PaperlayerView exact revision on arXiv

Source-linked AI summary

Techno-economic analysis of PV-battery systems in Switzerland

Xuejiao Han, Jared Garrison, Gabriela Hug

arXiv:2103.16298v1eess.SYcs.SY

TL;DR

The study asks how PV-battery economics vary across Swiss customer groups and uncertainties in costs, prices, tariffs, and load profiles. It optimizes static investments over 2020–2050 and finds that viability and payback vary by group, while residual-load dynamics point to a need for fast-ramping flexibility.

  • Problem

    Prior work leaves limited analysis of how PVB economics vary across customer groups and how costs, load profiles, prices, and tariffs affect viability.

    Source: C. Status of PV and BESS in Switzerland · p. 3

  • Method

    A static investment optimization model evaluates Swiss PVB systems across customer groups and investment years from 2020 to 2050.

    Source: B. Literature review · p. 3, B. Rooftop potential and data clustering · p. 4, §VII · p. 19

  • Results

    PVB investments already yield better NPV than PV alone for some groups; payback fluctuates through 2035 and then decreases, while optimal sizes increase over time.

    Source: §VII · p. 19

  • Takeaways & Limitations

    Residual load becomes more dynamic at hourly and daily scales as PV penetration increases, emphasizing the need for flexible resources with fast ramping capabilities.

    Source: 2) Baseline results - load · p. 16

    Synthetic load profiles do not capture real-world consumption diversity, and annual demand is assumed constant, omitting possible increases from EVs and other electrification.

    Source: §VI · p. 18

Abstract

from arXiv · show

This paper presents a techno-economic optimization model to analyze the economic viability of a PV-battery system for different customer groups in Switzerland clustered based on their annual electricity consumption, rooftop size, annual irradiation and location. The simulations for a static investment model are carried out for years 2020-2050 and a comprehensive sensitivity analysis is conducted to investigate the impacts of individual parameters such as costs, load profiles, electricity prices and tariffs, etc. Results show that while combining PV with batteries already results in better net present values than PV alone for some customer groups today, the payback periods fluctuate between 2020 and 2035 due to the mixed effects of policy changes, costs and electricity price developments. The optimal PV and battery sizes increase over time and in 2050 the PV investment is mostly limited by the rooftop size. The economic viability of PV-battery system investments varies between different customer groups and the most attractive investment (i.e., that has the shortest payback period) is mostly accessible to customer groups with higher annual irradiation and electricity demand. In addition, investment decisions are highly sensitive to payback periods, future costs, electricity prices and tariff developments. Lastly, the grid impact of the PV-battery system deployments are investigated by analyzing the residual Swiss system load profiles. The dynamics of residual load profiles caused by the seasonal, daily and hourly patterns of the solar generation emphasizes the need for flexible resources with fast ramping capabilities.

I. INTRODUCTION

Changing incentives and falling battery costs make self-consumption more relevant, while prior PVB studies leave gaps in operating strategies, customer heterogeneity, sensitivity analysis, and grid impacts. The paper addresses these questions for Swiss customer groups and examines their economics and potential deployment.

A. Motivation

  • Falling battery costs and injection tariffs below retail tariffs strengthen the case for self-consumption, while batteries can increase PV self-consumption and support other applications.The paper notes that combining battery applications could make storage investments attractive today.Source: A. Motivation · p. 1

B. Literature review

  • Prior studies often use rule-based battery operation, with limited treatment of battery C-rates, load-profile effects, and grid impacts.The review also notes limited analysis of regional and customer-group differences in prices, tariffs, and trade-offs.Source: B. Literature review · p. 2
  • The proposed static investment optimization model assesses PVB economics across Swiss customer groups from 2020 to 2050, representing differences in rooftops, irradiation, electricity use, load profiles, and region.It minimizes total investment and operating costs over a 30-year horizon.Source: B. Literature review · p. 3

C. Status of PV and BESS in Switzerland

  • The study asks how PVB economics and optimal system sizes vary across customer groups, how cumulative investments may develop, and which uncertainties drive viability.It also considers potential challenges and opportunities for investors, retailers, system operators, and policy-makers.Source: C. Status of PV and BESS in Switzerland · p. 3

II. DATA

The data and assumptions define 30-year investment evaluations, rooftop-based customer groups, and the modeled PVB system and its costs.

