Topic:"Forecasting Methodology" in M04819

tepayer and taxpayer funded programs is required for ENSC's annual financial statements; hence, that allocation by general ledger account is reviewed by ENSC's auditors. - 3. ENSC's CAM is used to establish the true-up adjustments that ens...

AI summary ENSC's cost allocation processes involve using a Cost Allocation Model (CAM) for true-up adjustments to ensure accurate recovery of ratepayer-funded program costs. Unlike regulated utilities, ENSC's CAM reflects unique characteristics, such as non-capital intensity and program-specific expenditures, while adhering to cost causality principles. The model is divided into two parts, with preliminary allocations based on planned costs rather than CAM inputs.

5.2 PRELIMINARY DSM RATE AND BILL IMPACTS Attachment 2 shows the potential impact on the annual DSM rate rider of the 2013-2015 DSM Plan by customer class. Since the 2012 DSM rate includes a true-up (balance adjustment) for 2010, the DSM r...

AI summary The document outlines preliminary DSM rate and bill impacts from the 2013-2015 DSM Plan, noting variability due to CAM allocation differences, expenditure reallocations, and future NSPI rate/load forecast changes. Attachments 2 and 3 detail rate rider impacts and bill effects by customer class, with caveats about preliminary budget estimates versus audited financial statements.

Conservativeness Bottom-up estimates tend to be quite conservative, notably because they don't capture all the small efficiency projects that we know have been occurring at NPPH. Top-down estimates tend to be close to reality because they...

AI summary Bottom-up estimates are more conservative as they miss small efficiency projects at NPPH, while top-down estimates use real data and are closer to reality. Estimates reflect avoided energy costs, not absolute savings.

Consistency In each step of our analysis we have endeavoured to maintain a high level of consistency. For example, when we generate regression equations for three different time periods, we make sure all three periods are based on the same...

AI summary The analysis emphasizes maintaining methodological consistency by using identical factors across three regression equations, even when statistical significance varies between periods. This approach ensures uniformity despite differing significance levels in individual equations.

on function reasonably approximates the actual real world situation. Figure 10 – Scatter plots of actual data vs linear regression model baselines, 3 reporting periods Figure 11 shows the actual reported values vs. those predicted by the m...

AI summary The analysis uses linear regression models to compare actual TMP plant energy consumption data with predicted values, showing the model's accuracy and indicating effective energy management by staff during Year 4. Figures highlight model overestimation and operational ranges, suggesting staff successfully managed energy performance.

8. CONCLUSIONS & RECOMMENDATIONS The top-down electricity savings estimate for the whole mill described in Section-5 of this Report is consistent with the M&V requirements defined in the SEP M&V Protocol and meets all the IPMVP criteria (a...

AI summary The top-down electricity savings estimate for the NPPH mill aligns with SEP M&V Protocol and IPMVP criteria, validating its accuracy. It closely matches the initial report but may overstate savings due to limited bottom-up data. The analysis concludes the top-down method provides a reliable energy savings representation.

1- Electric Space Heating Households - by channel and heat distribution systems As can be seen, the vast majority of Nova Scotia's electrically heated homes are not being reached by the existing comprehensive energy efficiency programs, an...

AI summary The text highlights that most electrically heated homes in Nova Scotia are not covered by existing energy efficiency programs, with many using zoned baseboard heating systems. Table 2 provides a qualitative assessment of market potential by technology and heat distribution system.

2 - Green Heating Systems Market Potential Target Technologies Zoned Heating (baseboard) Central Heating (furnaces, boilers) Solar Thermal Small Small Ground Source Heat Pumps (GSHP) marginal Small Ductless Air Source Heat Pumps (DHP) LARG...

AI summary The table shows market potential for various green heating technologies, with Ductless Heat Pumps (DHP) having large potential for zoned heating. While individual technologies have limited potential, their combined use offers significant electricity and GHG savings.

5.2 PRELIMINARY DSM RATE AND BILL IMPACTS Attachment 2 shows the potential impact on the annual DSM rate rider of the 2013-2015 DSM Plan by customer class. Since the 2012 DSM rate includes a true-up (balance adjustment) for 2010, the DSM r...

AI summary The document outlines preliminary impacts of the 2013-2015 DSM Plan on rates and bills, noting variations due to factors like the Cost Allocation Model (CAM), program cost reallocations, and NSPI forecast changes. Attachments 2 and 3 detail rate rider impacts and bill effects by customer class, with caveats about preliminary estimates.

4.4Period of Analysis, Energy & Conditions Since TMP refining energy (electricity) consumption represents a full two-thirds of the NPPH mill's electric consumption a first analysis was done on TMP refining energy. Because of its relative i...

AI summary The analysis focuses on TMP refining energy consumption at NPPH, which accounts for two-thirds of the mill's electricity use. Detailed records from power transmitters enabled statistical modeling of energy use relative to production, leading to a CUSUM chart for monitoring efficiency trends pre- and post-2009 energy projects.

Conservativeness Bottom-up estimates tend to be quite conservative, notably because they don't capture all the small efficiency projects that we know have been occurring at NPPH. Top-down estimates tend to be close to reality because they...

AI summary Bottom-up energy efficiency estimates are conservative as they exclude small projects at NPPH, while top-down estimates using real data are more accurate. Estimates reflect avoided energy costs over time, not absolute energy savings.

Consistency In each step of our analysis we have endeavoured to maintain a high level of consistency. For example, when we generate regression equations for three different time periods, we make sure all three periods are based on the same...

AI summary The analysis emphasizes maintaining consistency by using identical factors across three time periods when generating regression equations, even if some factors show varying significance levels (e.g., not highly significant in one equation but significant in others).

8. CONCLUSIONS & RECOMMENDATIONS The top-down electricity savings estimate for the whole mill described in Section-5 of this Report is consistent with the M&V requirements defined in the SEP M&V Protocol and meets all the IPMVP criteria (a...

