8.1 Context: Why Lowering Peak Demand Can Unlock Faster Service and Lower Cost
Peak-load reduction is one of the most powerful levers in a power access acceleration project because utilities must plan for the highest demand the facility may place on the system, not only its average use. A facility that averages 6 MW but occasionally spikes to 14 MW can create a larger grid problem than a facility that runs steadily at 8 MW. The peak may determine feeder capacity, transformer size, substation loading, protection requirements, customer contribution, bridge power sizing, and the timeline for energization.
For many facilities, the initial peak is higher than it needs to be. It may reflect simultaneous startup of equipment, unmanaged HVAC operation, overlapping shifts, uncoordinated battery or EV charging, refrigeration cycles, compressor behavior, process batch timing, or automation systems that all ramp at once. Some of these peaks are real operating requirements. Others are artifacts of poor sequencing, conservative assumptions, or lack of controls.
Lowering peak demand can change the utility conversation. A utility that cannot serve 20 MW today may be able to serve 11 MW under a phased or limited operating plan. A customer that can hold demand below a defined cap may qualify for earlier partial service, interruptible service, or a limited operating agreement. Even when full service still requires upgrades, a lower near-term peak can reduce the size and cost of bridge power.
This workstream is not about starving the business of power. It is about using power more intelligently. The objective is to preserve the highest-value operations while reducing avoidable peaks, shifting discretionary loads, and making the facility more predictable. The best peak-reduction programs are designed jointly by operations, engineering, controls, and finance. They ask what the business truly needs in each hour, not just what the installed equipment could draw if everything ran at once.
Importantly, peak reduction is different from energy efficiency, although the two can overlap. Energy efficiency reduces total consumption. Peak reduction reduces the highest point of demand. A facility may use the same total energy in a day but create a lower grid impact if some work moves from a constrained afternoon window to overnight or if equipment starts are staggered.
The work should also distinguish a real production constraint from a control problem. Some peaks are unavoidable because the process requires them. Others exist because schedules were never coordinated across departments. The second category can often be improved quickly and with limited capital.
The financial benefits can also be meaningful. Many commercial and industrial tariffs include demand charges based on monthly peak demand. Reducing peak load may therefore lower ongoing utility bills, not just accelerate access. Peak reduction can also reduce generator size, battery size, transformer capacity, electrical room requirements, and emissions from temporary generation.
The critical executive mindset is to treat peak demand as a manageable operating variable. It should not be accepted as a fixed number until the team has tested whether equipment can be sequenced, loads can be shifted, controls can cap demand, and business processes can be redesigned around available power.
8.2 Methodology: Identifying Avoidable Peaks, Sequencing Equipment Starts, and Shifting Discretionary Loads
The methodology begins with the load model from Chapter 2 and the criticality analysis from Chapter 4. The team should identify not only the size of the peak, but what creates it, how long it lasts, how often it occurs, and whether the business truly requires it.
Step 1: Establish the baseline demand curve. The demand curve shows facility load across time. For an existing facility, it may come from interval meter data, submeters, controls systems, or utility data. For a new facility, it is built from equipment schedules, operating assumptions, weather sensitivity, production plans, and commissioning scenarios. The baseline should show hourly demand for typical days, peak days, seasonal periods, and ramp-up phases.
Step 2: Locate the peak events. The team should identify the highest demand hours and determine whether they occur daily, weekly, seasonally, or only during special events. A two-hour peak that occurs during shift change requires a different solution than a continuous afternoon cooling peak or a seasonal refrigeration peak.
Step 3: Decompose the peak by load category. The peak should be broken into major contributors, such as HVAC, refrigeration, process equipment, compressed air, conveyors, robotics, pumps, data center IT load, charging, lighting, and office loads. This prevents the team from pursuing generic energy savings while ignoring the few loads that actually create the constraint.
Step 4: Identify avoidable coincidence. Coincidence occurs when multiple loads operate at the same time. Some coincidence is necessary; some is avoidable. The team should look for equipment that starts together, processes scheduled for convenience rather than need, charging that overlaps with production peaks, or HVAC recovery that occurs at the same time as shift startup.
Step 5: Sequence equipment starts. Large motors, chillers, compressors, conveyors, pumps, and process systems can create short peaks during startup. Sequencing starts over several minutes or hours can reduce the maximum draw without changing total work performed. This may require control changes, operator procedures, variable frequency drives, or revised commissioning routines.
Step 6: Shift discretionary loads. Some loads can move to lower-demand hours. Examples include EV charging, battery charging, water pumping, ice making, certain batch processes, nonurgent packaging, warehouse replenishment, and some cleaning or maintenance activities. The test is not whether the load is useful; it is whether it must run during the peak window.