A. General assumptions

  • Each investment-year simulation assumes a 30-year PVB lifetime, a 4% WACC, and battery replacement when battery life is shorter than the system lifetime.The model also calculates the remaining value of the last reinvested battery at the end of the system lifetime.Source: A. General assumptions · p. 3

B. Rooftop potential and data clustering

  • Nearly 4 million rooftop entries are clustered by irradiation, roof size, estimated electricity consumption, and canton, with each customer group represented by its median values.The modeled rooftop potential is 224 km2, or 37 GW, across 57,200 groups.Source: B. Rooftop potential and data clustering · pp. 3–4

C. Parameters of the PVB system

  • Each customer group is modeled with PV, a battery, a hybrid inverter, load, and grid; the battery is AC-coupled for greater operational flexibility despite requiring an additional inverter.Only lithium-ion batteries are considered, and future PV and battery costs use published projections.Source: C. Parameters of the PVB system · p. 4

D. Load and generation profiles

The study constructs customer load and PV generation profiles and models Swiss subsidies, tariffs, and tax rebates to represent investment incentives and grid remuneration.

D. Load and generation profiles

  • Synthetic household load profiles are scaled to match estimated annual demand, using predefined household settings for different consumption categories.The demand estimate is based on warm-water consumption.Source: D. Load and generation profiles · pp. 4–5
  • PV generation profiles use historical hourly irradiation data from canton-representative stations, scaled by annual irradiation category, with perfect generation forecasts assumed.Generation is calculated from module area, efficiencies, performance ratio, and irradiation.Source: D. Load and generation profiles · p. 5

E. Policies and regulations

  • The policy model includes capacity-based investment subsidies and tax rebates, while the feed-in-tariff scheme is expected to expire in 2022.The investment subsidy can cover up to 30% of investment costs and is valid until 2030, with a possible extension to 2035 noted.Source: E. Policies and regulations · pp. 5–6
  • Baseline grid exports are remunerated at regional DSO injection tariffs, assumed to decline 10% annually, but never below the annual Swiss wholesale-price floor.The model includes these tariffs because they value exported PV and affect the benefits of storing generation for later self-consumption.Source: E. Policies and regulations · p. 6

F. Scenarios

The scenarios test how uncertain costs, prices, tariffs, load profiles, and financing assumptions affect PVB investment decisions. Most sensitivities are one-at-a-time simulations for Zurich in 2050, while the Baseline covers all cantons from 2020 to 2050.

F. Scenarios

  • Future costs, electricity prices, tariffs, and policies are uncertain, so the study varies key assumptions individually in Zurich in 2050 against a 2020–2050 all-canton Baseline.Financial assumptions are also simplified as constant across PV categories, although investors may differ in financing, payback preferences, and non-economic motivations.Source: F. Scenarios · p. 6, 2) Load profile scenarios · p. 6

1) PV and battery cost scenarios:

  • High- and low-cost scenarios average 15% higher or lower PV costs and 54% higher or lower battery costs than the Baseline.The differences vary by year and size category.Source: 1) PV and battery cost scenarios · p. 6

2) Load profile scenarios:

  • Synthetic individual profiles peak in the morning and evening and show higher weekend use than the aggregated profile, motivating a sensitivity scenario that substitutes aggregated cantonal profiles.The study compares resulting PV self-consumption and battery evaluations under the aggregated-profile case.Source: 2) Load profile scenarios · pp. 6–7

3) Electricity price scenarios:

  • The Baseline raises retail prices 1% annually, excludes hourly wholesale access, and lowers injection tariffs 10% annually until they match the wholesale price, which rises 1.5% annually.Nine scenarios test wholesale-market price access, while another assumes zero injection tariffs and no hourly wholesale access.Source: 3) Electricity price scenarios · p. 7