AI summary The top-down electricity savings estimate for NPPH aligns with SEP M&V and IPMVP criteria, validating its accuracy. It closely matches the initial report but may be higher due to limited data for bottom-up analysis. The estimate is deemed a reasonable reflection of actual energy savings achieved.

Plug-in Lamp Bulbs Although there are considerably fewer plug-in lamp bulbs as compared to permanent light fixture bulbs, usage of plug-in lamp bulbs on a typical day is similar to the permanent light fixture bulb use pattern. Across Nova...

AI summary Plug-in lamp bulbs are less common than permanent light fixtures in Nova Scotia, with residents averaging six units. 45% have fewer than five, 39% have five to nine, and 16% have ten or more. Daily usage patterns mirror those of permanent fixtures despite lower prevalence.

II. The Individual Program Evaluations The developments of the Econoler estimates of energy savings and demand reduction attributable to each DSM program were carefully reviewed. As indicated above (Recommendation SV1), our opinion is that...

AI summary The document reviews Econoler's estimates of energy savings and demand reduction from DSM programs, finding them reasonable. It supports evaluator recommendations on impact assessment methods and results, excluding others like program marketing. Key focus is on validating energy savings estimates and endorsing specific evaluation approaches.

2011 Load Forecast Prepared April 2011

AI summary The 2011 Load Forecast, prepared in April 2011, outlines the projected electricity demand for Nova Scotia. It serves as a regulatory document for energy planning and decision-making processes.

Figure 1 Annual Net System Requirement In addition to annual energy requirements, NSPI also forecasts the peak hourly demand for future years. The forecast methodology uses forecast energy requirements and expected load shapes (hourly cons...

AI summary NSPI forecasts annual energy and peak demand using load shapes adjusted for customer changes. Net System Peak is projected to decline 1.5% annually until 2021, driven by DSM programs. Without DSM, growth would be 1.0% annually. Historical analysis informs load shape derivations.

6 7 1 3 NSPI annually develops a forecast of energy sales and peak demand requirements to assess the 4 effects of customer, demographic and economic factors on the future provincial system load. It is a fundamental input to the overall pla...

AI summary Nova Scotia Power Inc. (NSPI) annually forecasts energy sales and peak demand to evaluate the impact of customer, demographic, and economic factors on future provincial system load. This forecast, developed in the winter of 2010-2011, covers 2011 to 2021 and uses average annual growth rates calculated for that period.

Forecast Models 11 10 12 Nova Scotia electric energy sales are modeled and forecast as three provincial customer sectors: 13 residential, commercial and industrial. Energy forecasts for sector electricity sales are calculated using econome...

AI summary The document describes the use of econometric models to forecast Nova Scotia electric energy sales across residential, commercial, and industrial sectors. These models use variables such as population, GDP, energy prices, and heating degree-days, with data sourced from the Conference Board of Canada's Economic Outlook.

Energy Forecast Details 13 14 15 16 For forecasting, modeling and sales reporting, Nova Scotia electric load is divided into three sector requirements: residential, commercial and industrial. The relative sizes of sector sales are shown in...

AI summary The document outlines the division of Nova Scotia's electric load into residential, commercial, and industrial sectors for forecasting and reporting purposes, with Figure 4 illustrating the relative sizes of sector sales.

Residential Sector Sales 2 1 In 2010, residential customers represented approximately 37 percent of total Nova Scotia energy sales. In addition to direct domestic customers of the Company, the sector also includes residential customers ser...

AI summary Residential customers in Nova Scotia accounted for 37% of energy sales in 2010, including those served by municipal utilities. Forecasting uses econometric models considering appliance indexes, heating loads, and electricity costs, alongside population trends and market share analysis to predict demand.

Figure 5 Persons per Residential Account 2 3 Within the residential sector forecast, large household appliances are modeled individually, considering age, efficiency trends, and acquisition rates. Since these improvements apply only to new...

AI summary The residential sector forecast models large household appliances based on age and efficiency, leading to gradual load reductions as older models are replaced. Electric space heating saturation is around 29% (2010) and 59% for water heating, with slow growth. Heating degree-days (HDD) from Shearwater Airport (2000-2009 average: 4,010 HDD) are used for weather effects, with Figure 6 showing past HDD variations.

Figure 7 Annual Energy – Residential Sector 2 4 5 6 7 1 Growth in this sector is expected to be relatively low. The 2012 load forecast for this sector is 4,437 GWh representing a 1.3 percent annual increase over 2010 actual sales adjusted...

AI summary Residential sector energy growth is projected at 1.3% annually (2012 forecast: 4,437 GWh), with DSM reducing this to 2.1% (4,514 GWh). Nova Scotia Power Inc. is highlighted as the key organization involved in the analysis.

Commercial Sector Sales 8 9 10 11 12 13 14 7 Energy sales to the commercial sector in 2010 represented 29 percent of Nova Scotia sales. This customer group includes restaurants, hotels, offices, recreational facilities, stores warehouses h...

AI summary Energy sales to Nova Scotia's commercial sector in 2010 accounted for 29% of total sales, influenced by GDP and DSM programs. An econometric model using GDP, RPDI, residential sales, and prior commercial data forecasts sector demand, with DSM effects noted in 2008-2010.

Figure 9 Annual Energy – Commercial Sector 16 17 18 19 20 Growth in this sector has averaged 0.5 percent over the past 5 years (also 0.6 percent when adjusted for weather). Driven by trends in wholesale trade, consumer confidence, and grow...

AI summary The commercial sector's energy use has grown 0.5% annually over 5 years, driven by trade, consumer confidence, and disposable income. Forecasts predict 3,355 GWh by 2012, with DSM reducing load growth by 2.0% over 10 years (vs. 1.0% without conservation).

Industrial Sector Sales 8 9 10 - In 2010, the industrial sector represented 34 percent of Nova Scotia total electricity sales. This group is comprised of customers who process raw materials or manufacture finished goods. It includes both p...

