8 Linear and Nonlinear Optimisation
4.3.9 Stochastic and Deterministic Optimization
4.3.10 Multi-objective Optimisation
4.4 Optimization Algorithms
4.4.1 Types of Optimisation Algorithms
4.4.2 Initialisation
4.4.3 Stopping Criteria
4.4.4 Gradient Methods
4.4.5 Genetic Algorithm
4.5 Stochastic Optimisation
4.5.1 Approach 1: Mean Value
4.5.2 Approach 2: Mean Objective Function
4.5.3 An Example
4.5.4 Example for Residential Networks
4.6 A Comparison of the LV Control Strategies
References
5 Model Predictive Control
5.1 Introduction
5.2 Model Predictive Control
5.2.1 Mathematical Formulation
5.2.2 Sampling Rate
5.2.3 Horizon
5.2.4 Computational Complexity of MPC
5.2.5 Example Application: Autonomous Drone Control
5.3 Stochastic Model Predictive Control
5.3.1 Stochastic MPC: Mathematical Framework
5.3.2 SMPC Parameters
5.3.3 Example for Residential Networks
5.4 Summary: Advantages of MPC and SMPC
Reference
6 Case Study: Storage Control for LV Applications
6.1 Designing a Control System
6.2 Residential Network Demand
6.2.1 Initial Scoping
6.2.2 Data Collection and Analysis
6.2.3 Parameter Selection and Control Algorithm Design
6.2.4 Generate Forecast Model
6.2.5 Simulation Results
6.3 Case Study 2: Network of RTG Cranes
6.3.1 Initial Scoping
6.3.2 The Electrified RTG Crane Demand Data
6.3.3 Parameter Selection and Control Algorithm Design
6.3.4 Generate Future Demand Profile
6.3.5 Simulation Results
6.4 Summary
References
Appendix A Guided Walk-Through: Energy Storage Control Task
A.1 Analyze the Data
A.2 Determine the ESS Model Parameters
A.3 Develop a PID Controller for the ESS
A.4 Develop an Optimal Control for the ESS
A.5 Develop Model Predictive Control Model
Appendix B Further Reading
B.1 Forecasting Algorithms
B.2 PID Controllers
B.3 Optimal Control
B.4 Model Predictive Control
Appendix C Public Dataset and Model Repository
C.1 Probabilistic Forecast Models
C.2 Deep Learning Models
C.3 DataSets
Appendix D Probability: Definition and Properties
D.1 Types of Events
D.2 Independent Events
D.3 Disjoint or Mutually Exclusive Events
D.4 Additive Rules of Probability
D.5 Multiplicative Rules of Probability
D.6 Random Variables
D.7 Probability Distributions
D.8 Discrete Probability Distributions
D.8.1 Binomial Distribution
D.9 Continuous Probability Distributions
D.9.1 Normal Distribution
D.9.2 Standard Normal Distribution
Appendix E Elementary Statistics
References
Index
