2.1 Context: Why the Load Request Must Be Credible, Detailed, and Defensible
The first major workstream in a power access acceleration project is to determine how much power the facility truly needs, when it needs it, and how that demand will change over time. Many companies begin with only a rough number: “we need 10 MW,” “we need 25 MW,” or “full buildout could require 80 MW.” That number may come from a developer, architect, equipment supplier, automation vendor, construction team, or internal planner. It may be based on nameplate ratings, rules of thumb, early design assumptions, or a desire to reserve future capacity.
For a utility, a rough estimate is not enough. The utility must understand not only the maximum requested load, but also the timing, shape, certainty, flexibility, and electrical characteristics of that load. It is not simply deciding whether to “hook up” the building. It is evaluating whether feeders, substations, transformers, protection systems, voltage levels, and sometimes upstream transmission infrastructure can serve the facility without harming reliability or other customers.
A vague or inflated request can create delay. If a company says it needs 30 MW, the utility will want to know whether that is needed on day one or at full buildout, whether it reflects connected equipment or realistic operations, whether the peak occurs during system peak hours, and whether any load can be shifted, curtailed, or phased. If the company cannot answer those questions, the utility may pause the review, request more information, or make conservative assumptions. Conservative assumptions often mean larger upgrades, higher cost, longer study timelines, and slower energization.
A credible load request becomes the basis for utility studies, service agreements, upgrade estimates, bridge power sizing, capital approvals, permitting, and operating plans. The goal is not to minimize the request artificially. The goal is to make the request accurate, phased, and defensible.
A utility-grade load forecast should answer four executive questions: What power do we need to start operating? What power do we need at phase one, later expansion, and full buildout? What drives our peak demand? Which loads are critical, flexible, deferrable, or uncertain?
Executives do not need to perform the engineering analysis themselves, but they must insist on rigor. A strong load forecast converts scattered assumptions into a fact base that can be used with the utility, engineers, vendors, regulators, and internal decision makers.
2.2 Methodology: Building the Hourly, Daily, Seasonal, and Ramp-Up Demand Profile
A power demand diagnostic should move quickly from a rough estimate to a structured, utility-facing forecast. The work can begin with imperfect data, as long as assumptions are documented and improved as better information becomes available.
The methodology typically follows eight steps.
Step 1: Define the operating scenarios. The forecast should begin with the business plan, not the equipment list. Define commissioning, day-one operations, minimum viable operations, phase one, later expansion, full buildout, emergency mode, and curtailed mode. Each scenario should have a clear meaning. “Phase one” may mean one production line, two shifts, half the automation system, or one data hall. Without this clarity, day-one needs and future expansion loads are easily mixed together.
Step 2: Collect the load data. Gather site plans, electrical drawings, one-line diagrams, equipment lists, vendor specifications, automation plans, HVAC schedules, production plans, commissioning schedules, and utility correspondence. For existing operations, collect utility bills, interval meter data, building management system data, production records, and seasonal operating history. For new facilities, the forecast may need to be built from design assumptions and comparable sites.
Step 3: Build the equipment-level load inventory. List major electrical loads by category, such as HVAC, refrigeration, process equipment, motors, pumps, fans, compressed air, conveyors, robotics, lighting, IT, data center equipment, EV charging, life safety, laboratory equipment, and future expansion. For each category, distinguish between rated capacity and expected operating demand. A motor, chiller, compressor, or charger may have a maximum rating that is much higher than its normal draw.
Step 4: Apply diversity and operating assumptions. Diversity reflects the fact that not every piece of equipment operates at full load at the same time. Refrigeration cycles on and off. HVAC depends on the weather. EV charging can often be scheduled. Production equipment may run by shift. Document operating hours, duty cycles, coincidence assumptions, seasonal factors, startup sequences, and utilization levels. Assumptions should be traceable to vendor data, engineering judgment, comparable facilities, or operating plans.
Step 5: Build the hourly load profile. A single peak number is not enough. The company and utility need to understand how demand changes across the day. At minimum, create profiles for a typical weekday, weekend day, summer peak day, winter peak day, commissioning day, full production day, and curtailed operating day. For larger facilities, build an 8760 model, which estimates demand for every hour of the year. This helps reveal whether the facility has steady load, short spikes, seasonal peaks, or periods when storage, load shifting, or curtailment could help.
Step 6: Model seasonal variation. Many facilities do not use power the same way all year. Cooling, refrigeration, holiday logistics peaks, production campaigns, and data center cooling loads can all create seasonal variation. Identify monthly peak demand, monthly energy use, weather-sensitive loads, seasonal operating changes, and whether the facility’s peak coincides with the utility’s system peak.
Step 7: Build the ramp-up forecast. Most new facilities do not reach full demand immediately. A warehouse may add automation modules over time. A factory may commission one line at a time. A data center may energize data halls as customers deploy. The ramp-up forecast should show expected demand by month or quarter, separating commissioning load, initial operations, phase one, later expansion, and full buildout. This may allow the company to request partial service earlier while the utility plans later upgrades.
Step 8: Classify loads by criticality and flexibility. Critical loads must remain powered to avoid safety, product loss, customer disruption, or major financial impact. Essential loads support normal operations but may tolerate limited interruption. Flexible loads can shift to other hours. Deferrable loads can move to a later phase. Curtailable loads can be reduced during grid events. This classification supports phased operations, battery sizing, backup generation, load shedding, demand response, and utility negotiation.
