Peak demand management is the discipline of monitoring, limiting and spreading out the highest average power (demand) a facility draws from the grid. An electricity bill is not driven by energy consumption (kWh) alone; how much power (kW) the facility requests from the grid at the same moment also shapes the final cost. Of two factories consuming exactly the same amount of energy, the one that spreads its load over time pays noticeably less, while the one that switches all of its large loads on simultaneously faces higher demand charges and may exceed its contracted power. In this article we walk through what peak demand is, how the 15/30-minute averaging logic works, why peaks occur, and how an energy monitoring system helps you bring them down step by step.
What Is Peak Demand?
Demand is the average power a facility draws from the grid within a defined time window, expressed in kW (or kVA). The critical detail is that it is an average, not an instantaneous reading. In line with common practice worldwide, electricity meters record consumption in fixed intervals — typically 15 or 30 minutes. The energy consumed in each interval is divided by the interval length to obtain that period’s average power. The interval with the highest value in the billing period is then recorded as the month’s peak demand (maximum demand).
A simple example: if a facility consumes 250 kWh within a single 15-minute interval, the average power of that interval is 250 ÷ 0.25 hours = 1,000 kW. Among the hundreds of intervals recorded during the month, suppose the highest one — say, at the start of the morning shift — reached 1,400 kW. The facility’s peak demand for that month is then 1,400 kW, even if it ran at an average of 800 kW for the rest of the month.
Instantaneous Power vs. Demand
The few seconds of high inrush current a motor draws at start-up do not, by themselves, set the demand value, because the average is taken over a 15- or 30-minute window. But when several large loads run together within the same window, the average of that window climbs and a peak is created. That is why peak demand management is less concerned with second-by-second fluctuations and far more concerned with a single question: which loads are running together within the same quarter hour?
Same Total kWh, Different Cost — Why?
An electricity bill is not a single line item. Alongside the cost of the energy itself, distribution charges and — depending on your tariff — power- and demand-related components appear on the invoice. We covered the line items in detail in our guide to industrial electricity bill analysis. For this topic, the essential distinction is this: energy charges depend on how much you consume, while demand-related charges depend on how fast, at the same time you consume it.
The hypothetical comparison below shows why two plants with identical monthly consumption can end up with different costs:
| Parameter | Plant A (spread load) | Plant B (clustered load) |
|---|---|---|
| Monthly consumption | 500,000 kWh | 500,000 kWh |
| Recorded peak demand | 900 kW | 1,500 kW |
| Load factor* | High | Low |
| Demand-related cost | Lower | Higher |
*Load factor: the ratio of average power to peak power. The closer it is to 1, the more evenly consumption is spread over time.
Both plants took the same total energy from the grid, but Plant B forced the network to reserve 1,500 kW of capacity for it. The grid operator has to keep that capacity available even outside the single quarter hour in which the peak occurred, and tariffs are designed to pass this cost on to the user. That is why reducing peak demand is one of the rare improvements that can lower your bill without saving a single kWh.
Contracted Power and What Happens When You Exceed It
Industrial consumers declare a contracted power (agreed demand) in their connection agreement with the utility. This value represents the upper limit the facility commits to drawing from the grid, and it serves as the reference both for sizing the connection infrastructure and for calculating demand-related billing components.
The optimization works in both directions:
- Exceeding contracted power: When the recorded peak demand exceeds the contracted value, additional charges or penalties may apply. How they are applied — and how much they cost — depends on your connection type and your tariff, so you should confirm the current terms with your supplier and network operator.
- Contracting too much power: Conversely, if the facility has committed to a power level it never comes close to, it may be carrying an unnecessarily high fixed cost for capacity it does not use.
The only reliable way to set the right contracted power is a data-driven review of the historical demand profile. What range have the peaks of the last 12 months fallen into? Do peaks come from a recurring operational habit, or from exceptional events that happen a few times a year? Without answers to these questions, any revision of contracted power is guesswork.
How Do Peaks Occur?
In industrial facilities, peak demand is most often the result of unplanned simultaneity. The most common scenarios are:
- Large loads starting at the same time: High-power equipment such as furnaces, compressors, chillers and presses starting within the same quarter hour creates a demand level none of them could produce alone.
- Shift start-up: When production lines, lighting, ventilation and process equipment are all switched on within minutes at the start of the morning shift, the day’s highest demand interval usually lands in the first half hour of work.
- Thermal loads drifting into sync: Thermostat-controlled furnaces, cooling plants and HVAC systems, left uncoordinated, can gradually start cycling on and off at the same time. This “unintended synchronization” produces recurring demand spikes.
- Recovery after an outage: Restarting every system simultaneously after a power cut or planned maintenance can write the month’s highest peak in a single event.
- Power factor and power quality issues: Where demand is assessed in kVA, a low power factor means a higher apparent power demand for the same active power.
The common thread: very few peaks are genuinely unavoidable. Most exist because nobody has a complete picture of what is running when.
The Role of a Monitoring System: You Cannot Manage a Peak You Cannot See
By the time the invoice arrives, peak demand is historical data you can no longer act on. Without knowing in which quarter hour, and because of which loads, the peak was set, every countermeasure is based on guesswork. This is where an energy monitoring system becomes the backbone of demand management, providing three critical capabilities:
- Real-time demand visibility: The system shows live how the average power of the current measurement interval is developing. Operators can see peak risk before the interval closes.
- Peak alerts: When demand approaches a defined threshold — for example, 90% of contracted power — the system raises an automatic alarm, so the team can hold back deferrable loads and keep the interval average under the limit.