AI summary In 2010, Nova Scotia's industrial sector accounted for 34% of total electricity sales, driven by manufacturing and resource industries. Large customers dominate consumption, with five major users accounting for two-thirds of sector energy use. Economic factors, DSM programs, and econometric models (using GDP and employment data) influence load forecasting. The 2009 sales drop reflected economic downturn impacts.

System Losses and Unbilled Sales - 28 The load forecast is developed using Nova Scotia Power "billed" sales rather than "accrued" - sales to provide a longer historical time series upon which to base the models. Billed sales refers - 30 to...

AI summary Nova Scotia Power Inc. (NSPI) uses 'billed' sales rather than 'accrued' sales for load forecasting due to the longer historical data availability. Billed sales reflect energy billed to customers, while accrued sales represent actual energy generated. System losses, including transmission/distribution losses (4% in municipal areas) and unbilled sales, are estimated at 6.6-6.7% of total Nova Scotia energy requirements.

Peak Demand 13 14 15 16 17 The total system peak is defined as the highest single hourly average demand experienced in a year. It includes both firm and interruptible loads and due to the weather-sensitive load component in Nova Scotia, th...

AI summary Peak demand in Nova Scotia is defined as the highest hourly average demand in a year, influenced by weather and customer behavior. DSM programs and price signals (e.g., ELI 2P-RTP) have reduced peak demand growth. The 2009/2010 peak was 124 MW lower than 2004 due to conservation and interruptions. Forecasts show a 1.5% annual decline in net system peak by 2021, attributed to DSM and conservation efforts.

- 5 usually close, due to the peak often being driven by cold temperatures. 1 Load Forecast 2 Appendices 3 4 5 1 Appenaix A 2 3 2010 NSPI Forecast 4 5 Residential Sector Econometric Model Detail

AI summary The document discusses the residential sector econometric model detail, focusing on load forecasting and appendices. The peak load is influenced by cold temperatures, and the document includes appendices with load forecasts from 2010.

$DOMENG = 302.4 \ AIDX + 0.2540 \ CHDD - 28.25 \ RREP + 0.1095 \ RRCGOODS + 0.4458 \ DOMENG_{-1}$ Forecast Model for DOMENG Dynamic regression

AI summary The document presents a dynamic regression forecast model for DOMENG, incorporating variables such as AIDX, CHDD, RREP, RRCGOODS, and the lagged value of DOMENG. This model is used to predict future values of DOMENG based on historical and current data.

Regression(5 regressors, 0 lagged errors) Term Coefficient Std. Error t-Statistic Percentile 2014 1.418 429 5,273 1339 14.05 -397 10,829 1,186 4,358.9 1,943 1.2 204 4,300 -1.4% 2015 1.404 425 5,351 1359 13.74 -388 10,877 1,191 4,299.6 1,91...

AI summary The text presents a regression analysis with five regressors and no lagged errors, showing coefficients, standard errors, t-statistics, and percentiles for various years from 2014 to 2021. The data appears to be related to energy demand or economic indicators, with statistical values decreasing slightly over time.

1 Commercial Sector Econometric Model Detail 2 3 4 5 $COMENG = 0.01906 \ RQTOS + 0.01362 \ RPDI + 0.2685 \ DOMENG + 0.4245 \ COMENG_{-1}$ 6 7 Forecast Model for ComEng 8 Regression(4 regressors, 0 lagged errors) 10 Coefficient Std. Error t...

AI summary The Commercial Sector Econometric Model for ComEng uses four regressors (RQTOS, RPDI, DOMENG, and lagged COMENG) with high statistical significance (t-statistics >3.0). The model achieves a near-perfect adjusted R-square (1.00) and low forecast error (MAPE 0.76%), though the sample size is small (30 observations).

Commercial Sector Model Fit using historical data Industrial Econometric Model Details 1 2 3 4 Small and Medium Industrial class models are shown below. 5 6 7 8 SM_{IND} = 0.01885 GDP_{Man} + 0.01278 NonRes_{Inv} + 0.7220 SM_{IND_{-1}} 9 1...

AI summary The document presents econometric models for Small and Medium Industrial sectors in Nova Scotia, using GDP, employment, and lagged variables. Models show high R-square values (0.98 and 0.95), low forecast errors (5.74 and 17.40), and statistical significance (p-values <0.01). These models aim to predict energy demand based on economic and employment indicators.

Industrial Model Input Variables and Contributions

AI summary The document examines input variables and their contributions to an industrial model, focusing on pricing mechanisms like ELI 2P-RTP and DSM, economic indicators (CPI, GDP), and energy demand factors (HDD, NSR). NSPI and OATT are highlighted as key entities influencing industrial energy pricing and demand management.

Figure 1 Annual Net System Requirement In addition to annual energy requirements, NSPI also forecasts the peak hourly demand for future years. The forecast methodology uses forecast energy requirements and expected load shapes (hourly cons...

AI summary NSPI forecasts annual energy requirements and peak hourly demand using historical load shapes adjusted for changes like new equipment. Growth in annual net system peak is illustrated in Figure 2, highlighting methodology reliance on energy forecasts and customer-specific load profiles.

T 4 4 • Intr utin ction 11111 vuu NSPI annually develops a forecast of energy sales and peak demand requirements to assess the effects of customer, demographic and economic factors on the future provincial system load. It is a fundamental...

AI summary NSPI annually forecasts energy sales and peak demand requirements to evaluate the impact of customer, demographic, and economic factors on future provincial system load. This forecast, developed in the winter of 2009, covers the period from 2009 to 2019 and is a key input for planning, budgeting, and operating activities.

Discussion of Major Inputs 1 2 The Gross Domestic Product (GDP) for Nova Scotia was estimated at $26,812 million (in constant 2002 dollars) in 2008, and is forecast to increase by just 1.0 percent in 2009 and 2.0 percent in 2010. This low...