4) Battery scenario:

  • Battery sensitivities use a 2020 installation cost of 14,700 CHF for a 13.5 kWh Powerwall 2 with Baseline cost reductions, and also test investment without batteries.The Swiss price adjustment reflects battery costs, especially labor, being higher than the global average.Source: 4) Battery scenario · p. 7

5) WACC scenarios:

  • The Baseline assumes a 4% WACC for all PVB investments, with sensitivities at 2% and 8%.The scenarios compare alternative cost-of-capital assumptions against the Baseline.Source: 5) WACC scenarios · p. 7

III. METHOD

The static optimization evaluates each region and year over a 30-year PVB-system lifetime, optimizing investment and hourly operation. It constrains PV deployment by rooftop potential and models PV generation and battery operation over time.

III. METHOD

  • For each region and examined year, the model optimizes investment and operation across a 30-year lifetime using all 8,760 hours of the examined year and projected parameters thereafter.It minimizes total investment and operating costs subject to investment, operational, and system power-balance constraints.Source: §III · pp. 7–8

A. Investment constraints

  • Each customer group combines irradiation, electricity-consumption, and roof-size categories, and PV capacity is bounded by rooftop potential across five candidate size categories.The model considers 2,200 possible customer-group combinations; available rooftop area is converted using 6 m2/kWp.Source: A. Investment constraints · p. 8

B. Operational constraints

  • PV generation is limited by module area, module and inverter efficiencies, performance ratio, and solar irradiation; the constraint permits curtailment.The rooftop area required per installed kWp is included in the generation limit.Source: B. Operational constraints · p. 8
  • Battery stored energy is bounded by depth of discharge and installed energy capacity, while charging and discharging are limited by installed power capacity.The storage level is linked across consecutive time steps using charging and discharging efficiencies.Source: B. Operational constraints · p. 8

C. Power balance constraints

The model balances PV generation across charging, serving demand, and grid injection, and supplies demand from PV, batteries, or the grid. It minimizes discounted lifetime costs while accounting for investment, operation, revenues, degradation, and battery replacement.

C. Power balance constraints

  • At each time step, PV output can charge the battery, serve load, or be injected into the grid, while demand can be met by PV, battery, or grid power.Self-consumption includes PV electricity used directly by the customer or indirectly through battery-to-load power.Source: C. Power balance constraints · p. 8

D. Formulation of optimization problem

  • The objective minimizes PVB-system investment and operating costs over its lifetime using net present value, with future revenues and costs discounted by WACC.The formulation applies the stated constraints across the entire system lifetime.Source: D. Formulation of optimization problem · p. 9
  • Investment costs include net PV costs after subsidies and tax rebates, plus battery costs tied to both energy and power capacity.Operating costs include variable and fixed components for the system.Source: D. Formulation of optimization problem · p. 9
  • Annual revenues combine grid-injection reimbursement and retail-tariff savings from self-consumed electricity, with degradation applied over time.Retail tariffs use high and low rates that vary by annual consumption category.Source: D. Formulation of optimization problem · p. 9
  • Battery replacement is included because battery lifetime is 13 years versus 30 years for the PVB system, and the final replacement can retain residual value.Replacement capacity is assumed equal to the initial battery capacity; a 13-year battery in a 30-year system requires two replacements, the last in year 27.Source: D. Formulation of optimization problem · p. 9

E. Technical and economic indicators

The section introduces three indicators used to evaluate investment decisions: self-consumption rate, self-sufficiency rate, and payback period.

E. Technical and economic indicators

1) Self-consumption rate:

The section defines self-consumption, self-sufficiency, and payback metrics, then describes how the study evaluates investment economics across years and customer groups.