After these steps, validate the forecast through comparison with similar facilities, vendor review, design engineer review, operations review, throughput reconciliation, and sensitivity testing. Each major assumption should receive a confidence level: high, medium, or low. It is better to acknowledge uncertainty than to present early planning assumptions as final facts.
The final output should be a utility-facing load package. At minimum, it should include an executive load summary, requested service date, phase-by-phase demand table, equipment-level load inventory, hourly profile, seasonal profile, ramp-up forecast, critical and flexible load summary, expected power factor, planned on-site generation or storage, major assumptions, open questions, and a single technical contact for utility follow-up.
A concise phase-by-phase view should show, for each phase, the date needed, expected peak demand, average demand, critical load, flexible load, and major assumptions. The phases should typically include commissioning, minimum viable operation, phase one, phase two, and full buildout.
Common failure modes include adding all nameplate ratings without diversity, treating future expansion as day-one demand, ignoring startup loads, overlooking power factor, using operating assumptions that conflict with the business plan, failing to separate critical from flexible load, and sending the utility an unvetted number.
2.3 Skills Required: Power Systems Engineering, Energy Analytics, Facility Operations, and Utility Documentation
A power demand diagnostic requires complementary skills. It should not be assigned only to a financial analyst, architect, construction manager, equipment vendor, or utility account representative. Each may contribute, but none usually owns the full picture.
The first required skill is power systems engineering. A power systems engineer understands how facility loads interact with electrical infrastructure and the utility grid. This person can interpret equipment ratings, voltage levels, transformers, switchgear, feeders, motors, backup systems, one-line diagrams, protection requirements, power factor, and power quality issues. For larger facilities, the engineer should often be a licensed Professional Engineer, especially where technical documents support permitting, utility submissions, or construction.
The second required skill is energy analytics. The analyst turns equipment data, operating schedules, production assumptions, weather sensitivity, and historical information into an hourly demand model. This role requires time-series analysis, scenario modeling, sensitivity testing, and clear visual communication.
The third required skill is facility operations knowledge. The forecast must reflect how the facility will actually run. Operations leaders know which equipment runs continuously, which equipment cycles, which processes are shift-based, which loads are flexible, and how the facility will ramp up. Without operations input, the forecast can look precise but be wrong.
The fourth required skill is utility documentation experience. A strong internal model must be translated into a format the utility can process. This includes service applications, load letters, technical exhibits, assumptions logs, meeting materials, and responses to utility questions. A former utility planner, interconnection specialist, owner’s engineer, or experienced electrical consultant can help avoid incomplete or inconsistent submissions.
Specialized facilities may require additional expertise. Data centers need knowledge of IT load, UPS systems, cooling architecture, redundancy, and data hall ramp-up. Cold-storage facilities need refrigeration expertise. Factories need process engineering input. EV fleet depots need managed charging expertise. Automated warehouses need knowledge of conveyors, robotics, sortation, and equipment sequencing.
Useful credentials may include a licensed electrical PE, Certified Energy Manager, commissioning experience, familiarity with the National Electrical Code, NFPA 70E electrical safety practices, IEEE power systems standards, and prior utility interconnection experience. Credentials matter, but practical experience with utilities and operating facilities matters more.
2.4 Terminology: Connected Load, Diversified Load, Peak Demand, Base Load, Load Factor, kW, kWh, kVA, and 8760 Model
Connected load is the total rated electrical capacity of all equipment connected to the facility’s electrical system. It is often based on equipment nameplates. Connected load is useful, but it usually overstates actual demand because not all equipment runs at full rating at the same time.
Diversified load is the estimated demand after accounting for how the facility actually operates. It reflects the fact that equipment cycles, shifts, stages, or runs at partial load. Diversified load is often more relevant to utility planning than connected load.
Peak demand is the highest level of power demand during a defined period. It is usually measured in kilowatts or megawatts. Peak demand matters because electrical infrastructure must be sized for the highest expected draw, not just average use.
Base load is the demand that runs continuously or nearly continuously. It may include life safety systems, core IT, refrigeration, process controls, security systems, minimum HVAC, and continuously running pumps or fans.
Load factor describes how steadily a facility uses power over time. A high load factor means demand is relatively steady. A low load factor means the facility has large peaks compared with average demand. Low-load-factor facilities may have more opportunity for peak shaving or load shifting.
kW, or kilowatt, measures real power: the useful power that runs equipment, motors, lights, servers, compressors, and chargers. One megawatt equals 1,000 kilowatts.
kWh, or kilowatt-hour, measures energy used over time. A facility using 1 kW for one hour consumes 1 kWh. kW is the rate of use; kWh is the quantity used over time.
kVA, or kilovolt-ampere, measures apparent power. Apparent power includes real power and reactive power. Utilities care about kVA because transformers and electrical equipment must be sized for apparent power, not only useful power. Poor power factor can increase kVA requirements.
The 8760 model is an hourly forecast for all 8,760 hours in a non-leap year. It shows how demand changes across daily cycles, seasonal conditions, ramp-up periods, and operating scenarios. It is useful for identifying peaks, sizing batteries or generators, evaluating solar, and understanding whether load flexibility can help.
Related terms include nameplate rating, the manufacturer’s rated equipment capacity; duty cycle, how often equipment runs; coincident peak, when the facility’s peak occurs at the same time as the utility’s system peak; power factor, how effectively electrical power is converted into useful work; curtailable load, load that can be reduced under defined conditions; firm load, load expected to be served without planned interruption; and load shed, the intentional reduction of selected loads to stay within a power limit.
