160°C Steam Heat Pumps: When Do They Beat Gas Boilers?

Industrial heat pumps can reach steam temperatures that once made electrification look impractical. The harder question is whether a factory can buy enough useful steam, at the right hours, to recover the cost of installation. A maximum outlet temperature does not answer that question.

On October 7, 2026, Fraunhofer ISE reported work on a test stand for process steam and pressurized hot water up to 160°C, alongside separate monitoring of five district-heating heat pumps. The announcement provides a timely catalyst for procurement analysis, but it does not publish a numerical performance map for the 160°C test stand. It cannot support a universal efficiency or payback claim. Source: Fraunhofer ISE, October 7, 2026.

The practical next question is: what electricity price, annual utilization, and guaranteed system efficiency make an industrial steam heat pump preferable to an existing gas boiler?

This article combines primary-source evidence with an independently calculated screening model. EveeStatistic did not run a heat-pump experiment. Supplier figures below are identified separately from assumptions and calculations. Prices are illustrative euros, not October market quotations or project bids.

First, distinguish temperature capability from useful performance

A heat pump upgrades heat from a lower-temperature source. Its economics depend on the electrical work needed to deliver useful heat across that temperature difference. The US Department of Energy's industrial heat-pump technical brief explains this relationship and the need to compare purchased energy saved against operating and capital costs. Its historical price examples should not be reused as current tariffs. Source: DOE technical brief, June 2003.

An instructive comparison appears in Svenska Rotor Maskiner's STC300 supplier sheet, dated September 2025 and hosted by the Heat Pumping Technologies programme. The document explicitly says its information lacks third-party validation. These are supplier-reported operating points, not independent measurements by this publication. Source: SRM supplier sheet.

The sheet concerns a steam-compressor technology and also describes a cascade application. Do not assign these COPs to a complete installation lifting low-temperature waste heat to steam: the upstream stage and auxiliaries require their own accounting. The sheet's installation plans are not evidence that commissioning subsequently occurred.

Our procurement inference is straightforward: specify both sides of the heat exchanger, the electrical boundary, and the complete heat-supply chain before comparing a vendor number with boiler efficiency. A warmer source and a different delivery condition describe a different engineering problem.

Define the boundary before calculating savings

For the following model, useful output is the net thermal energy delivered to a common steam-header boundary, referenced to incoming feedwater. Electrical input includes the complete heat-pump train and its incremental pumps and auxiliaries. The source heat is assumed otherwise wasted and available at no purchase cost.

That last assumption must be tested. A recoverable stream can have an opportunity cost if another process already uses it. Conversely, avoided cooling costs may add value. Neither benefit nor penalty is included here.

Define annual system performance as:

Both numerator and denominator are energy, in MWh. This ratio is not the arithmetic average of instantaneous COP readings. An excellent efficiency reading during a quiet hour cannot compensate for large electrical consumption during poor operating conditions.

The gas comparator is an existing, serviceable boiler. Its capital expenditure is treated as sunk. We assume 90% efficiency from fuel energy to useful output at the same boundary, using a consistent fuel heating-value basis for both efficiency and price. This is a scenario assumption, not a claim about every industrial boiler.

The comparison therefore asks whether to add a heat pump now. For a greenfield project or mandatory boiler replacement, include the avoided boiler investment and revise the result.

The operating threshold is necessary, but insufficient

Let electricity cost $p_e$ be euros per MWh of electricity, gas cost $p_g$ euros per MWh of fuel, and boiler efficiency $\eta_b$. Ignoring fixed costs initially:

The energy-only break-even electricity price is:

At an assumed gas price of €45/MWh fuel, boiler efficiency of 0.90, and system COP of 3.0, gas steam costs €50/MWh useful heat. The heat pump reaches energy-cost parity at €150/MWh electricity.

That threshold looks generous. It does not pay for equipment, integration, financing, or maintenance. Treating it as an investment decision would confuse cheaper operation with a financially justified retrofit.

Add capital recovery and utilization

Our base case assumes 1 MW of useful heat capacity, €1.2 million installed investment, €24,000 annual incremental fixed operating cost, a 15-year economic life, and an 8% real discount rate. All five are analyst-selected assumptions. Installed investment is intended to include integration and electrical works; actual quotations must replace it.

Assume constant real prices, no subsidy, no carbon charge, no salvage value, and no midlife replacement. Electricity demand charges are excluded and must be added where applicable. The retained boiler's unchanged standing costs cancel; we do not credit boiler maintenance savings.

Using a capital-recovery factor:

Here $K$ is installed euros, $F$ annual incremental fixed euros, and $Q_{annual}$ delivered MWh per year. For 6,000 equivalent full-load hours, the 1 MW installation supplies 6,000 MWh/year. Annualized capital plus fixed cost is approximately €164,195, adding €27.37/MWh before electricity.

