Most industrial heat recovery projects in Mexico do not reduce the CFE bill. They reduce the fuel bill. The energy that comes back from a heat exchanger is heat, and what it saves is the natural gas or LP that no longer has to be burned. A plant director who approves the project expecting a percentage off the electricity account will not find it there. The decision here is narrow. It is whether to measure your two hottest exhaust streams before the next capital cycle, and what that measurement has to record.
Two variables decide whether a stream is worth measuring at all. The first is temperature. The U.S. Department of Energy, in the Waste Heat Recovery technology assessment of its 2015 Quadrennial Technology Review, sorts industrial waste heat into bands, with 450 to 1,200 degrees Fahrenheit treated as medium grade and anything above 1,200 degrees treated as high grade. The second variable is annual operating hours, because recovered energy is a rate multiplied by time. A furnace exhaust at 900 degrees running 8,000 hours a year and the same exhaust running one shift are not the same asset.
That same DOE assessment puts the heat wasted in U.S. industry above a 300 degree Fahrenheit reference at 276 TBtu a year, against 1,478 TBtu a year measured against a 77 degree reference. On our read that leaves more than 80% of the counted waste heat below 300 degrees. This is the trap in the subject. The heat is everywhere, the steel to move it is cheap enough to quote, and the value of what it displaces is often too small to carry the capital. Heat recovery is a fuel project. Only cogeneration and pressure letdown recovery produce electrical output, and only those two reach the CFE account at all.
What clears the bar for recovery
The clearest published screening case is a boiler economizer. The DOE Advanced Manufacturing Office, in Steam Tip Sheet 3 dated January 2012, states that an economizer can often cut a boiler's fuel requirement by 5 to 10%. Its worked example is a 45,000 pound per hour boiler at 150 psig with a 500 degree Fahrenheit stack, running 8,400 hours a year on fuel priced at 8.00 dollars per MMBtu. On those assumptions DOE reports an annual saving of 386,400 dollars and a payback of under two years.
Two of those inputs are the ones a Mexican plant has to replace before the example means anything. Your operating hours are yours. Your delivered gas price is not DOE's 2012 assumption, and the gap between the two is where most optimistic paybacks are built.
The screen itself takes an afternoon. List every exhaust stream above 250 degrees Fahrenheit. Record its temperature, its flow and its hours per year. Then name the fuel or utility the recovered heat would displace and the point in the process where it would land. A stream with no sink inside the plant has no value however hot it is, and that is the failure mode a temperature reading on its own will not catch. This is also the difference between a walk through and what an energy audit actually looks for. Reheat furnaces and continuous annealing lines put the largest streams on one site, which is why heat recovery in Mexican steel operations reaches sizes that food and automotive plants rarely see.
The saving lands on the gas account
Convert before you compare. One gigajoule is 277.8 kilowatt hours. Divide a gas price in pesos per gigajoule by 277.8 and you have the pesos per kilowatt hour that recovered heat is worth on a thermal basis, before boiler or burner efficiency is applied. Set that against what a kilowatt hour costs the plant on CFE's Gran Demanda en Media Tensión Horaria schedule, which prices energy separately in punta, intermedia and base hours and bills demand on top.
For the gas benchmark, CRE has published the Índice de Referencia Nacional de Precios de Gas Natural al Mayoreo every month since 18 August 2017, in pesos per gigajoule and in dollars per million Btu. The U.S. Energy Information Administration described the index and its first value in a note dated 30 August 2017. The index is a wholesale reference. A delivered price adds transport and marketing, so build the case on the invoice and use the index as the sanity check.
On the electricity side, CENACE reports an average local marginal price of 613.63 pesos per MWh in the day ahead market across the Sistema Interconectado Nacional for the week of 18 to 24 August 2024. The real time average for the same week was 679.75 pesos per MWh. Those are energy prices only. A retail industrial account carries capacity, transmission, distribution and operation charges above them. The exact spread in any one month is not the point. The point is that a heat exchanger is priced off the first number and never off the second.
One consequence belongs in the capital request in writing. A heat exchanger on a paint shop oven does not reduce the plant's electrical demand by one kilowatt. The demand component of the bill does not move. A finance reviewer who has seen solar and efficiency cases will assume otherwise unless the paper says so.
