Advanced Load Forecasting & Methodology
27 - 31 May 2024
Sandton
Johannesburg South Africa
Cost per Delegate
R19,999.00
Overview
This course will look at the latest statistical and mathematical forecasting frameworks used by industry practitioners to tackle and manage the issue of load forecasting that cuts across short and long time scales.
The course will discuss practical applications to solving forecasting challenges, supported by real life examples from large control areas. The weather impacts on the load forecasts and the methodologies employed to quantify the weather effect while building a repository of weather normal data will be addressed.
A good load forecast methodology must improve its forecasting accuracy and support a consistent load forecasting process. Load forecasting methods widely used in the power industry has evolved significantly with modern day advancement with the adoption of Artificial Intelligence techniques such as Machine Learning.
With the increased penetration of inverter-based resources, the operation of electric grids has grown in complexity, leading to load forecasts which are updated more frequently than once a day. Furthermore, several jurisdictions have adopted a smaller granularity than the hourly load forecasts in the effort to reduce the forecasting uncertainties.
On the generation side, fuel forecasting professionals must meet energy requirements while making allowance for the uncertainty on both demand and supply.
Training Objectives
Upon completion of this course, the participants will be able to:
• Review the current approaches to all aspects of load forecasting concepts and methods.
• Get equipped on market segmentation and econometric framework for long-term forecasting.
• Develop expertise to build weather normal repositories.
• Explore Artificial Neural Networks and Probabilistic forecasting methods to manage forecasting uncertainties in short timeframes.
• Identify the main drivers in load forecasting and develop necessary mechanisms to assess new factors or how changes in the weather patterns affect the load profile and its forecasting uncertainty.
• Evaluate the impact of cross sectoral growth e.g. electrical vehicles.
• Learn about the current methodologies adopted by large power companies worldwide.
Target Audience
• Energy Forecasting Professionals
• Energy Planners and Energy Outlook Forecasters and Plant Operators
• Fuel Procurement Professionals
• Planners and Schedulers of Thermal Generating Units
Bulk Electricity System
• NIST Domains
• Generation System with Renewables and Storage o Base Load Generation
- Peaking Plants (Gas, Coal generating units)
-Ramping Rates
• Transmission System and Congestion Management Transactive Energy Model
-Prosumer
• Examples of Load Profiles
• Standards and Requirements
Load Forecasting
• Load Profiles
• Coincidental Peak Load
• Forecasting Methodologies
• Load Forecast Uncertainty
• MAPE Index
• Time Scale: Short-Term Load Forecasting
• Time Scale: Long-Term Forecasting
Industry Applications
• Methodology Adopted by CAISO
• Load Forecasting in NERC, NPCC region
• ITron - Metrix ND
Weather Normalization and Coincident Peak (PJM)
• Weather Feeds
• Weather Normal
• Data Data Cleaning
• Weather Normal Data compared to Actual Load
Short-Term Forecasting Methodologies
• AI Applications and Machine Learning for Load Forecasts
Load Forecast Algorithms
• Like Day or Similar Day Lookup Algorithms
• Rotation Algorithms
• Decision Tree Regression
• Forecast Algorithm Summary
Univariate Frameworks
• Lag Structures and Load Forecast Equations
• Moving Average
• Exponential Smoothing
• Exponential Smoothing Working Example
Multivariate Frameworks
• Regression
• Neural Network Models
• Support Vector Regression
• Artificial Neural Networks – Machine Learning
Explanatory Variables Based on Calendar Conditions
• Calendar Conditions
• Sun Rise and Sun Set
• Non Linear Response Load and Weather
• Daily versus Coincidental Hourly
Long Term Forecasting Methodologies
• Econometric Models
• Key Assumptions
• Market Segmentation
• Methodology
Load Forecasting Processes
• Annual Gross Energy Demand Forecast End-Use Forecasting Model (EUF)
• EUF Modules
• EUF Market Segmentation
• EUF Energy Usage
• EUF Customer Growth
• EUF Equipment Choice
• Scenario Forecasting
• Calibration, Consultation and Professional Judgement
• Hourly Gross and Hourly Net Energy Demand Forecast
• Conservation Programs and Conservation Regulations
• Industrial Conservation Initiatives Embedded Generation
• Grid Level Demand Base Year - New
• Demand Simulation Weather Model Methodology for the Base Year
• Net Level Demand Base Year
Weather Impacts on Load Forecasting
• Weather Response
• Weather Derivatives
• How Can Data Be Used for Decision-making?
• What Data Will Be Needed in the Future?
• What Data should be collected(Advanced)?
Power Electronics
• Capabilities of Smart Inverter
• Impact of BESS on Load Profile
Energy Market Scenarios
• Congestion Management
• Strategies for Coal or Gas Generation
Factors Influencing the Load Profile
• Electrical Vehicles
• Demand Response and Interoperability
• Controllers for Local Energy Networks
• Residential Energy Management Systems
• Financial Impacts of Smart Energy Technologies
System Operation with a Smart Meter Ecosystem
• Utility Data Analytics
• Industry Ecosystem
• Smart Meters and Advanced Metering Infrastructure (AMI) for Energy Storage
• Advanced Grid Controls and Sensors
• Accessible Energy Data for End-use Demand Response
• Innovative Policies and Adaptive Social infrastructure (still in development)
End of the Workshop
For Training arrangements call us on the detail below
TANZANIA: +255 749 50 26 78
SOUTH AFRICA: +27 694 31 79 73
KENYA: +255 749 50 26 78
DUBAI: +27 694 31 79 73