The following values are calculated scenarios, not measured project outcomes. Every row uses the same gas baseline, COP 3.0, and 6,000 MWh annual output.

At €90 electricity, the heat pump saves €20 per MWh on energy but loses the investment comparison after capital recovery. At €60, its positive modeled margin is modest enough that an omitted connection cost or deterioration in performance could reverse the decision.

Utilization can matter more than a better COP

Rearranging the cost equation gives an investment-screening threshold:

The table below shows the maximum affordable electricity price, in €/MWh electricity, for different analyst-selected operating scenarios. All other base assumptions remain unchanged.

In this model, moving from 4,000 to 6,000 hours at COP 3 increases the affordable electricity price by about €41/MWh. Improving COP from 3 to 3.5 at 6,000 hours adds only about €11/MWh. The inference is specific to these assumptions: dependable utilization deserves at least as much attention as an attractive design-point COP.

Hours cannot be assumed independently of price. Running only during cheap electricity periods reduces output and spreads fixed costs over fewer MWh. Use an hourly simulation with matched heat demand, source availability, electricity price, and system COP before making the investment decision.

For example, a plant reporting 8,000 production hours does not necessarily offer 8,000 heat-pump hours. Its recoverable stream might disappear during cleaning, or its steam demand might peak when that stream is cold. Equivalent full-load hours must come from delivered energy divided by rated capacity, not merely the calendar time equipment is switched on.

A practical specification and acceptance method

The HPT Annex 58 Task 4 work separates specification, laboratory testing, and site testing. It emphasizes that guaranteed COP applies to defined operating conditions. That is the appropriate procurement discipline, rather than selecting equipment from an isolated efficiency figure. Source: HPT Annex 58 Task 4, accessed October 9, 2026.

The following is our proposed implementation sequence:

1. Collect synchronized operating data. Record source inlet and outlet temperatures, flow, steam demand, feedwater conditions, and electricity consumption. Include production changes and planned shutdowns. Preserve timestamps so that favorable source conditions are not incorrectly paired with unrelated demand.
2. Measure useful output consistently. Calculate steam duty from mass flow and the enthalpy difference between delivered steam and incoming feedwater. Specify pressure, temperature, steam quality where relevant, and treatment of blowdown, condensate, and storage changes. A temperature reading alone is not a heat meter.
3. Request a complete performance map. Cover the site's observed source conditions and load range. Separate upstream stages, compressor power, auxiliaries, startup energy, and standby consumption. Require vendors to identify excluded equipment and the measurement uncertainty attached to guarantees.
4. Replay the operating year. Calculate electricity consumption and cost interval by interval. Enforce available source duty, minimum load, startup constraints, and the required steam service. Compare the retrofit against the same baseline production year; do not assume an electrically cheap hour is automatically operationally usable.
5. Contract for site acceptance. Agree test conditions, calibrated instrumentation, tolerance bands, and remedies before purchase. Include representative part-load and source-temperature cases, not only the most favorable operating point. Keep process reliability and safety acceptance separate from the financial spreadsheet.

A reproducible spreadsheet needs columns for timestamp, useful heat, system electricity, electricity price, available source duty, and baseline gas use. Annual electricity cost is the sum of interval consumption multiplied by interval price. This avoids silently substituting a flat tariff for a volatile procurement contract.

What could overturn the conclusion?

The model is deliberately narrower than a bankable feasibility study. It excludes taxes, carbon prices, demand charges, grants, cooling credits, and major repairs. Different heat-source economics or avoided boiler replacement could favor the heat pump; lower utilization, expensive integration, or source interruptions could favor continued boiler operation.

Neither the supplier table nor the Fraunhofer announcement validates our assumed annual COP. A real project needs measured or contractually supported performance for its own boundary. The analysis also makes no quantified carbon claim: that requires a stated electricity emissions factor, fuel basis, and treatment of upstream emissions.

Most importantly, the base result is a threshold, not a forecast. Its value is exposing which evidence to buy next: a load-duration profile, a complete-system performance guarantee, an installed quotation, and an electricity contract that can coexist with the required utilization.

Architectural / Strategic Summary

• Buy guaranteed useful steam, not a maximum temperature: define source conditions, delivery conditions, and the complete electrical boundary.
• Separate evidence classes: supplier operating points, laboratory announcements, and modeled project economics answer different questions.
• Include capital recovery: our illustrative threshold falls from €150/MWh electricity on energy cost alone to €67.90/MWh at 6,000 hours after fixed costs.
• Validate utilization and price together: selective operation can lower energy prices while undermining capital recovery.
• Advance only with site-specific evidence: replace every scenario assumption with measured demand, credible performance guarantees, and priced integration work.