Energy-Management System Optimization: Baselines, Controls, Tariffs, and Verification
Optimizing an energy-management system (EMS) means defining the site objective, measuring the baseline, applying controls that respect comfort and process constraints, and verifying that savings actually persist. This guide covers the steps in order and the evidence each step requires. An EMS can reduce demand charges, shift load to cheaper periods, and coordinate solar, batteries, and generators — but every claimed saving needs a baseline and a measurement-and-verification method to be credible.
Define the site objective
Write the objective as a measurable statement: reduce peak demand below a target, shift a share of load out of a rate period, minimize imported energy, or keep critical loads running during outages. The objective decides which meters, controls, and tariffs matter. A vague objective (“become more efficient”) cannot be verified.
Interval-data baseline and meter validation
Build a baseline from interval data (15-minute or hourly) covering at least a full year where possible, normalized for weather and production. Validate meters and sensors before trusting them: check calibration, CT orientation, time synchronization, and that each measured value corresponds to the circuit it claims to represent. A baseline built on bad data makes every later comparison meaningless.
Control boundaries and constraints
Define what the EMS may control (schedules, setpoints, battery dispatch, generator starts, solar curtailment) and what it may not (safety systems, medical equipment, process-critical loads unless explicitly designed). Comfort and process constraints — temperature bands, ventilation requirements, production schedules — are hard limits, not suggestions.
Demand charges, time-of-use rates, and equipment schedules
Read the current utility rate: demand-charge structure, time-of-use periods, and any export or demand-response programs. Match equipment schedules to the rate periods where the change is cost-effective and allowed. Load-shifting only pays when the tariff delta exceeds the efficiency, comfort, and equipment-life costs of shifting.
HVAC controls, battery dispatch, solar curtailment, and generator integration
HVAC is usually the largest controllable load; pre-cooling or pre-heating shifts load only within the building’s thermal mass and comfort band. Battery dispatch should follow the rate structure and the battery’s cycle-life and warranty terms. Solar curtailment (reducing inverter output) is rarely optimal unless export is limited or prices are negative. Generators should run only when they are the least-cost or reliability-required option; fuel, maintenance, and emissions count as costs.
Measurement and verification
Measurement and verification (M&V) compares post-implementation energy against the adjusted baseline. The International Performance Measurement and Verification Protocol (IPMVP) and similar frameworks define options from whole-facility analysis to isolated-meter measurement. Report savings with the baseline period, adjustment factors, and uncertainty — not as a single unqualified number.
Alarm performance, failed sensors, and manual override
Alarms must detect real faults without flooding operators with noise; test alarm response with injected faults. Define failed-sensor behavior: what the EMS does when a meter stops reporting (hold last value, switch to a fallback schedule, alarm) and how operators are notified. Manual override must always be possible, logged, and time-limited so a bypass does not become permanent.
Cybersecurity
An EMS that can change schedules and dispatch equipment is an operational system: change default credentials, segment the network, keep firmware current, and control remote access. Treat the EMS like the utility and safety infrastructure it touches, not like a consumer gadget.
Commissioning and savings persistence
Commission the EMS against its design: every schedule, setpoint, dispatch rule, and alarm verified in operation. Savings persistence requires ongoing review — quarterly check that the baseline assumptions still hold, the tariff has not changed, and controls still match operations. Most savings leakage comes from changed operations, not failed hardware.
Worked facility example
Illustrative example with transparent assumptions: a small commercial building with 40 kW average demand, an 8-hour 2–10 p.m. peak period priced $0.04/kWh above off-peak, and a 50 kWh battery limited to one discharge cycle per day. Shifting 30 kWh/day out of the peak period saves 30 × $0.04 = $1.20/day in energy cost, before battery losses (typically 10–15%) and any demand-charge effects. If the same battery also reduces peak demand by 10 kW under a $12/kW-month demand charge, that adds $120/month. The example is arithmetic, not a promise: the actual tariff, load shape, and battery terms decide the result.