1) Self-consumption rate:

  • The self-consumption rate (SCR) is PV electricity directly or indirectly consumed by the system owner divided by total PV generation.Source: 1) Self-consumption rate · p. 9

2) Self-sufficiency rate:

3) Payback period:

  • Payback period (PBP) equals investment cost divided by yearly cash flow; shorter PBPs indicate more attractive investments.Source: 3) Payback period · p. 10

IV. RESULTS

  • The baseline model evaluates separate greenfield investments every five years from 2020–2050, minimizing costs over a 30-year lifetime with hourly operation simulated for each investment year.Unlike a dynamic multi-period model, it does not optimize investment timing or produce investment pathways; sensitivities are run only for Zurich in 2050.Source: §IV · p. 10
  • Results proceed from one representative customer group to Zurich-wide groups and then Switzerland, with baseline results followed by sensitivities in the first two analyses.Source: §IV · p. 10

A. Results for one representative customer group

For the representative Zurich customer group, optimal PV and battery sizes grow over time as costs fall, while dispatch responds to PV output and tariff periods. Sensitivity results show that costs, load profiles, and electricity-price assumptions affect investment outcomes.

1) Baseline results - investment:

  • PBP generally falls from above 13 years in 2025 to below 10 years in 2050, while optimal PV and battery sizes and NPV generally rise; subsidy expiration produces a 2030–2035 setback.Early profitability is driven by high SCR, whereas later investments benefit more from falling costs; larger systems can be profitable even beyond household demand.Source: Table VII · p. 10, 1) Baseline results - investment · p. 10

2) Baseline results - dispatch:

  • The battery generally charges when PV exceeds load and discharges when load exceeds PV, but on Sundays it can charge from cheap grid electricity for later peak-tariff use.The modeled Sunday off-peak tariff and Monday–Saturday 6:00–22:00 peak period explain this exception; the baseline uses dual retail tariffs and a constant injection tariff.Source: 2) Baseline results - dispatch · pp. 10–11

3) Sensitivity scenario results:

2) Baseline results - dispatch:

  • Battery capacity rises from 5.7 kWh/1.0 kW in 2030 to 10.3 kWh/2.3 kW in 2050, while PV rises from 2.3 kW to 6.0 kW.The 2050 dispatch has lower grid purchases and higher PV injections than 2030, with more pronounced grid power dynamics as battery capacity per PV kW falls.Source: 2) Baseline results - dispatch · p. 11, Figure 3 · p. 11

3) Sensitivity scenario results:

  • Higher costs reduce optimal battery size and NPV and lengthen PBP, while lower costs have the opposite effects; PV size remains unchanged, and Tesla-based battery pricing yields even less battery investment.Removing batteries also produces lower NPV and longer PBP than the baseline; outcomes are particularly sensitive to future PV and battery costs and electricity prices.Source: 3) Sensitivity scenario results · pp. 11–12
  • An aggregate load profile keeps PV size unchanged but reduces battery size and C-rate, because its flatter demand better matches PV and achieves similar SCR and SSR with a smaller battery.The aggregate-profile scenario consequently has higher NPV and shorter PBP than the baseline.Source: 3) Sensitivity scenario results · p. 12
  • Higher wholesale prices increase PV investment and NPV but reduce SCR and battery capacity per PV unit, whereas higher retail prices increase PV and battery sizes and NPV while shortening PBP.Wholesale-price effects are generally smaller than retail-tariff effects; wholesale price also raises battery C-rate as PV investment grows.Source: 3) Sensitivity scenario results · p. 12
  • With zero injection tariff and no wholesale-market access, PV size is unchanged, battery size rises slightly, and NPV falls slightly while PBP lengthens slightly.Source: 3) Sensitivity scenario results · p. 12

B. Results for all customer groups within the canton of Zurich

Across Zurich customer groups, investment sizes and cumulative capacity rise over time, but payback periods fluctuate through 2035 as policy and market factors interact. Sensitivity results show differing responses across PV and battery investments, and the Switzerland analysis classifies investments by recovery period.

