A Smart Mobile Solution with Novel MCP, Peak Demand and Hourly Grid GHG Emissions with Multiple Integration Points
G.Michor1; G.Doucett1; A.Sahebalam, PhD2 ; C.Ding, PhD3
1Screaming Power, 2Ryerson University, Screaming Power, SOSCIP, 3Ryerson University, SOSCIP
Class-A customers can manage their Global Adjustment (GA) costs by reducing demand during peak periods. These customers pay GA based on their percentage and contribution to the top five peak Ontario demand hours. These customers need to anticipate the top peak hours to be able shift their demand and as a result reduce their consumption at the opportune time.
There are several methods to anticipate the top peak hours:
• Time of year: Some part of peaks occurs during a sustained heat wave in the summer or cold snap in the winter.
• Time of day and days of the week: Peaks tend to occur on weekdays when businesses are in operation
• Specific tools on the IESO’s website: On the Class-A global adjustment page at www.ieso.ca/peaktracker, customers will find the current day and week-ahead demand.
The Class-A customers need a smart and automated tools to predict and estimate the possible peak at the right time and to help them make best decision to shift their loads to the best time of day. Indirect emissions (e.g., electricity, district steam, district heating, district cooling, etc.) are one of the main areas of CO2e emissions. Electricity is one of the most significant sources of emissions; therefore, it is essential to calculate and report these emissions correctly. To accurately calculate CO2e emissions, a clear understanding of standards, their relationships, the similarities / dissimilarities among them is essential. In reducing and–or managing CO2e emissions, we need to measure and find ways for objective comparison, but there are challenges. The forecasting of CO2e emissions is critical in preventing the impact of climate change. Understanding how to reduce emissions, as it can be directly related to energy conservation and energy management, will save energy and in turn money.
Methods and Materials
Market Clearing Price (MCP) & Peak Tracker
•Combinational Model: ARIMA (AutoRegressive Integrated Moving Average) and ML (Machine Learning) regression
•Training dataset and Model are updated every 5 mins
•Threshold updates are dynamic and patterns are changed
•Different patterns are dependent on demand and weather
•Market load and price are forecasted every 5 minutes
•Customers easily manage Global Adjustment exposure
•Customer can implement demand reduction strategies
•Consumers can avoid high energy prices
•Forecasts missing / late data from IESO
•Help customers react to price signals
•Creates future trends for price
•Provides dynamic Top 10 Peaks’ Demand Region
Live CO2e Prediction/Forecasting
•Ontario and Market Generation Output mix
•The hourly generator energy output and capability report for generating facilities in Ontario
•The IESO’s hourly total energy and operating reserve scheduled, Market demand and related historical reports.
•The hourly energy consumption
•The hourly historical weather data
The CO2e Intensity is replaced with an Intensity function. This Intensity is a function of grid generation and monthly historical CO2e inside the grid:
Where Κi is a coefficient for fuel type i; αi ,t is the portion of generated power by fuel i at time t ; βi is monthly emissions per energy production for fuel type i which is obtained from historical data.
Forecasted Upper and Lower Intensities:
Load Shift Recommender:
Live Decision Maker (DM) for shifting the load to the best time of the day:
• DM solves an optimization problem (dynamically) and finds the best time to shift load while minimizing a cost function
• DM also provides optimal load (cooling, heating, …) and optimal amount for shifting
• DM results in optimal savings
• User can also select a point (time) manually to shift the load while DM gives the cost for this point
• Track and forecast MCP with according to the dynamic threshold
• Track possible peak demand and send an alert one hour ahead.
• Recommender System for load shifting
• Predicted and forecasted CO2e with 95% Confidence
We present a new Smart Mobile Solution, platform and algorithm for forecasting and tracking Market Clearing Price (MCP) and Peak Demand (PD), and a novel live Decision Maker (DM) for load shifting. The Class A customers pay Global Adjustment (GA) based on their percentage and contribution to the top peak demand periods. A smart and automated tool can estimate the possible peak at the right time and help them to make the optimal decision to shift their loads to the best time of the day. The shift loader needs to know market price and can forecast the MCP, especially during peak time, to make responsible decisions and provide optimal recommendations. The developed platform solves a linear optimization problem while creating and updating a dynamic predictive model every five minutes to find the best time to shift loads. The platform also computes/predicts Hourly CO2e and Live CO2e, and forecasts emitted CO2e by the electrical power grid. In reducing and/or managing CO2e emissions, it is essential to measure and find ways for objective comparison of different standards, the presented method/solution minimizes the challenges in doing this.
“This research is funded by the SOSCIP TalentEdge Post-doctoral Fellowship Program in partnership with Ontario Centres of Excellence (OCE). Computations were performed on the SOSCIP Consortium’s Cloud Data Analytics and GPU computing platforms. SOSCIP is funded by the Federal Economic Development Agency of Sothern Ontario, the Province of Ontario, IBM Canada Ltd., Ontario Centres of Excellence, Mitacs and Ontario academic member institutions.”