August 5, 2015 1 Robert Broadwater [email protected] Model-Centric Smart Grid for Big Data.

August 5, 2015

1

Robert [email protected]

Model-Centric Smart Grid for Big Data

Utility Data Sets

2

1-Analysis models (load forecasting, transmission, radial distribution, heavily meshed network, power flow, fault, reliability, transient, outage management)2-Substation models3-GIS 4-Renewable generation5-Device settings (control settings, protective devices)6-Customer load (monthly, demand, AMI)7-SCADA/EMS/PMU data8-Outage data9-As-Is Equipment List10-Weather Data…..

Model-Centric Smart Grid

• The model-centric approach employs a holistic, construction detail, model of the physical system – “Integrated System Model (ISM)”

• All measurement data, including weather data, is related to the ISM– Changes paradigm of “pushing data to algorithms”

to “pushing algorithms to data”

ISM Features • Model which includes everything needed for

simulating scenarios that engineers, operators, and field personnel talk about

• Allows any data/measurement set to be attached• Allows any calculation to be attached• Different calculations may work together as a team• Community maintained and shared model

– Emerging problem to solve – use ISM– Move from “craftsman modeler” to “manufactured model”

• Proactive modeling with Circuit Server4

Integrated System Model

Merge different construction models together, relating all measurements

“Aha” understanding

Model for holistic solutions, not point or scenario based solutions

The Best Equivalent Is No Equivalent

6

Every model simplification leads to elimination of scenarios

Model-Based Decisions

7

Problem Domain

Solution DomainModel

Makes it possible to find a solution that satisfies all scenarios

Our ability to solve a problem depends upon the model we have to solve the problem

Point solutions or scenario based solutions; Alphabet soup of systems with scattered data

Finding Solutions for the Hard Problems

Big Data

Big Model

Big Analysi

s

Physical System Model

Analysis extends beyond that which is possible with Data Analytics

Model-Centric Smart Grid Equation

Performance Analysis + Economic Analysis + Lab Testing + Field Validation = Model-Centric Smart Grid

Reliability, Efficiency, Capacity, Protection, Controllability

Silo’ed Organizations with Disjoint Models

10

Suppose data sets contain terabytes?

ISM “Living Model” Organization

11

Push algorithms to data

Eyes of all experts on the same model

Moves modeling from “age of modeling

craftsman” to “manufactured models”

created and used by many

Incremental CBA

12

Phase-Balance(no capacitors)

Phase Balance for Time Varying Load

Capacitor Design(capacitors on local control)

Cap Design for TimeVarying Load

Auto Reconfiguration, Monte Carlo

Efficiency policy goal

Efficiency and energy reduction policy goals

Efficiency and energy reduction policy goals

Reliability goal at least cost

Base System(not optimized,

some capacitors)

CoordinatedControl

Coordinated Control

Distribution Automation(blue sky days)(storm conditions)

“Dependency Ordering” of Investments

Big Analysis: Algorithms Working Together

13

Power Flow

Restoration AnalysisProtection / Coordination

Load Estimation

Fault Analysis

Customer Load Data

Model Validation

SCADA Measurements

Monte Carlo Driven Reliability Analysis

Contingency Analysis

Outage Data

© Copyright Electrical Distribution Design, Inc. 2015 Confidential

Matrix Analysis with Edge-Node Graph

Graph Trace Analysis for the ISM

14

2

1

5 4

2

3

1 3

4

1

5 4

2

3

Transform

Graph Trace Analysis with Edge-Edge Graph

• Edge knows neighboring edges• Topology continuously maintained• Algorithms with topology iterators • Write KVL and KCL directly• Processing time required for

configuration changes is independent of system size

Edges

1 2 3 4 5

Nodes

1 1 0 1 0 -1

2 -1 1 0 0 0

3 0 -1 -1 1 0

4 0 0 0 -1 1

Computer Processing

TopologyIterators

Global View Local View

Common Analysis Architecture

15

App 1

App 1 Model

App 2

App 2 Model

Core Models

Topology Management

Creation of simplified models

Pushing measurement data

Point Solution

Interfaces?

SCADA Data

Customer Load Data

ISM Analysis Architecture

16

ISM edge-edge topology

SCADA Measurements

App 1

App 2

Mass Storage

Memory

ISMCustomer Loads

Weather Measurements

Topology iterators, sharing of results, measurements

Interface provided by ISM to applications

ISM Model Management for Distributed Computation Environment

17

ISM Model Server

Client: Fault Location

Client: Reconfiguration

Model Queue

Analysis Processes

Supports distributed computations

Summary: Some Uses of ISM

• Effects on transmission system of high renewable penetration at distribution level

• Automated renewable generation screening analysis• Weather dependent load forecast that takes into account

renewable generation forecast• Storm outage prediction with radar-weather data, …• System reliability from analysis team of Monte Carlo, Power

Flow, and Restoration• Distribution solutions versus transmission solutions• Time series driven, CBA of smart grid investments

18

Generic Programming Roots of GTA

19

Container with Iterators CS Algorithms Generic Programming

Algorithms that process edges or components of graph

ISM with Topology Iterators Engineering Algorithms

Graph Trace Analysis

Generic analysis independent of system type - electric, gas, fluid, etc.

Algorithms that process objects in container, independent of object type