Tenerife banana residues: electricity feasibility case#

This source-backed screening evaluates 5,000 wet tonnes/year of raw banana pseudostems in an assumed digestion/CHP project in Tenerife. It combines published island context with adjacent-island laboratory measurements and explicit engineering and financial assumptions. It is not an operating plant, a contracted feed supply, a detailed design or a supplier quotation.

Open the executable results gallery for five figures, a downloadable Python script and a notebook. The reproducible analysis is in analysis/tenerife_banana/.

Evidence and assumptions#

Input

Value

Status and source

Tenerife banana-fruit production, 2024

186,771 t

Published context: CES Canarias, Table 40. Not a residue inventory.

Raw pseudostem total solids / wet mass

6.2%

Laboratory measurement: Durán Hernández et al., Table 3.

Volatile solids / total solids

67.8%

Same raw-pseudostem sample and table.

Methane potential / VS fed

0.36329 Nm³/kg VS

Blank-corrected laboratory BMP, Table 6; 0 °C and 101.325 kPa.

Feed quantity

5,000 wet t/year

Selected project scale, not a supply estimate.

Realized fraction of laboratory BMP

70%

Scale-up assumption; stress-tested at 50% and 90%.

CHP nameplate / electrical efficiency

50 kW / 30%

Engineering assumptions; no measured part-load curve.

Heat recovery / methane capture

45% / 98%

Engineering assumptions, applied separately.

Digester volume / temperature

600 m³ / 37 °C

Assumptions; HRT/OLR are diagnostics, not stability criteria.

CHP outages

30 days/year

Two assumed 15-day shutdowns; digestion and plant loads continue.

Initial investment

EUR 650,000

Itemized assumption, not a quote.

Export / import electricity value

EUR 0.07 / 0.18 per kWh

Scenario assumptions, not current tariffs or contracts.

Analysis period / real discount rate

20 years / 8%

Constant 2026 EUR, unlevered and pre-tax.

The laboratory material was collected in Gáldar, Gran Canaria, and tested in Madrid. Transferring these properties to Tenerife requires local sampling. The publication’s accepted manuscript distinguishes BMP from full-scale performance. The case uses raw pseudostem values; it applies no steam-explosion enhancement and includes neither fruit, rachis, leaves nor a co-substrate.

The production figure is a pinned 2024 context observation, not a claim to the latest harvest data. Fruit tonnes are never converted into available residues. No collection fraction, residue-to-fruit ratio or island-wide generation potential is inferred. The 5,000 t/year supply remains to be established with growers.

Physical calculation#

VS_fed = wet_mass * total_solids_fraction * volatile_solids_fraction_ts
CH4_potential = VS_fed * laboratory_BMP_per_VS_fed
CH4_produced = CH4_potential * realization_fraction
CH4_captured = CH4_produced * capture_fraction

The dedicated BMPProcess keeps the VS-fed basis distinct from the original DigestionProcess, whose yield is per VS destroyed. No VS-destruction fraction is inferred from BMP. The potential methane energy must not exceed the configured feedstock LHV. The assumed 15 MJ/kg dry LHV is an energy plausibility bound, not a calorimetry measurement of these samples.

The established digester supplies methane uniformly over a representative 365-day year. Each day uses the existing generator and auxiliary model. The daily heat demand is feed sensible heat plus vessel loss:

H_feed = wet_kg_per_day * cp_kJ_per_kg_K * max(T_digester - T_feed, 0) / 3600
H_vessel = UA_kW_per_K * max(T_digester - T_ambient, 0) * 24

Both feed and ambient temperature are assumed to be 20 °C; this is not a weather replay. Recovered heat meets digester duty first, with electric backup during outages and other deficits. Mixing, fixed plant loads and shredding persist on outage days. Additional auxiliaries are 5% of gross generation. No heat is sold.

There is no microbial startup, inhibition, nutrient-balance, storage, seasonal supply, reactor sizing or detailed gas-cleaning model. Unused captured fuel is reported; the case does not assume that it is stored, flared or released. An actual installation requires a defined gas-management system. Methane emissions, carbon savings and lifecycle environmental impacts are not calculated.

