Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

EMS Optimization Formulation

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Microgrid Energy Management System Development

Using Optimization-based Methods

= G

Jonathan LeSage, PhD


Senior Application Engineer - Energy and Automation
MathWorks, Inc.

© 2019 The MathWorks, Inc.


1
Demo – Microgrid EMS
Using Model-Based Design and Optimization

HEMS example – microgrid with loads

PV Panels:
- MPPT Control
- Power electronics

Decision Logic: Battery System:


- Store/draw power - Charge controls
- Use external grid - Discharge controls
- Optimize energy - Power electronics

2
Why energy storage with solar?

Adapted from: Smart EnergySystems Website


http://www.smart-energy.ag/products/ac-gekoppelte-speicherlosung-smartenergy-ac/?lang=en
3
Peak Demand Shift using Energy Storage

Adapted from: Smart EnergySystems Website


http://www.smart-energy.ag/products/ac-gekoppelte-speicherlosung-smartenergy-ac/?lang=en
4
Implementation of Energy Management Logic

Ref: Liu 2011 - A Hybrid AC/DC Microgrid and Its Coordination Control
5
Factoring in Variable Electricity Cost

Adapted from: Smart EnergySystems Website


http://www.smart-energy.ag/products/ac-gekoppelte-speicherlosung-smartenergy-ac/?lang=en 6
Optimization-based Peak Demand Shift

▪ Objective function: what do we want to minimize/optimize?


– Minimize total cost of variable priced electricity

𝐶𝑡𝑜𝑡 = ෍ 𝐶𝑔𝑟𝑖𝑑 (𝑘) ∙ 𝐸𝑔𝑟𝑖𝑑 (𝑘)


𝑘=0

▪ Constraints: how does our system operate?


– Power input/output to battery

𝐸𝑏𝑎𝑡𝑡 𝑘 = 𝐸𝑏𝑎𝑡𝑡 𝑘 − 1 + 𝑃𝑏𝑎𝑡𝑡 (𝑘)∆𝑇


– Power balance

𝑃𝑝𝑣 𝑘 + 𝑃𝑔𝑟𝑖𝑑 𝑘 + 𝑃𝑏𝑎𝑡𝑡 𝑘 = 𝑃𝑙𝑜𝑎𝑑 𝑘

Sources Loads
7
Linear program-based Optimization

▪ Standard form for the Linear program (LP) - linprog

𝐴∙𝑥 ≤𝑏
𝑇
min. 𝑓 𝑥 such that
𝑥 𝐴𝑒𝑞 ∙ 𝑥 = 𝑏𝑒𝑞

▪ Define states (x) necessary for LP optimization


– Pgrid(1:N) – Power from grid used from time step 1 to N
– Pbatt(1:N) – Power from battery
– Ebatt(1:N) – Energy stored in battery

𝑇
𝑥 = 𝑃𝑔𝑟𝑖𝑑 (1: 𝑁) 𝑃𝑏𝑎𝑡𝑡 (1: 𝑁) 𝐸𝑏𝑎𝑡𝑡 (1: 𝑁)

8
Linear program-based Optimization
▪ Equivalent constraint
beq
Aeq

𝑃𝑙𝑜𝑎𝑑 (1: 𝑁) − 𝑃𝑝𝑣 (1: 𝑁)


𝐼𝑁×𝑁 𝐼𝑁×𝑁 0𝑁×𝑁
x= 𝐸𝑏𝑎𝑡𝑡 (1)
𝑂𝑁×𝑁 Υ𝑁×𝑁 Φ𝑁×𝑁
0𝑁−1
0 0 0 1 0 0
Υ3×3 = ∆𝑇 0 0 Φ3×3 = −1 1 0
0 ∆𝑇 0 0 −1 1
▪ Inequality constraints
A b

𝑂𝑁×𝑁 𝐼𝑁×𝑁 𝑂𝑁×𝑁 𝑃𝑚𝑎𝑥


𝑂𝑁×𝑁 −𝐼𝑁×𝑁 𝑂𝑁×𝑁 −𝑃𝑚𝑖𝑛
x≥
𝑂𝑁×𝑁 𝑂𝑁×𝑁 𝐼𝑁×𝑁 𝐸𝑚𝑎𝑥
𝑂𝑁×𝑁 𝑂𝑁×𝑁 −𝐼𝑁×𝑁 −𝐸𝑚𝑖𝑛
9
10
Policy Comparison

Heuristic Optimized

11

You might also like