At 9:42 a.m. on a Monday in January, at a food plant in the Bajío region, two large compressors started simultaneously with a heat-treatment oven entering its cycle. The peak measured by CFE in that 15-minute interval was 1,450 kW —28% above the plant's average consumption for the rest of the month—. That Monday at that hour, the plant ended up committed to a peak demand charge that would bill for the next 30 days, even though for the rest of the month the plant did not exceed 1,150 kW. The over-cost of the incident was approximately MXN 92,000, recurring monthly unless actively managed.
This scenario is among the most expensive in Mexican industry, and one of the least understood. The CFE bill does not charge only for kWh consumed: it also charges for measured peak demand, which is determined by a single 15-minute interval per month. This article develops how it is measured, why a single peak dominates the monthly bill, and what strategies work in Mexican plants to control it.
What peak demand is in CFE billing
In CFE's industrial tariffs —GDMTH, GDMTO, DIST, and other medium and high voltage tariffs—, the bill has three main components:
- Energy charge — MXN per kWh consumed
- Contracted demand charge — MXN per kW of nominal demand contracted with CFE
- Measured peak demand charge — MXN per kW of the actual peak recorded in the month
The measured peak demand charge applies to the highest peak of the last 12 months in some tariffs, or to the current month's peak in others —depending on scheme and state—. In any case, it is a charge proportional to an instantaneous peak value, not to average consumption.
This means two plants with the same kWh consumption can pay very different peak demand charges, depending on whether their consumption is flat or has pronounced peaks.
How it is measured: 15-minute intervals
CFE measures demand in 15-minute intervals. The month's peak demand is the highest average recorded in any 15-minute interval during the billing period.
This has a critical implication: a 30-second peak does not generate the charge, but a sustained 15-minute peak does. This opens two different strategies:
- If your plant's peaks are very short (seconds to a few minutes), smart startup management can prevent them coinciding within the same 15-minute interval.
- If peaks are sustained, structural strategies are required: load scheduling, storage, or process adjustments.
To quantify this in your plant, you need the demand curve at 15-minute resolution —a data set CFE delivers to Qualified Users, and which for Basic Supply requires installing a smart meter or temporary analyzer—. It is one of the key data points a good energy audit must capture.
Why a single monthly peak dominates the bill
A numeric example illustrates the effect. A plant under GDMTH tariff pays approximately MXN 350 to 450 per kW of measured peak demand, per month (exact figure depends on state and tariff year).
If the plant has average demand of 1,100 kW but a monthly peak of 1,450 kW, the 350 kW of excess generate an additional charge of approximately MXN 140,000 per month, recurring. Flattening that peak —reducing it to 1,200 kW— releases approximately MXN 100,000 per month without touching total kWh consumption.
It is one of the few places where the behavioral lever delivers six-figure annual savings with zero equipment investment.
Control strategies: active vs passive management
Two complementary approaches exist:
Passive management — operational load design
Consists of structuring operation so peaks do not occur. Includes:
- Staggered startup scheduling — preventing large compressors, ovens, and main motors from starting simultaneously.
- Batch sequencing — separating high-consumption processes into different time slots.
- Operational limitation of redundant equipment — running 3 out of 4 units instead of 4 simultaneously, when capacity allows.
- Maintaining conservative setpoints — preventing backup systems from starting under momentary dips.
This strategy requires no investment, only procedure and discipline.
Active management — real-time control systems
When operation is complex, an active system is justified:
- Demand response controllers — equipment that monitors demand in real time and triggers actions (shutting down non-critical loads, modulating processes) when demand approaches a setpoint.
- BMS systems with demand-limiting logic — useful in plants with dominant HVAC where modulation is possible without affecting production.
- Energy storage as a buffer — batteries or thermal storage that absorb peaks.
Active management requires investment and careful operation, but it is the only path when peaks are inherent to the process and cannot be scheduled.
Staggered load scheduling: the most profitable option
For plants with rigid schedules and predictable startups, this is the best-ROI option. The methodology:
- Map every major startup —compressors, ovens, chillers, large welding equipment—.
- Document peak startup power vs nominal operating power.
- Design schedule matrix preventing simultaneity of major startups within the same 15-minute interval.
- Implement via PLC, BMS, or operating procedure —based on available infrastructure—.
- Verify result in next bill —if measured peak fell, the strategy works; if not, processes were missed—.
To go deeper into KPIs that help monitor this management, read Essential Energy KPIs —load factor is a direct indicator of demand management health—.
Real-time monitoring technologies
Without continuous measurement, demand management is blind. Options ordered by cost:
- Submeters on main circuits with telemetry — modest option, sufficient for mid-sized plants.
- BMS system with energy module — integrates monitoring with automated control.
- Third-party energy management platform — specialized software connected via open protocols.
- Integrated solution with supplier (for Qualified Users) — some suppliers offer monitoring platforms as part of service.
Measurement should reach at least the main circuit and dominant loads —submetering everything is unnecessary, but covering what moves the needle is essential—.
Energy storage as a buffer: when it makes sense
Storage (mainly lithium batteries) has dropped in price but remains significant capex. It makes sense in specific scenarios:
- Plants with short, predictable peaks where the cost of primary equipment does not allow operational adjustment.
- Plants with critical production where flattening peaks via operational adjustment affects output.
- Plants where storage serves dual purpose (backup + demand management).
In most mid-sized Mexican industrial plants, operational and scheduling measures deliver more savings per peso invested than storage. Worth modeling case by case.
How much you can save: realistic range
Achievable peak demand reduction depends on how much non-structural peak the plant has. Typical ranges in Mexican industrial plants:
- Plants with load factor below 40% — potential reduction of 15% to 30% in measured peak demand.
- Plants with load factor between 40% and 60% — potential reduction of 5% to 15%.
- Plants with load factor above 60% — small margin; peak is already well managed.
Translated to the bill, peak demand charge reduction can represent 3% to 8% of total electricity bill, without touching kWh consumption.
Common errors managing peak demand
- Assuming that reducing total consumption reduces the peak — not necessarily. A plant can consume fewer kWh and still have high peaks if startups concentrate.
- Not differentiating contracted demand from measured peak demand — they are different charges, with different logics.
- Implementing without verifying in subsequent billing — without closing the measurement loop, you don't know if the strategy worked.
- Waiting for the peak at month-end — the peak is built interval by interval. Management is continuous, not reactive.
Next step
Peak demand management is one of the most profitable behavioral levers. It fits within the broader framework described in the Strategic Guide to Industrial Energy Optimization. To complement the technical lever, read Top 10 Measures with Best ROI and Power Factor: Why CFE Penalizes You.
If you want a peak demand–specific analysis of your plant —with quantification of recoverable savings and a management plan—, request an evaluation. We work with real data from your latest billing, not generic estimates.
