Demand spikes happen when large loads overlap for a short time and set a high maximum demand for the billing period. A factory can have normal monthly units and still pay extra because one poor start-up sequence pushed MD upward.
This is a scheduling and control problem before it is a capex problem. First find the overlap, then decide whether equipment changes are needed.
How do peaks form even when units look normal?
Units measure energy over time. Demand measures how high the load rises during a short billing measurement window defined by the metering system and tariff.
That difference matters. A short overlap can hurt demand without adding many units.
Typical peak-forming events include:
- furnace or oven heating during compressor recovery
- multiple injection moulding machines starting together
- chilled water pumps, AHUs and process loads starting after lunch
- borewell, transfer pump and production motor overlap
- welding sets used during a heavy production batch
- DG changeover and restart confusion after a power cut
The owner sees a bill and asks why payable amount rose. The engineer should ask what happened during the highest-demand period.
For the bill concepts, read contract demand and maximum demand with the contract demand guide.
Why do simultaneous starts create avoidable MD?
Many factories restart everything after a break because it feels efficient. In reality, the meter sees a pile-up.
A few minutes of convenience can set the billing demand. Compressors recover pressure, pumps fill tanks, heaters warm up, machines return to auto mode and lighting or HVAC comes back together.
This is common after:
- morning start
- lunch break
- power failure
- weekly shutdown
- maintenance isolation
- product changeover
The fix may be as simple as a start sequence. Start utilities first, then process loads, or stagger heavy loads based on actual process need. The correct sequence depends on safety and production, so do it with the plant team.
How do batch loads and utilities overlap?
Batch processes create natural peaks. Heating, mixing, melting, drying and pressing do not draw steady power across the day. Utilities can make the peak worse if they run at the same time.
For example, an oven may need warm-up, while the compressor is recovering from a leak-heavy line and the cooling water pump is running in hand mode. Each load is explainable alone. Together they create a demand event.
Ask these questions:
- Which loads are process-critical at the same moment?
- Which support utilities can wait safely?
- Is any load running because of manual habit?
- Does a timer start equipment at the same clock time?
- Does the demand peak match a shift change or break?
This is where production planning and energy management meet. See load factor for production planning if peaks are frequent across the month.
What soft scheduling fixes should be tried first?
Start with low-risk operational fixes. Do not buy a demand controller before understanding the process.
Useful actions include:
- stagger start-up of heavy motors and heaters
- avoid compressor pressure recovery during major process starts
- move water transfer to lower-load periods where possible
- schedule maintenance welding away from peak production starts
- stop non-essential HVAC or exhaust during short high-load windows
- review auto-start timers that all trigger together
Some fixes are only minutes of discipline. Others need control wiring or operator training.
A demand alarm can help, but alarms without authority become background noise. Someone must know what to shed or delay when demand approaches the limit.
Tips from the field
- Ask for the exact date and time of maximum demand from meter data before changing schedules.
- Put machine start times beside the demand curve, because the overlap is easier to see visually.
- Watch the first hour after lunch, as many plants restart utilities and process loads together.
- Separate essential process starts from convenience starts, then stagger only the convenience loads.
- Check whether compressor recovery after leaks is adding to the peak.
- Review automatic timers after power cuts, because many controls restart at the same time.
When is equipment control justified?
Equipment control is justified when scheduling discipline is not enough or when the process has repeatable peaks that controls can manage safely.
Options include soft starters, VFDs, staged heater control, demand monitoring, automatic load sequencing and utility interlocks. Each has a place, but none should be sold as a magic fix.
Before approving controls, confirm:
| Question | Why it matters |
|---|---|
| Is the peak repeatable? | One-off events do not justify complex controls. |
| Can the load be delayed safely? | Some process loads cannot be interrupted. |
| Who will respond? | Monitoring without action does not reduce MD. |
| Will quality suffer? | Demand savings must not create scrap or rejection. |
How do you prove demand scheduling worked?
Compare maximum demand before and after the scheduling change across similar operating periods. Also check whether production, product mix and power interruptions were comparable.
The monthly bill should show the demand effect, but interval data gives the better proof. If the peak shape becomes flatter and the plant still meets output, the scheduling change has done its job.
Demand work is often invisible to operators because the machine still runs. Make the result visible in the monthly review, so the start sequence does not slowly drift back to old habits.