1) Baseline results:

  • Cumulative PV and battery capacity rises from 1.4 GW and 0.4 GWh/0.1 GW in 2020 to 3.6 GW and 7.7 GWh/1.6 GW in 2050, against 4.7 GW modeled PV potential.Weighted-average investment sizes grow, while NPV rises and weighted-average PBP fluctuates through 2035 before declining; totals include every investment with positive NPV, however small.Source: Table IX · p. 12, 1) Baseline results · pp. 12–13
  • Zurich’s cumulative investments in 6–10 kW and 10–30 kW PV units grow significantly, while smaller-unit capacity shifts toward larger categories as investment costs fall.PV investments below 6 kW first rise and then fall; investments above 30 kW generally rise, with a small decline from 2030 to 2035.Source: Figure 4 · p. 13, 1) Baseline results · p. 13
  • Across 2,200 Zurich customer groups, PV and battery capacities generally increase alongside capacity with shorter PBPs, but 2030 and 2035 curves intersect or overlap, especially for PV.The reported explanation is the mixed effect of cost reductions and subsidy expiration at the end of 2030.Source: Figure 5 · p. 13, 1) Baseline results · p. 13

2) Sensitivity scenario results:

  • In 2050, PV capacity is highly sensitive to electricity-price scenarios, while battery capacities vary notably, reaching their minimum under aggregate loads and maximum under low costs.Among price scenarios, PV integration is lowest at SP1 and highest at SP9; battery capacity is highest at SP7 and lowest at SP3.Source: Figure 6 · p. 14, 2) Sensitivity scenario results · p. 14

C. Results for Switzerland

  • The Switzerland analysis covers 26 cantons with 2,200 customer groups in each region and considers baseline results for 2020, 2030, 2040, and 2050.Source: C. Results for Switzerland · p. 14
  • The national analysis defines fast recoverable investments as PBP below 10 years and moderately fast recoverable investments as PBP below 15 years.Investments are optimized for NPV and may have PBP up to the system’s 30-year lifetime, though the text notes most investors would expect a shorter recovery period.Source: C. Results for Switzerland · p. 14

1) Baseline results - investment:

Optimal PV-battery investments grow over time, but their payback and viability differ across customer groups. The resulting load profiles become more seasonal and dynamic, while investment economics remain sensitive to assumptions and market developments.

1) Baseline results - investment:

  • The investment decision is especially sensitive to the acceptable payback period during 2030–2040; by 2050, almost all investments have payback periods below 10 years.The 15-year-payback investment results are almost always at the top edge of the box because most investments have payback periods under 15 years.Source: 1) Baseline results - investment · p. 14
  • In 2050, fast-recoverable investments are more evenly distributed across regions and PV categories, with the most attractive investments concentrated among groups with higher irradiation and electricity demand.Optimal PV size is generally limited by rooftop size.Source: 1) Baseline results - investment · p. 15, Figure 8 · p. 15, Figure 9 · p. 16

2) Baseline results - load:

  • Residual load remains high in winter but falls increasingly in summer as viable PV installations accumulate, with a stronger seasonal effect under the less stringent 15-year payback threshold.The assumed original Swiss load is constant over time.Source: 2) Baseline results - load · p. 15, Figure 10 · p. 16
  • By 2050, residual load plunges and recovers sharply on sunny days, with hourly and daily variation intensifying as PV penetration and the range of included investments increase.On winter Sundays, 2020 residual load can be lower than or similar to later years because batteries absorb PV generation while grid electricity supplies load at the low Sunday tariff.Source: 2) Baseline results - load · pp. 15–16

3) Baseline results - self-consumption:

  • Self-consumption rate peaks in 2040 even as PV generation increases, because later investments are more driven by lower system costs than by increasing that rate.The table covers fast- and moderately fast-recoverable investments from 2020 to 2050.Source: 3) Baseline results - self-consumption · p. 16, Table X · p. 16
  • In 2050, required retail tariff increases are 12.6 cent/kWh for fast- and 15.4 cent/kWh for moderately fast-recoverable investments, a simplified estimate that excludes investment rebound.These increases correspond to 67% and 82% above the 2020 average household tariff, respectively.Source: 3) Baseline results - self-consumption · p. 17