The cogeneration route and what it takes to enter
Cogeneration is the case where recovered energy does become electricity, and in Mexico it carries its own permitting. Under the Ley de la Industria Eléctrica, published in the DOF on 11 August 2014, a power plant with capacity of 0.5 MW or more requires a generation permit from CRE. Article 3 of the same law treats efficient cogeneration as clean energy only where the system meets efficiency criteria issued by CRE and emissions criteria set by SEMARNAT.
CRE updated the values behind that test in Acuerdo A/018/2023, published in the DOF on 26 May 2023 and in force from 29 May 2023. The acuerdo revised the reference efficiency values applied to each capacity range and the loss factors applied by voltage level. What that means on the ground is that accreditation is a calculation on measured heat and power output rather than a label that comes with the equipment. It has to be maintained, and it decides how the electrical output is treated.
The cogeneration decision therefore has three gates rather than one. There is the thermal case, which is the same case an economizer has to pass. There is the generation permit, which starts at 0.5 MW. And there is the efficiency accreditation, which sets the value of the output. Sizing all three at once is what separates a cogeneration project and its economics from a heat exchanger with a longer bill of materials.
What degrades and what that costs to run
Recovery equipment adds surface, and surface fouls. On a dirty exhaust, particulate and soot build on the gas side of the exchanger and heat transfer falls long before anything is visible. The instrument that tells you is the temperature differential across the unit. Record inlet and outlet on both sides from commissioning, and treat a falling differential at constant load as the cleaning trigger rather than waiting for a calendar interval.
Two further operating costs belong in the model. An exchanger raises pressure drop on the exhaust path, which loads the induced draft fan and can require it to be resized. And flue gas cooled far enough will condense, which turns acid attack into a materials selection question rather than a maintenance one. Neither is a reason to skip the project. Both are reasons the maintenance line in the model should not be zero.
Where this shows up in reporting
A recovery project does not by itself satisfy any Mexican obligation, and no article should suggest that it does. It does change what a large site reports. Conuee's criteria for Usuarios de Patrón de Alto Consumo, published in the DOF on 15 November 2018, capture users consuming more than 45 GWh of electricity or more than 100,000 barrels of oil equivalent of fuel in a year. Those users report production, import, export and consumption data to Conuee between 1 March and 30 June each year, together with the energy saving measures they carried out and the results those measures produced. The measurement a payback case needs is the same measurement that report needs.
Find out whether your waste heat is worth recovering
Mexico Energy Partners measures the temperature and flow of a plant's main exhaust streams and identifies which of them clear the threshold for recovery. The review states what the recovered energy would displace and at what unit cost, then returns an installed cost band and a payback range. To start, you supply twelve months of electricity and fuel invoices, the process flow diagram, and the annual operating hours of the equipment producing the exhaust. We do not quote a percentage before the streams are measured, and the first output you receive is the stream inventory, not a proposal.
Sources
- U.S. Department of Energy, Quadrennial Technology Review 2015, Waste Heat Recovery Systems technology assessment, 2015. Waste heat temperature bands and the 276 TBtu and 1,478 TBtu annual waste heat figures at the 300 and 77 degree Fahrenheit reference temperatures.
- U.S. Department of Energy, Advanced Manufacturing Office, Use Feedwater Economizers for Waste Heat Recovery, Steam Tip Sheet 3, January 2012. Economizer fuel saving range, worked example and payback.
- Congress of the Union, Ley de la Industria Eléctrica, DOF, 11 August 2014. Generation permit threshold of 0.5 MW and the treatment of efficient cogeneration as clean energy.
- Comisión Reguladora de Energía, Acuerdo A/018/2023, DOF, 26 May 2023. Updated reference efficiency values and loss factors for cogeneration systems.
- CENACE, Reporte semanal del Mercado Eléctrico Mayorista, Sistema Interconectado Nacional, semana 34 de 2024, 18 to 24 August 2024. Average local marginal prices in the day ahead and real time markets.
- U.S. Energy Information Administration, Mexico publishes first monthly natural gas price index after moving to competitive market, 30 August 2017. Start date, units and construction of the IPGN.
- Comisión Nacional para el Uso Eficiente de la Energía, Disposiciones administrativas que establecen los criterios para determinar cuándo un usuario cuenta con un patrón de alto consumo de energía, DOF, 15 November 2018. UPAC thresholds and the annual reporting window.