AI summary The document discusses the economic outlook for Nova Scotia in 2008-2010, highlighting slow GDP growth, challenges in manufacturing, mining, and forestry sectors, reduced tourism, and the impact of the global economic downturn on various industries. It also includes forecasts for housing starts, retail sales, and population growth.

Residential Sector Sales 242526 27 28 29 In 2008, residential customers represented approximately 36 percent of total Nova Scotia energy sales. In addition to direct domestic customers of the Company, the sector also includes residential c...

AI summary Residential customers accounted for 36% of Nova Scotia energy sales in 2008, with seasonal residences comprising 6.5%. Modeling uses econometric methods, population forecasts, appliance saturation rates, and variables like RREP (Real Sector Customer Price per kWh) and CHDD (composite heating degree-days). Factors include economic trends (RCGOODS), price elasticity, and household growth due to urbanization.

Residential Sector Energy Residential Sector Residential Sector Year With DSM Growth Rate Without DSM Growth Rate GWh % GWh % 2001 3,741.2 1.9 3,741.2 1.9 2002 3,828.9 2.3 3,828.9 2.3 2003 4,010.5 4.7 4,010.5 4.7 2004 4,113.5 2.4 4,113.5 2...

AI summary The table presents residential sector energy consumption (GWh) with and without DSM programs from 2001-2019, showing fluctuating growth rates. Without DSM, consumption trends mirror with-DSM data but with slight variations. The text notes a projected 0.8% annual residential load decline (with DSM) versus 0.5% increase (without DSM) over the 10-year forecast period.

In 2008, the industrial sector represented 36 percent of Nova Scotia total electricity sales. This group is comprised of customers who process raw materials or manufacture finished goods. It includes both primary resource industries such a...

AI summary The industrial sector in Nova Scotia accounts for 36% of total electricity sales, with a few large customers consuming most of the energy. Load forecasting combines econometric modeling and customer data, accounting for economic factors like GDP and customer migration between rate classes.

Rate Class Sales 1 2 Forecast sales by sector are allocated into 13 rate classes for revenue forecasting purposes. The following section describes these rate classes and their expected energy requirements for the forecast period. In most c...

AI summary The document outlines the allocation of forecast sales by sector into 13 rate classes for revenue forecasting. Load growth trends and customer migration between classes influence energy requirements, which are calculated using historical data and migration patterns.

Medium Industrial - 30 This class is applicable to any industrial customer having a regular demand of at least 250 kVA, - 31 but less than 2,000 kVA. As of December 2008, there were 196 customers in this class, - 32 representing about 4.6...

AI summary The Medium Industrial class includes customers with 250–2,000 kVA demand, comprising 4.6% of NSPI sales. Sales are projected to decline 7.3% over 10 years without conservation and DSM programs, which could limit the decline to 1.2% growth.

Commercial Sector Econometric Model Detail 1 23 4 5 $COMENG = 0.02531 RQTOS + 0.01268 RPDI + 0.2806 DOMENG + 0.3676 COMENG_{.1}$ 6 7 Forecast Model for ComEng 8 Regression(4 regressors, 0 lagged errors) 9 10 Coefficient Std. Error t-Statis...

AI summary The document presents an econometric model for forecasting commercial energy demand (COMENG) using variables like RQTOS, RPDI, DOMENG, and lagged COMENG. The model shows high statistical significance (p<0.001) with R²=0.9977, low forecast error (26.08), and strong fit (MAPE=0.76%). It uses 30 data points with no lagged errors.

Industrial Econometric Model Details 1 2 3 4 Small and Medium Industrial Classes are summed and modeled together. 5 6 7 $IND = 0.015016 RQTOS + 0.52599 IND_{-1} - 35.513 MIGRATE$ 8 9 10 Forecast Model for IND 11 Regression(3 regressors, 0...

AI summary An econometric model forecasts industrial demand (IND) using RQTOS, lagged IND values, and MIGRATE. The model shows strong performance (R²=0.9888, MAPE=1.48%) with statistically significant coefficients. Small and medium industrial classes are aggregated for analysis.

Table A1: Energy Requirement – 2009 NSPI Forecast

AI summary Table A1 presents the 2009 energy requirement forecast by Nova Scotia Power Inc. (NSPI), incorporating variables like heating degree-days, consumer price index, and GDP to model energy demand across residential, commercial, and industrial sectors under different pricing structures.

Annual Net System Requirement, High and Low Scenarios

AI summary The document analyzes Nova Scotia's Annual Net System Requirement (NSR) under high and low scenarios, involving entities like NSPI and programs such as DSM. It references economic indicators (CPI, GDP) and energy pricing models (RTP, 2P-RTP) to assess system demand and cost implications.

1 2 3 4 2 3 4 5 6 7 8 9 10 11 12 13 13 15 16 17 18 19 20 21 Appendix D 22 23 Forecast Sensitivity by Major Variable 24 25 Appendix D: Forecast Sensitivity by Major Variable

AI summary Appendix D provides a forecast sensitivity analysis by major variable, likely examining how different factors influence future projections. This section is part of a regulatory proceeding document and is used to assess the impact of various variables on forecasts.

2006 IRP Load Forecast Prepared September 2006

AI summary The document titled '2006 IRP Load Forecast' was prepared in September 2006 as part of a Nova Scotia regulatory proceeding. It outlines load forecasting efforts, though specific details or analysis are not included in the provided text.

Commercial energy sales are forecast using the same models and assumptions as described in the 2006 NSPI Load Forecast report. For the IRP forecast period beyond 2015 shown in the Load Forecast Report, the following assumptions have been m...

AI summary Commercial energy sales are forecast using models from the 2006 NSPI Load Forecast report. The economic forecast from the Conference Board of Canada provides data up to 2025, and a 1.6% growth rate is applied to extend the forecast to 2029.

Industrial energy sales are forecast using the same models and assumptions as described in the 2006 NSPI Load Forecast report. For the IRP forecast period beyond 2015 shown in the Load Forecast Report, the following assumptions have been m...

