Heat recovery on industrial boilers

Industrial plants reject vast amounts of heat to atmosphere — boiler stacks at 200°C, dryer exhausts at 120°C, kiln exhausts at 150°C, compressor heat at 80°C. Capturing it usually means installing a heat exchanger between the rejected stream and a co-located thermal demand: combustion air pre-heat, feedwater pre-heat, building heating, or process water. The math is straightforward; the engineering is in finding co-located demand.

Where waste heat lives

Every industrial plant has a heat-rejection inventory. The biggest streams:

The economics scale with two factors: temperature differential and flow rate. High ΔT × high flow = lots of available kWh-equivalent. Low ΔT or intermittent flow = limited recovery potential even when the absolute temperature looks attractive.

The recovery calculation

Recoverable heat from a stream is the standard heat-transfer equation:

Q_recoverable = m × Cp × ΔT × hours × ε × realization

Worked example. A natural-gas boiler stack with 0.8 kg/s flue-gas flow, exhaust at 220°C, target post-recovery 130°C (ΔT = 90°C), boiler operating 7,500 hours/year with a feedwater economizer (ε = 0.65):

Q = 0.8 × 1.05 × 90 × 7,500 × 0.65 × 0.75 = 276,413 MJ/yr ≈ 76,800 kWh-equivalent/yr   (or 7,000 m³ of natural gas saved)

At a delivered NG price of $0.35/m³, that's $2,450/year in fuel savings. The economizer itself costs $30–60K installed, putting payback at 3–5 years on a small boiler. On a 50 t/h industrial steam boiler, the same calculation produces $30–60K/year in savings against a similar capex — payback < 18 months.

Which streams qualify

The library identifies heat-recovery candidates when:

• Source stream temperature ≥ 100°C — below this, ΔT is small enough that capex doesn't pay back unless the stream is enormous.
• Source stream is continuous or near-continuous (≥ 4,000 hours/year). Intermittent streams (batch ovens, occasional bake-out cycles) rarely justify the heat-exchanger capex.
• A co-located thermal sink exists: boiler feedwater, combustion air, process hot water, or building-heating loop.
• The sink demand is roughly proportional to the source supply — i.e. the hot water demand is during the same hours the dryer exhaust is hot.
• Physical routing is feasible — the source and sink are within ~50m and there's no insurmountable constraint (food-grade segregation, ATEX zones, pressure mismatches).

Common configurations

When it doesn't pay back

• Source temperature near or below sink temperature — the available ΔT collapses and you're paying capex for nothing.
• Intermittent source operation — boiler that fires twice a day for 30 minutes generates <500 hours of recovery time per year.
• Fouling-prone streams — wood-dryer exhausts laden with resins, foundry stacks with particulates. Cleaning costs eat the savings unless self-cleaning designs are specced.
• Corrosive flue gas — high-sulfur fuels condense to sulfuric acid below ~150°C exhaust, requiring corrosion-resistant economizer materials (stainless or higher) that double capex.
• No co-located sink — the cleanest waste-heat opportunity is useless without somewhere for the heat to go that wants it at the right time.
• Steam pressure mismatch — recovering 200°C kiln exhaust into a 350°C steam loop requires raising the recovered heat by another energy input, often making the math worse.