Baseline results#

These are model outputs under the stated inputs, rounded for display. The gallery recomputes the result from the pinned JSON; committed reference outputs retain full precision.

Metric

Result

Methane produced

53,449 Nm³/year

Gross electricity

143.36 MWh/year

Plant auxiliaries

52.32 MWh/year

Electric backup heating

10.15 MWh/year

Positive electricity delivered/exported

94.75 MWh/year

Electricity imported

13.86 MWh/year

Net electricity (deliveries minus imports)

80.89 MWh/year

Net electricity per wet tonne

16.18 kWh/t

Annual operating cash flow

−EUR 134,862

20-year NPV

−EUR 1,999,062

Benefit/cost ratio

0.032

Gross LCOE per delivered kWh

EUR 2.219/kWh

Simple and discounted payback

Not reached within 20 years

Total gate fee required for zero NPV

EUR 40.72/t of incoming feed

The baseline does not support a standalone electricity-sales project under these assumptions. The break-even fee is an arithmetic requirement, not evidence of a payer or a waste-management contract. A different feed mix, lower shared costs or an electricity user can change the result and must be modeled explicitly.

Cost boundary and sensitivity#

The input file itemizes receiving/shredding, digestion/civil works, gas treatment, CHP/heat recovery, electrical connection and engineering/contingency. Operating costs cover labour/monitoring, maintenance, insurance/administration, collection, transport, residue handling and imported electricity. The transport rate assumes the empty return is included; only loaded one-way tonne-km are charged. Residue handling is a cost per incoming tonne, not a prediction of digestate mass.

Year 10 includes EUR 40,000 replacement cost and year 20 includes EUR 30,000 closure cost. The model does not include land purchase, taxes, financing, grants, price escalation, heat/digestate/carbon sales or an avoided-disposal credit. Land, nutrient replacement, inoculum, permitting conditions and any omitted site costs need local verification; their omission does not strengthen feasibility.

The cost-benefit methods price exports and imports separately. The two-dimensional sensitivity varies the value of positive electricity deliveries (EUR 0.05–0.30/kWh) and a hypothetical gate fee (EUR 0–60/t), keeping import price and all other inputs fixed. These are deterministic stress tests, not probability bounds or market forecasts. A further sensitivity varies BMP realization.

Reproduce and audit#

git clone https://github.com/julianascencio/x2power-simulator.git
cd x2power-simulator
python -m pip install -e '.[plots]'
python -m analysis.tenerife_banana.run --plot

For numeric output only, install the package without extras and omit --plot. Use --inputs path/to/inputs.json and --output-dir path/to/output for variants. The analysis is repository code, not part of the installed core wheel: run it from the checkout. The downloadable gallery script/notebook needs that same checkout.

Outputs include the full inputs and their SHA-256 fingerprint, annual metrics, 365 daily energy rows, 21 annual cash-flow rows, both sensitivities and five PNG/SVG figures. No network calls, credentials or random sampling are used. Source facts are transcribed into the input file with source IDs, locators and retrieval date; complete third-party publications are not copied into the repository.

Before progressing: measure Tenerife feed properties and collectable quantities, verify nutrient requirements and stable digestion, measure logistics and existing residue use, request installed-cost quotes, and validate an offtake or coincident self-consumption profile. Model co-digestion separately instead of attributing its potential to banana pseudostems alone.

References#

  1. CES Canarias, Informe Anual 2024, Table 40, printed p. 86, PDF p. 88. Original report. The table attributes production figures to ASPROCAN.

  2. Durán Hernández et al. (2026), Methane production from the co-digestion of banana crop residues and the organic fraction of municipal solid waste: influence of steam explosion. DOI: 10.1186/s13068-026-02794-y. Accepted manuscript: §2.1 and §2.3, Table 3 (PDF p. 19) and Table 6 (PDF p. 24).

  3. NREL (2020), LCOE in SAM using FCR and Single Owner Financial Models. Financial convention, not a source of the example’s prices, costs or discount rate.