AI summary Industrial energy sales are forecast using models from the 2006 NSPI Load Forecast report. Economic data from the Conference Board of Canada is used up to 2025, and a 0.8% growth rate is applied to extend the forecast to 2029.

• Industrial load: - reduced by 1,700 GWh annually from 2007 onward for the low case, the equivalent to closing a major paper mill. - increased by 500 GWh annually from 2008 onward for the high case, the estimated load of a major industria...

AI summary Industrial load projections show a potential reduction of 1,700 GWh annually from 2007 (equivalent to closing a major paper mill) and a possible increase of 500 GWh annually from 2008 (equivalent to a major industrial expansion or new industry). These scenarios reflect contrasting energy demand trajectories for Nova Scotia.

• Economic variables: - the annual growth rates of the major economic indicators used in the base forecast are reduced by 50 percent for the low case. This range was judged to be suitable and is within the variation observed in provincial...

AI summary The text discusses adjusting the annual growth rates of major economic indicators by 50% for low and high cases, aligning with recent provincial GDP variations.

- 5 ample opportunity for the UARB and stakeholders to ensure oversight and accountability. 1 Request IR-5: 2 3 Please provide ENSC's views of the impact of a three year program versus a one year 4 program on projected Total Resource Cost...

AI summary The text discusses ENSC's response to requests regarding the impact of a three-year program versus a one-year program on Total Resource Cost (TRC) and Program Administration Cost (PAC) tests for 2013-15. ENSC suggests that a three-year program would reduce costs and improve TRC and PAC ratios. Additionally, it references load forecasts from 2006 to 2011 and compares them with those from earlier Integrated Resource Plans (IRPs).

2011 Load Forecast Prepared April 2011

AI summary The 2011 Load Forecast document, prepared in April 2011, outlines the projected electricity demand for Nova Scotia. However, the provided text lacks detailed analysis or specific data points, suggesting further information is required for a comprehensive understanding.

Figure 1 Annual Net System Requirement In addition to annual energy requirements, NSPI also forecasts the peak hourly demand for future years. The forecast methodology uses forecast energy requirements and expected load shapes (hourly cons...

AI summary NSPI forecasts annual energy and peak hourly demand using historical load shapes adjusted for changes like new equipment. Net System Peak is projected to decline by 1.5% annually until 2021 due to DSM programs, contrasting with a 1.0% growth rate without DSM. This reflects the impact of demand-side management on reducing peak demand.

Figure 2 Annual Net System Peak (Winter-ending) The hourly peak demand in the year 2010 occurred in February and was 2,114 MW with temperatures of approximately -13°C (Winter peaks are typically set when cold temperatures drive residential...

AI summary Figure 2 details the 2010 winter peak demand of 2,114 MW at -13°C, driven by heating loads. The 2012 forecast projects 2,301 MW under typical winter temperatures, highlighting temperature-driven demand patterns.

6 7 1 3 NSPI annually develops a forecast of energy sales and peak demand requirements to assess the 4 effects of customer, demographic and economic factors on the future provincial system load. It is a fundamental input to the overall pla...

AI summary NSPI annually develops forecasts of energy sales and peak demand requirements, using information available at the time, to assess the effects of customer, demographic, and economic factors on future provincial system load. This forecast covers the period 2011-2021 and uses average annual growth rates calculated between these years.

Forecast Models 11 10 12 Nova Scotia electric energy sales are modeled and forecast as three provincial customer sectors: 13 residential, commercial and industrial. Energy forecasts for sector electricity sales are calculated using econome...

AI summary The document discusses the forecasting models used for Nova Scotia electric energy sales, divided into residential, commercial, and industrial sectors. Econometric models, particularly multiple linear regression equations, are used to predict future energy loads based on variables such as population, GDP, oil prices, and heating degree-days. Data sources include the Conference Board of Canada's Economic Outlook.

Losses 8 9 7 System losses have averaged 6.7 percent of NSR over the past five years and are expected to remain in the 6.6 to 6.7 percent range over the 10 year forecast period. 11 12 10

AI summary System losses have averaged 6.7% of NSR over the past five years and are projected to remain between 6.6% and 6.7% over the next decade.

Energy Forecast Details 13 14 15 16 For forecasting, modeling and sales reporting, Nova Scotia electric load is divided into three sector requirements: residential, commercial and industrial. The relative sizes of sector sales are shown in...

AI summary Nova Scotia electric load is segmented into residential, commercial, and industrial sectors for forecasting, modeling, and sales reporting. Figure 4 illustrates the relative sizes of sector sales, highlighting their proportional contributions to overall demand.

Residential Sector Sales 2 1 In 2010, residential customers represented approximately 37 percent of total Nova Scotia energy sales. In addition to direct domestic customers of the Company, the sector also includes residential customers ser...

AI summary In 2010, residential customers accounted for 37% of Nova Scotia energy sales, including municipal utility customers and seasonal residences. The sector allows detailed modeling via econometric methods using variables like appliance indexes, heating loads, and CHDD (Composite Heating Degree-Days) to forecast demand and penetration rates.

Commercial Sector Sales 8 9 10 11 12 13 14 7 Energy sales to the commercial sector in 2010 represented 29 percent of Nova Scotia sales. This customer group includes restaurants, hotels, offices, recreational facilities, stores warehouses h...

AI summary Energy sales to Nova Scotia's commercial sector in 2010 accounted for 29% of total sales, influenced by GDP, real personal disposable income (RPDI), and demand-side management (DSM) effects. An econometric model using real GDP, RPDI, residential sales, and prior commercial sales forecasts energy demand, with input from major customer surveys.

Figure 9 Annual Energy – Commercial Sector 16 17 18 19 20 Growth in this sector has averaged 0.5 percent over the past 5 years (also 0.6 percent when adjusted for weather). Driven by trends in wholesale trade, consumer confidence, and grow...