- Historical peak analysis: Overlaying the demand profiles of past months reveals at which hours, on which feeders and with which equipment combinations peaks occur. Contracted power reviews and automation scenarios are built on this analysis.
The ATS Energy Monitoring System combines panel- and machine-level measurements with live demand tracking, threshold-based peak alerts and historical demand reports on a single platform — turning peak management from an educated guess into a measurable process.
Peak Reduction Strategies
Once the demand profile is visible, the main strategies you can apply are the following:
Load Shifting
Load shifting means moving energy-intensive tasks that are not time-critical from the facility’s natural peak hours to lower-load hours. Charging stations, water pumping, tank heating, certain batch processes and test or quality equipment are typical candidates. Without changing total consumption, load shifting can reduce both the peak demand and — under time-of-use tariffs — the average unit cost.
Sequential (Staggered) Start-Up
At shift start or during recovery after a shutdown, large loads are brought online in sequence, minutes apart, instead of all at once. A simple start-up procedure — which line switches on at which minute — is the fastest peak reduction measure in most plants and usually requires no investment at all. Soft starters and variable frequency drives support this strategy further by smoothing start-up loads.
Moving Deferrable Loads Out of Peak Periods
Every facility has loads that can tolerate a delay of minutes or even hours: the compressor’s next cycle when the air receiver is full, the chiller’s next start while the cold store is within its temperature band, forklift charging, wastewater transfer pumps. When these loads are held back while demand is climbing and released in low-demand intervals, the peak profile flattens noticeably. For facilities on time-of-use tariffs, scheduling these loads outside the most expensive hours adds a further cost advantage.
Demand-Limiting Automation Scenarios
In advanced applications, the monitoring system acts as a demand controller: when live demand approaches the defined limit, loads are automatically shed in stages according to a predefined priority list. Comfort loads go first (a temporary HVAC setback), then deferrable process loads — while critical production loads are never touched. When the interval closes, loads return to normal automatically. Because it does not depend on a human watching a screen, this approach keeps the demand limit consistently protected throughout the month.
Example Scenario: The Effect of Staggered Start-Up
The following calculation is entirely hypothetical; unit prices and tariff structures vary by facility, so you should validate the real impact with your own tariff and measurement data.
Assume a metalworking plant consumes 400,000 kWh per month. Monitoring data shows that the month’s peak demand hits 1,300 kW every morning in the 07:30–07:45 interval, when two compressors, the induction furnace and the ventilation system all start simultaneously — while demand never exceeds 950 kW for the rest of the day.
- Measure: A start-up sequence is defined: ventilation at 07:15, the first compressor at 07:25, the furnace at 07:35, the second compressor at 07:50.
- Result: Since all large loads no longer start within a single interval, the monthly peak drops from 1,300 kW to roughly 1,000 kW. Total consumption is unchanged.
- Impact: Demand-related billing components are now based on ~1,000 kW instead of 1,300 kW, and a data-driven revision of a contracted power originally sized for 1,300 kW can be put on the agenda. Because this improvement is achieved with little or no investment, it feeds directly into the payback calculation of the monitoring system itself — a topic we explored in our article on energy monitoring system ROI and payback.
KPIs to Track in Peak Demand Management
To make peak management sustainable, we recommend tracking the following indicators regularly; for the full set of energy indicators, see our guide to energy monitoring KPIs and metrics:
- Monthly maximum demand (kW/kVA): The headline indicator; the goal is that it does not rise unless production does.
- Load factor: Average power ÷ peak power. A rising load factor shows consumption is being spread more evenly over time.
- Peak-to-contracted-power ratio: How close you get to the contracted limit; values above 90% should be defined as an alert threshold.
- Number and duration of exceedances: How many intervals per month cross the threshold; the target is to drive this toward zero.
- Time-of-day distribution of peak intervals: Which shifts and hours the peaks cluster in — the key input for prioritizing strategies.
- Share of consumption in peak tariff periods: Under time-of-use tariffs, the percentage of consumption still falling in the most expensive hours.
Frequently Asked Questions
What is the difference between peak demand (kW) and consumption (kWh)?
Consumption (kWh) is the total amount of energy used over a period; demand (kW) is the average power that shows how fast that energy is drawn from the grid. The bill is affected by both: a facility consuming the same kWh with higher simultaneity faces higher demand-related charges.
Why is demand measured as a 15- or 30-minute average?
Second-long inrush currents are not meaningful for grid capacity planning; what stresses the network is sustained high power demand. Meters therefore divide consumption into fixed-length intervals and record each interval’s average power. Depending on the metering infrastructure and local practice, the window is typically 15 or 30 minutes.
What happens if I exceed my contracted power?
If the recorded peak demand exceeds your contracted power, additional charges or penalties may apply; their scope and amount depend on your connection agreement and tariff. The soundest approach is to confirm the current terms with your supplier and network operator, and then review your contracted power against your historical demand data.
Does reducing peak demand slow down production?
Not when it is designed properly. Peak management does not aim to curtail production loads; it aims to spread deferrable and shiftable loads over time. Staggered start-up, coordination of thermal loads and automation scenarios flatten the demand profile without touching critical processes.
Do I need a monitoring system for peak alerts?
The monthly invoice shows only the single highest peak of the month — not which day, which interval or which loads caused it. Intervening before an interval closes requires live demand data and threshold-based alerts, which is only possible with a monitoring system measuring at panel and machine level.
To make your facility’s demand profile visible, activate peak alerts and optimize your contracted power based on real data, contact Atasayın Enerji — we will evaluate the ATS Energy Monitoring System on site at your facility.