Load curve & consumption profile

The load curve represents the evolution of the power demand (kW) over time (hourly, 10-minute, 30-minute, etc.), while the consumption profile describes recurring patterns (typical day, week, season). Together, they reveal the base load (unavoidable consumption), the peaks (power surges), the troughs, seasonality (summer/winter), and the impact of usage patterns (HVAC, process, lighting). Two key figures to remember are: the cumulative energy (kWh) under the curve and the maximum power (kW), which determines the contract and equipment size. The load factor (kWh / (Pmax × duration)) indicates whether the installation is being used efficiently or at very high peak demand.

Load curve: definition and interpretation

A standard reading begins with aggregation by time interval (10/15/30 min) followed by a comparison of day/week/season. The following are identified:

- Base: continuous consumption (servers, ventilation, standby).

- Start-ups & peaks: morning starts, DHW, cooling, air conditioning during heat waves.

- Time zone differences: weekends, night, periods of inactivity.

- Seasonality: heating (winter), air conditioning/cooling (summer).

These elements guide:

- sizing (HVAC power, sections, storage),

- the contractual choice (subscribed power, peak/off-peak hours, dynamic signals),

- control (load shedding, load shifting, preheating, precooling),

- the opportunity for self-consumption/storage and BMS/EMS (start-up rules, setpoints, alarm thresholds).

Advantages, limitations and points of attention of load curves

Interests

• Optimize your bill: adjust your contracted power and take advantage of peak/off-peak hours.
• Reducing peaks: clipping via control/storage, smoothing by sequencing.
• Better sizing: avoid oversizing (CAPEX) and underperformance.
• Target the gains: identify unnecessary operations (nights, weekends).
• Negotiate: support a CPE or PPA with solid data.

Boundaries

• Incomplete data (breakdowns, bad timing) = biased analyses.
• Usage variability (occupancy rate, weather): requires standardization.
• Risk of false causality without monitoring of technical states (GTB).
• Limited gains if the loads are not controllable or critical.

Points to consider

• Measurement quality: calibration, clock synchronization, no stable time.
• Correlation: linking curves to weather, occupancy, HVAC instructions.
• EMS strategies: load shedding thresholds, offset rules, preheating.
• Contracts: penalties for exceeding limits, pricing options, minimum/maximum power.
• Safety & comfort: no erasure that degrades IAQ, comfort or process.
• Indicators: load factor, off-peak energy, peak/base ratio.

Anecdote — “The morning peak that was costing Lille dearly”

In Lille, an office building was experiencing excessive power consumption. The power curve showed a sharp peak at 8:30 a.m.: HVAC, hot water, and lighting all started simultaneously. By sequencing the restarts (ventilation at 7:50 a.m., preheating at 8:05 a.m., lighting at 8:25 a.m.) and slightly lowering the initial temperature setting, the peak was reduced by 18%. The result: reduced contracted power, avoided penalties, and unaffected comfort. The moral of the story: sometimes, a 30-minute difference is worth more than a new chiller.