AI summary The commercial sector in Nova Scotia is projected to grow at 0.5% annually (0.6% weather-adjusted) through 2012, driven by retail trade activity linked to consumer confidence and disposable income. Demand Side Management (DSM) is expected to reduce annual load rates by 2.0% over 10 years, contrasting with a 1.0% increase without conservation efforts. Figure 10 illustrates these forecasts.

Industrial Sector Sales 8 9 10 - In 2010, the industrial sector represented 34 percent of Nova Scotia total electricity sales. This group is comprised of customers who process raw materials or manufacture finished goods. It includes both p...

AI summary The industrial sector in Nova Scotia accounted for 34% of total electricity sales in 2010, driven by manufacturing and resource industries. Large customers dominate energy consumption, with DSM effects significantly impacting usage. Economic factors like GDP and employment influence demand, modeled via econometric equations using historical sales data and economic indicators.

System Losses and Unbilled Sales - 28 The load forecast is developed using Nova Scotia Power "billed" sales rather than "accrued" - sales to provide a longer historical time series upon which to base the models. Billed sales refers - 30 to...

AI summary NSPI uses billed sales for load forecasting due to longer historical data, distinguishing between billed and accrued sales. System losses, including transmission/distribution losses and unbilled sales, are estimated at 6.6-6.7% of total energy requirements.

Small General 11 12 - 13 Prior to 2004, this class comprised commercial sector customers whose annual energy - 14 consumption was less than 12,000 kWh. This threshold was changed to 32,000 kWh/yr by - 15 January 2005. This moved some custo...

AI summary The Small General class's customer threshold increased from 12,000 kWh to 32,000 kWh/yr in 2005, shifting some customers from the General class, affecting load distribution. In 2010, it had 23,436 customers consuming 235 GWh, forecasted to 219 GWh in 2012.

1 Commercial Sector Econometric Model Detail 2 3 4 5 $COMENG = 0.01906 \ RQTOS + 0.01362 \ RPDI + 0.2685 \ DOMENG + 0.4245 \ COMENG_{-1}$ 6 7 Forecast Model for ComEng 8 Regression(4 regressors, 0 lagged errors) 10 Coefficient Std. Error t...

AI summary The Commercial Sector Econometric Model forecasts ComEng using variables like RQTOS, RPDI, DOMENG, and lagged COMENG. The model shows high statistical significance (t-statistics >3.0) and a near-perfect fit (Adj. R-square = 1.00). Forecast accuracy is strong (MAPE = 0.76%, MAD = 20.48), with low residual errors (Durbin-Watson = 2.03).

Commercial Sector Model Fit using historical data Industrial Econometric Model Details 1 2 3 4 Small and Medium Industrial class models are shown below. 5 6 7 8 SM_{IND} = 0.01885 GDP_{Man} + 0.01278 NonRes_{Inv} + 0.7220 SM_{IND_{-1}} 9 1...

AI summary The document presents econometric models for small and medium industrial sectors, incorporating GDP, employment, and prior consumption. Models demonstrate high adjusted R-square values (0.98 and 0.95) and low forecast errors (5.74 and 17.40), with statistical tests (Ljung-Box) indicating no significant autocorrelation.

1 2 3 4 5 6 7 8 9 .0 1 1 2 3 4 13 4 5 16 17 8 9 20 Appendix C 21 22 Forecast Sensitivity by Major Variable rorceast sensitivity by major variable 23 0/1 Forecast Sensitivity by Major Variable

AI summary The text presents a table titled 'Forecast Sensitivity by Major Variable' from a regulatory proceeding document. The table appears to analyze how different variables impact forecast outcomes, though the content is not fully visible due to the inclusion of a figure and incomplete data.

2009 Load Forecast Prepared March 2009

AI summary The 2009 Load Forecast, prepared in March 2009, outlines energy demand projections for Nova Scotia. It references various economic and weather factors influencing load, including GDP, HDD, and consumer spending metrics.

Figure 1 Annual Net System Requirement In addition to annual energy requirements, NSPI also forecasts the peak hourly demand for future years. The forecast methodology uses forecast energy requirements and expected load shapes (hourly cons...

AI summary NSPI forecasts annual net system requirements and peak hourly demand using energy requirements and load shapes adjusted for customer-specific changes. Methodology incorporates historical data and expected load profiles, with growth trends illustrated in Figure 2.

T 4 4 • Intr utin ction 11111 vuu NSPI annually develops a forecast of energy sales and peak demand requirements to assess the effects of customer, demographic and economic factors on the future provincial system load. It is a fundamental...

AI summary NSPI annually develops energy sales and peak demand forecasts to assess the impact of customer, demographic, and economic factors on provincial system load. This forecast, produced in winter 2009, covers 2009-2019 and is a key input for planning, budgeting, and operations.

Residential Sector Energy Residential Sector Residential Sector Year With DSM Growth Rate Without DSM Growth Rate GWh % GWh % 2001 3,741.2 1.9 3,741.2 1.9 2002 3,828.9 2.3 3,828.9 2.3 2003 4,010.5 4.7 4,010.5 4.7 2004 4,113.5 2.4 4,113.5 2...

AI summary The table shows residential sector energy consumption (GWh) with and without DSM programs from 2001-2019, highlighting growth rates. Forecasts indicate a 0.8% annual decline in residential load with DSM, versus a 0.5% increase without DSM effects over the 10-year period.

Large General This class comprises large commercial sector customers (malls, universities, hospitals, etc) whose regular maximum demand is 2,000 kVA or more. As of December 2008, there were 18 customers in this class representing 3.6 perce...

AI summary The Large General class includes large commercial customers (malls, universities, hospitals) with 2,000 kVA or more. As of December 2008, 18 customers (3.6% of NSPI sales) were in this class. Annual load growth is projected at 0.2% with conservation/DSM programs and 1.0% without.

