Tool Demonstration: Demand Forecasting PACE D 2.0 RE Team Agenda - - PowerPoint PPT Presentation

Tool Demonstration: Demand Forecasting

PACE D 2.0 RE Team PARTNERSHIP TO ADVANCE CLEAN ENERGY DEPLOYMENT (PACE-D 2.0 RE) TECHNICAL ASSISTANCE PROGRAM April 2020

Agenda

• Demand Forecasting
• Demand Forecasting Tool: Why?
• About Tool
• Parameters Considers for Demand Forecasting
• Results
• Online Demo

Slide No. 2

Demand Forecasting

• Prediction of future energy demand requires an

intuitive and wise judgment

• The forecast needs to be revised at regular

intervals to take care of new policies and changes in socio-economic trends.

• The demand forecast is used as a basis for system

development, and for determining tariffs for the future.

• Over-forecasts lead to more plant than is

required – Unnecessary capital expenditure

• Under-forecasts prevent optimum economic

growth – Lead to installation of many costly and expensive to-run generators.

Long T erm Forecasting:

• Plays a fundamental role in

economic planning of new generating capacity and transmission networks.

• Spans over 5 to 20 years.

Medium T erm Forecasting:

• Used mainly for the scheduling
• f fuel supplies, maintenance

program, financial planning and tariff formulation

• Spans over 1 month to 5 years

Slide No. 3

Demand Forecasting Tool: Why?

Slide No. 4

About Tool

• The demand forecasting can

be performed at DISCOM level for all categories.

• The various consumer

categories, like residential, commercial, industrial etc., can be considered for forecasting. The methods that have been provided in the software to arrive at the best forecast values are: Univariate:

• CAGR
• Trend Analysis

Multivariate:

• Econometric Method
• ARIMA
• ANN

PEUM:

Decomposes the sales of electricity into its elemental component of consumption Slide No. 5

Parameters Considered for Demand Forecasting

• Demand is forecasted under two scenarios:

✓ Business As Usual ✓ Scenario with Drivers

Business As Usual Scenario with Drivers

• Based on the energy sales and econometric

data, the demand is forecasted for all the consumer categories.

• CAGR, Trend, and Econometric for long

term forecasting

• ARIMA and ANN for medium term

forecasting

• In this impact of drivers is considered on

BAU scenario to forecast the demand.

• Drivers: Open Access (OA), Captive Power

Plants (CPP), Distributed Energy Sources (DER), and Electric Vehicles (EVs).

Further, sensitivity and probabilistic analysis is done to study the variation in demand.

Slide No. 6

Results: Long Term Forecasting (APDCL)

5000 10000 15000 20000 25000 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 NET DEMAND (MU)

APDCL: Net demand requirement in MU

Net demand 1000 2000 3000 4000 5000 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 PEAK DEMAN (MW)

APDCL: Peak demand in MW

Peak demand

• Long term demand

forecasting: 2020 to 2040

• Based on the average load

factor of previous 3 years, the peak demand is estimated.

• On an average, % deviation
• f demand projections w.r.t

energy sales approved under AERC MYT Order 2018 is 4.8%.

Source: BAU Report

Slide No. 7

Results: Medium Term Forecasting (APDCL)

200 400 600 800 1000 1200 Mar Dec Sep Jun Mar Dec Sep Jun Mar Dec Sep Jun Mar Dec Sep Jun Mar Dec Sep Jun 2024 2023 2022 2021 2020

APDCL-Monthly net demand in MU

Net demand 500 1000 1500 2000 2500 3000 Mar Dec Sep Jun Mar Dec Sep Jun Mar Dec Sep Jun Mar Dec Sep Jun Mar Dec Sep Jun 2024 2023 2022 2021 2020

APDCL-Monthly peak demand in MW

Peak demand

Source: BAU Report

Slide No. 8

Results: Hourly Load Profiles (APDCL)

Source: BAU Report

The load profile for the day of each month having peak demand is shown for the year 2020.

Slide No. 9

Probabilistic Analysis

6000 7000 8000 9000 10000 11000 12000 13000 14000 15000 16000

• 100%
• 80%
• 60%
• 40%
• 20%

0% 20% 40% 60% 80% 100% Total Energy Sales (MU) Variation in the Standard Deviation of Independent Variables (%)

Probabilistic Energy Sales at Varying Standard Deviation of Independent Variable for the Year 2030.

Risk Based Resource Plan Identification

Source: Probabilistic Analysis Report

Slide No. 10

Tool Highlights: Configuration of DISCOM

• The DISCOM and associated

consumer categories can be configured as a one-time activity.

• The historical energy sales
• bserved for each consumer

category can be uploaded into the tool.

• The SCADA data can be

directly imported into the tool for capturing the hourly load profile and the load factor observed.

Slide No.11

Tool Highlights: Scenario Creation

Several scenarios can be created in the tool to analyse various aspects and carry out sensitivity studies to understand the impact of various policies and drivers on the total demand.

Slide No. 12

Tool Highlights: Forecast Results

The results obtained for each category by different forecasting methods can be visualized both graphically and in tabular form to identify the most suitable forecast results

Results obtained for LT4 Commercial

Slide No. 13

Training Videos

• Introduction to Tool
• Pre-requisites
• Logging-in and

DISCOM configuration Getting Started

• Configuration of
• Dependent

Variables

• Independent

Variables

• Load Profile
• T&D Losses

Data Modeling

• Forecast Methods
• Scenario-specific data

configuration

• Execution

Scenario Creation

• View & Analyse

Results Summary

• Category-wise Fitted

Curve

• Consolidated Results
• Detailed PDF Report
• Probability Analysis

Analysis of results

• Policy Configuration
• Drivers Configuration

➢ Distributed Energy Resources ➢ Open Access ➢ Captive Power Plants ➢ Electric Vehicles

Impact of Policies & Drivers 1 2 3 4 5

Slide No. 14

Brief Demonstration of the Tool & Discussion

Slide No 15

Video Tutorials Link Getting Started

https://drive.google.com/drive/folder s/1mE35s7G7X7- z4Oa4q_jsPelUnUsqM9kq?usp=sharin g

Video Tutorial Link for first video is provided

Slide No 16

The video can be streamed on your respective laptop and mobile..

Your Feedback, Questions are Welcome…

Thanks!