Renewable Integration and Smart Grid Technologies

Paul Scott

ANU/NICTA PhD Student

Introduction

The distributed and variable nature of renewables is a great challenge for classical electricity networks
Technologies like storage and demand response will enable change
Optimisation is essential for network design and operation

Overview

The Network
Voltage Regulation
Balancing Supply/Demand
Distributed OPF

The Network

Transmission

Meshed topology
Sensors and control
13% of retail costs in NSW

Distribution

Tree topology
Oversized to be robust
45% of retail costs in NSW

Lines

\[\begin{aligned} S &= VI^* \end{aligned} \] \[\begin{aligned} p_i &= g v_i^2 - g v_i v_j \cos(\theta_i - \theta_j) - b v_i v_j \sin(\theta_i - \theta_j)\\ q_i &= -b v_i^2 + b v_i v_j \cos(\theta_i - \theta_j) - g v_i v_j \sin(\theta_i - \theta_j) \end{aligned} \]

Safe Operation

Voltage bounds: \( \underline{v} \leq v \leq \bar{v} \)
Thermal limit: \( p^2 + q^2 \leq \bar{s}^2 \)
Frequency bounds: \( \underline{\omega} \leq \omega \leq \bar{\omega} \)
Transient stability
Oscillatory stability

Voltage Regulation

Distributed Generation

Distribution network was designed for flows in one direction
High PV penetration can upset this assumption
If controlled they can provide a significant opportunity
Inverters with adjustable power factors

Inverters

PV Demo

Inverter Control

Balancing Supply/Demand

Renewables

Conventional

Controllable
Centralised
~300 generators
Wholesale market

Renewables

Variable output
Distributed
Millions of generators
Wholesale/retail market

Conventional Dispatch

Generators submit bids to supply power
Every 5min AEMO dispatches generators
Cheapest generation dispatched first
Generators get paid at spot market price
NEM has ~300 generating units

Dispatch

Demand Response/Storage

Challenges

Incentives
Cooperation
Strong time coupling
Consumer privacy
Scalability

Distributed OPF

Replace function of dispatch
For micro or conventional grids
Receding time horizon
Ensure voltage and power within limits
Provides prices signal down to households
Incentives for distributed demand response/storage
Scalable

Distributed OPF

OPF Demo

Optimisation at NICTA

Optimisation Platforms
Constraint Programming
MILP, MINLP
Artificial Intelligence
Scheduling
Routing
Planning
Diagnosis

\[ \begin{align} &\min_x f(x)\\ &\text{ s.t. } g(x) \leq 0\\ %&x\in \mathbb{R}^n\\ %&f: \mathbb{R}^n \longrightarrow \mathbb{R}\\ %&g: \mathbb{R}^n \longrightarrow \mathbb{R}^m\\ \end{align} \]

Energy Systems at NICTA

Tractable power flow models
Self-healing networks
Microgrid simulation
Demand response
Power system restoration following disaster
Optimal line switching
Optimisation with transient stability assurances
PV forecasting (ML)

Links

Paul Email

Paul Homepage

NICTA Energy Systems