Small Industrial This class comprises small industrial, farming and processing customers whose regular demand is less than 250 kVA. This class was made up of 2,260 customers as of December 2008, and had sales representing 2.2 percent of NS...

AI summary The Small Industrial class includes customers with demand under 250 kVA, comprising 2,260 customers in 2008, accounting for 2.2% of NSPI sales. Energy requirements are projected to decline 0.3% annually with conservation/DSM programs or grow 1.2% without them.

Medium Industrial - 30 This class is applicable to any industrial customer having a regular demand of at least 250 kVA, - 31 but less than 2,000 kVA. As of December 2008, there were 196 customers in this class, - 32 representing about 4.6...

AI summary The Medium Industrial class includes customers with 250–2,000 kVA demand, comprising 4.6% of NSPI sales (196 customers in 2008). Sales are projected to decline 7.3% over 10 years without conservation/DSM programs, but may grow 1.2% with them.

Peak Demand The Total System Peak is defined as the highest single hourly average demand experienced in a year. It includes both firm and interruptible loads and due to the weather-sensitive load component in Nova Scotia, the Total System...

AI summary The Total System Peak is defined as the highest hourly average demand in a year, influenced by weather-sensitive loads in Nova Scotia. The 2007/2008 peak of 2,192 MW at -14°C was 46 MW lower than the 2004 peak despite colder temperatures. Forecasting methods use historical data and DSM impacts, projecting a 1.3% annual decline in Net System Peak by 2019 due to conservation efforts.

Forecast Model for DOMENG Regression(5 regressors, 0 lagged errors) Term Coefficient Std. Error t-Statistic Significance AIDX 323.774438 53.421556 6.060745 0.999998 CUSTHDD 0.250221 0.038338 6.526670 0.99999 RRCGOODS 0.107744 0.014962 7.20...

AI summary This section presents a forecast model for DOMENG Regression with five regressors and no lagged errors. The table shows the coefficients, standard errors, t-statistics, and significance levels for variables such as AIDX, CUSTHDD, RRCGOODS, RREP, and DOMENG1, highlighting their statistical significance in predicting demand energy.

Commercial Sector Econometric Model Detail 1 23 4 5 $COMENG = 0.02531 RQTOS + 0.01268 RPDI + 0.2806 DOMENG + 0.3676 COMENG._{I}$ 6 7 Forecast Model for ComEng 8 Regression(4 regressors, 0 lagged errors) 9 10 Coefficient Std. Error t-Statis...

AI summary The document presents a regression model for forecasting commercial energy consumption (COMENG) using variables like RQTOS, RPDI, DOMENG, and lagged COMENG. The model shows high statistical significance (t-statistics >5) and a strong fit (R-square 0.9977). Forecast accuracy metrics include MAPE 0.76% and RMSE 24.28.

Table A1: Energy Requirement – 2009 NSPI Forecast

AI summary Table A1 presents the 2009 energy requirement forecast by Nova Scotia Power Inc. (NSPI), incorporating factors like demand-side management, appliance saturation, and economic indicators such as GDP and personal disposable income. The table likely includes projections for residential, commercial, and industrial energy consumption.

1 2 3 4 2 3 4 5 6 7 8 9 10 11 12 13 13 15 16 17 18 19 20 21 Appendix D 22 23 Forecast Sensitivity by Major Variable 24 25 Appendix D: Forecast Sensitivity by Major Variable

AI summary This section introduces Appendix D, which focuses on forecast sensitivity by major variable, indicating an analysis of how different factors may influence forecasting outcomes in a regulatory proceeding.

2006 IRP Load Forecast Prepared September 2006

AI summary The 2006 IRP Load Forecast, prepared in September 2006, outlines energy demand projections for Nova Scotia, incorporating factors such as weather, economic indicators, and demand-side management programs.

Residential Sector Sales Residential energy sales are forecast using the same models and assumptions as described in the 2006 NSPI Load Forecast report. For the IRP forecast period beyond 2015 shown in the Load Forecast Report, the followi...

AI summary Residential energy sales are forecast using models from the 2006 NSPI Load Forecast report, with assumptions including economic data from the Conference Board of Canada, heating oil price escalation, and rising electric space heating adoption. The forecast extends to 2029, with load growth projected at 2.2% after 2025.

Commercial energy sales are forecast using the same models and assumptions as described in the 2006 NSPI Load Forecast report. For the IRP forecast period beyond 2015 shown in the Load Forecast Report, the following assumptions have been m...

AI summary Commercial energy sales are forecast using models from the 2006 NSPI Load Forecast report. The economic forecast from the Conference Board of Canada is used up to 2025, after which a 1.6 percent growth rate is applied to extend the forecast to 2029.

Industrial energy sales are forecast using the same models and assumptions as described in the 2006 NSPI Load Forecast report. For the IRP forecast period beyond 2015 shown in the Load Forecast Report, the following assumptions have been m...

AI summary Industrial energy sales are forecast using models from the 2006 NSPI Load Forecast report. The forecast assumes an economic growth rate of 0.8% for industrial loads from 2025 to 2029, extending the original forecast period beyond 2015. Table 3 provides the annual load forecast and growth rates up to 2029.

• Industrial load: - reduced by 1,700 GWh annually from 2007 onward for the low case, the equivalent to closing a major paper mill. - increased by 500 GWh annually from 2008 onward for the high case, the estimated load of a major industria...

AI summary Industrial load projections show a potential reduction of 1,700 GWh annually from 2007 (equivalent to closing a major paper mill) and a possible increase of 500 GWh annually from 2008 (equivalent to a major industrial expansion). These scenarios reflect contrasting outcomes for industrial energy demand in Nova Scotia.

• Economic variables: - the annual growth rates of the major economic indicators used in the base forecast are reduced by 50 percent for the low case. This range was judged to be suitable and is within the variation observed in provincial...