A. From investors’ perspective

  • Some customer groups already have better net present value from PV-battery systems than PV alone, while payback periods fluctuate through 2035 and decrease significantly afterward.Profitability and optimal system size vary with consumption, location, irradiation, and rooftop size; results are especially sensitive to costs, tariffs, and electricity prices.Source: A. From investors’ perspective · p. 17
  • The optimized investment decisions depend heavily on scenario assumptions and exclude non-economic motivations, so the authors do not claim they will be realized.Sensitivity analyses examine how assumptions such as payback period and unit costs affect potential investments.Source: A. From investors’ perspective · p. 17

B. From retailers’ perspective

The analysis links PV-battery deployment to changing prosumer purchases and increasingly dynamic residual loads, with implications for tariffs, flexibility, and policy. Its conclusions are bounded by assumptions about demand, investment behavior, clustering, and future electricity prices.

VII. CONCLUSIONS

  • By 2050, PV generation self-consumed by end-consumers could reach 24.0 TWh in the fast- and 27.3 TWh in the moderately fast-recoverable case, while changing purchases may require rethinking tariff and subsidy design.The figures equal 37.9% and 43.1% of 2019 Swiss demand, respectively.Source: B. From retailers’ perspective · p. 18, §VII · p. 19

B. From retailers’ perspective

  • Lower annual net consumption alongside more dynamic residual loads and higher peak net-load could increase grid burden, while electrification of heating and transport makes future load effects uncertain.A capacity-based grid tariff could support pricing and dispatch that better account for grid impacts.Source: B. From retailers’ perspective · p. 18

C. From system operators’ perspective

  • Greater PV penetration increases residual-load seasonality and hourly and daily variation, indicating needs for seasonal flexibility and fast-ramping resources.The paper identifies hydro storage, pumped hydro, batteries, and demand-side management as examples of flexible resources.Source: C. From system operators’ perspective · p. 18

D. From policy-makers’ perspective

  • The policy discussion links fluctuating payback periods and residual-load dynamics to flexibility needs and lower DSO revenues, and calls for considering market rules that support both renewable investment and flexibility.Sensitivity analyses are presented as a way for policy-makers to identify investment drivers.Source: D. From policy-makers’ perspective · p. 18

VI. LIMITATIONS AND FUTURE WORK

  • Synthetic load profiles may not capture regional and sectoral consumption diversity, while approximated and constant annual demand omits possible electrification growth.The authors suggest future work use realistic, evolving, region-specific building demand patterns.Source: §VI · p. 18
  • Limited rooftop clusters represented by median values may poorly represent highly varied or sparse groups, particularly those with high consumption or large roofs.More clusters would increase computational complexity and simulation time.Source: §VI · p. 18
  • The investment model maximizes NPV without representing non-economic factors or heterogeneous investor traits such as risk profiles and peer effects.The authors note that adding these features would require more input data and implementation overhead.Source: §VI · p. 18
  • Wholesale-price scenarios use historical prices scaled by multiplication factors, so they miss price suppression from PV injections and reflect generation changes in surrounding countries.The authors also note that finer time resolution could enable more granular operational and investment assessments.Source: §VI · pp. 18–19

APPENDIX A INPUT DATA OF A LOW, BASELINE AND HIGH COST

The appendix lists low, baseline, and high cost scenarios for PV and battery units, followed by canton-level injection and retail tariff data.

SCENARIO FOR PV UNITS

SCENARIO FOR BATTERY UNITS

APPENDIX C ASSUMED DSO INJECTION TARIFF BY CANTON

  • A table provides estimated PV injection tariffs in cent/kWh for each Swiss canton.Source: Table XVII · p. 24

APPENDIX D ASSUMED RETAIL ELECTRICITY TARIFF BY CANTON

  • A table provides estimated 2020 base retail tariffs in cent/kWh by Swiss canton and consumption group, with peak and off-peak tariffs assumed at 107% and 71% of the base tariff.The appendix also includes a table of electricity consumption categories.Source: Table XVIII · p. 25, Appendix D · p. 25, Table XIX · p. 25
Loading 2103.16298v1…