Step 7: Use thermal flexibility where available. Facilities with cooling, refrigeration, or chilled water may be able to pre-cool, store thermal energy, or temporarily relax noncritical temperature bands within safe and compliant limits. This must be designed carefully, especially where product quality, humidity, worker comfort, or regulated conditions matter.
Step 8: Implement demand controls. Manual procedures are useful, but automated controls are often required for sustained performance. A building management system, energy management system, microgrid controller, or industrial control system can monitor demand and reduce selected loads before a cap is exceeded. Control logic should respect the load priority stack and avoid unsafe interruptions.
Step 9: Model operating cases. The team should test several demand caps, such as 5 MW, 8 MW, 12 MW, or 15 MW, and describe what the facility can do under each. Each case should show business output, critical loads served, loads shifted, loads curtailed, operational compromises, labor impact, customer impact, and risk.
Step 10: Convert the plan into a utility-facing proposal. If the company can operate under a defined peak limit, it should present the utility with a credible load-shaping plan. The proposal should include the demand cap, operating windows, controls, metering, curtailment process, responsible operators, and evidence that the facility can comply.
Step 11: Govern performance after launch. Peak reduction is not a one-time spreadsheet exercise. Once the facility operates, the team should monitor actual demand, compare it with the forecast, investigate exceptions, update controls, and report performance. If the company has committed to a utility demand cap, governance is essential.
A practical peak-reduction opportunity list should include each opportunity, affected load, estimated peak reduction, implementation cost, operating impact, timeline, reliability risk, owner, and whether the measure is manual or automated. The best opportunities are usually those that reduce peak without reducing revenue-generating output.
Before implementation, each measure should be tested in an operating review. The review should ask who controls the load, how the change is triggered, what happens if the control fails, how operators are trained, and how the measure is reversed. Peak reduction only works if the field team can execute it repeatedly under real operating pressure.
8.3 Skills Required: Operations Improvement, Controls Engineering, Industrial Engineering, Facility Energy Analytics
Peak-load reduction requires both operating judgment and technical analysis. Operations improvement expertise is needed to understand how work is actually performed. An operations specialist can identify whether a process can be resequenced, whether a shift can start differently, whether manual workarounds are practical, and whether a proposed load shift will create bottlenecks elsewhere.
Control engineering is critical because many peak-reduction measures depend on automatic or semi-automatic control. Controls engineers design logic for equipment sequencing, demand caps, load shedding, alarms, operator overrides, and safe restart. They also integrate building systems, industrial controls, chargers, batteries, generators, and metering where relevant.
Industrial engineering helps translate electrical demand into throughput, labor, queueing, utilization, and cost. A load-shaping plan that reduces demand but disrupts production flow may not be acceptable. The industrial engineer tests whether revised operating windows still meet customer, quality, and productivity requirements.
Facility energy analytics is required to build and interpret the demand curve. The analyst identifies peak drivers, models weather and production sensitivity, evaluates scenarios, and measures actual results after launch. For complex facilities, this role may need submetering knowledge and familiarity with building management systems or energy management platforms.
The team may also require process engineers, refrigeration specialists, data center operators, fleet charging specialists, maintenance leaders, EHS, quality, finance, and utility-facing advisors. Finance estimates the value of peak reduction through lower bridge power cost, lower utility charges, or earlier revenue. EHS and quality ensure that no measure compromises safety, product integrity, or compliance.
8.4 Terminology: Peak Load, Coincidence Factor, Demand Curve, Ramp Rate, Load Shifting, Peak Shaving, and Diversity Factor
Peak load: The highest level of electrical demand during a defined period. Peak load is often the number that drives utility capacity requirements, demand charges, and bridge power sizing.
Coincidence factor: A measure of how many different loads occur at the same time. A high coincidence factor means many loads peak together; a lower factor means peaks are spread out.
Demand curve: A time-based view of electrical demand, often shown by hour or shorter interval. The curve helps identify when peaks occur and which operating patterns create them.
Ramp rate: The speed at which electrical demand increases or decreases. A fast ramp can create operational or grid challenges even if the peak is short.
Load shifting: Moving electrical demand from one time period to another, usually from constrained or high-cost hours to lower-demand hours.
Peak shaving: Reducing the highest demand level by controlling loads, using batteries, changing operations, or dispatching on-site resources during peak periods.
Diversity factor: A measure of the difference between total connected equipment capacity and the actual combined demand expected at one time. Diversity reflects the fact that not all equipment runs at full output simultaneously.
Demand cap: A defined maximum level of demand that the facility agrees or attempts to stay below through controls and operating procedures.
Managed charging: Scheduling or controlling chargers so vehicles, batteries, or equipment do not all charge during peak periods.
Pre-cooling: Cooling a space, product, or thermal system before a peak period so cooling load can be reduced temporarily during constrained hours.