AI summary The low-case forecast reduces annual growth rates of major economic indicators by 50%, while the high-case forecast increases them by 50%. These adjustments align with observed variations in Nova Scotia's provincial GDP over recent years.

Gross and net standard savings Since energy savings come from a product energy regulation, full compliance is assumed. - Evidence from other jurisdictions indicates that the market is close to saturation for CFLs, so the - NOMAD is assumed...

AI summary The analysis assumes full compliance with product energy regulations, leading to energy savings. Market saturation for CFLs in other jurisdictions suggests NOMAD (naturally occurring market adoption) is zero.

Gross and net standard savings Since energy savings come from a product energy regulation, full compliance is assumed. - Evidence from other jurisdictions indicates that the market is close to saturation for CFLs, so the - NOMAD is assumed...

AI summary The text assumes full compliance with product energy regulations for energy savings. It cites evidence from other jurisdictions indicating the CFL market is saturated, leading to a zero NOMAD assumption.

E A U NERGY EFFICIENCY RESOURCE SSESSMENT MODEL (RAM T PDATE OOL)

AI summary The document discusses the Energy Efficiency Resource Assessment Model (RAM Update Tool) in the context of a Nova Scotia regulatory proceeding. Efficiency Nova Scotia Corporation (ENSC) is involved in updating this model, which is critical for evaluating energy efficiency resources.

Technology cost (Cell Block: G40:AG500 in Cost-Benefits Res; G40:AG800 in Cost-Benefits C&I): The multiplication by Net to Gross that existed in the calculation was taken out. This is because the technology cost was already being applied t...

AI summary The calculation removed the Net to Gross (NTG) multiplication to avoid double application, as technology costs were already applied to net savings numbers. This adjustment ensures accurate cost-benefit analysis in the specified cell blocks.

Other Changes: There are a few other places in the Model where TRC Costs are pulled together. These also had to be changed in order to take out the costs for incentives paid to free riders. These changes occurred in the following places: D...

AI summary The document describes modifications to TRC Costs in the Model, specifically removing incentives for free riders in DSM Tech Tables (Res: Tech Tables 1-4; C&I: Tech Tables 11-13). TRC costs in Columns H, N, and T of these tables were adjusted to exclude free rider incentives.

SUMMARY OF PROFESSIONAL EXPERIENCE 1986– Present President, Resource Insight, Inc. Consults and testifies in utility and insurance economics. Reviews utility supply-planning processes and outcomes: assesses prudence of prior power planning...

AI summary The individual has extensive experience in utility and insurance economics, including reviewing utility supply planning, rate design, conservation programs, and advising regulatory commissions. They have worked as President of Resource Insight, Inc., Research Associate at Analysis and Inference, Inc., and Utility Rate Analyst for the Massachusetts Attorney General, focusing on topics like demand forecasting, cost allocation, and energy conservation.

EXPERT TESTIMONY 1. MEFSC 78-12/MDPU 19494, Phase I; Boston Edison 1978 forecast; Massachusetts Attorney General; June 12 1978. Appliance penetration projections, price elasticity, econometric commercial forecast, peak demand forecast. Joi...

AI summary Expert testimony from multiple Massachusetts regulatory proceedings (1978-1979) covering demand forecasting, appliance efficiency, rate design, and reliability. Testimonies involve Boston Edison, Massachusetts Attorney General, and joint experts like Susan C. Geller. Topics include economic models, peak demand projections, and nuclear economics.

2 Q DO THESE ISSUES JUST AFFECT DSM? A No. The impact is broader. Several proceedings affect NSPI rates and must consider the3 same information: This DSM application; the recent Load Retention rate application;4 the new PWCC load retention...

AI summary The issues extend beyond DSM, affecting multiple NSPI rate-related proceedings, including the Integrated Resource Plan, renewable energy projects, and Fuel Adjustment Mechanism. Consistent and reliable information is crucial, and forecasting future rates could help ratepayers plan for energy-saving opportunities.

2013-2015 DSM Plan (Appendix A) 1. The Parties agree with the proposed investments and Programs planned for 2013 and 2014 in the DSM Plan as filed, preserving all rights respecting future positions which may be taken respecting DSM plannin...

AI summary Parties agree with the proposed 2013-2015 DSM Plan investments and programs for 2013-2014 but retain rights to challenge future aspects of DSM planning, cost allocation, forecasting, investment levels, and program measures.

Economists forecast rosier outlook

AI summary Economists predict a more positive outlook, though specific details are not provided in the text. The context involves Nova Scotia regulatory proceedings, with references to Efficiency Nova Scotia Corporation (ENSC), Electricity Demand Side Management Plan (DSM), and Associated Press (AP).

11 Request IR-22 - Reference: Appendix C, Footnote 1 to Attachment 1-4, Table 2(b) (2013), Attachment 1-9,12 - Table 2(b) (2014) and Attachment 1-14, Table 2(b) (2015), states that the breakdown of13 - Municipal Class sales is based on a 2...

AI summary The request questions whether ENSC used the most current forecasts from MEUNSC or NSPI for 2012 and 2013–2015, and asks for updated Tables 2(b) based on current forecasts. It highlights discrepancies in data accuracy and requests revised documentation.

NON-CONFIDENTIAL _________________________________________________________________________________________________________ Request IR-28: Please provide an explanation of why Figure 4.1 - 2013-2015 DSM Plan Savings and Investment including...

AI summary The Consumer Advocate (CA) requests clarifications on discrepancies in ENSC's revised DSM Plan, including unexplained increases in investment without commensurate energy savings, higher sub-program costs, comparisons with past IRPs, and definitions of terms like 'No New DSM' and 'Without Future DSM' in load forecasts. The CA questions the rationale for cost differences and data inconsistencies.

PA3 Score: 100% -[FR4_P/(FR4_P+FR4_NP)]

AI summary The PA3 score is calculated as 100% minus the ratio of FR4_P to the sum of FR4_P and FR4_NP, indicating a performance metric or evaluation methodology used in the proceeding.