Time-of-use tariffs are, conceptually, straightforward: electricity costs more during periods of high system demand and less during periods of low demand. In practice, the rate schedules that implement TOU pricing are anything but simple — they vary by season, by day type, by service voltage, by utility territory, and occasionally by rate version within the same utility. For fleet charging managers, the gap between understanding TOU in principle and actually scheduling against it in practice is where a significant portion of energy cost savings either materializes or evaporates.
This article covers the structure of commercial TOU rate schedules, the practical mechanics of scheduling fleet charging around them, and the limitations that fleet operators need to account for when building or evaluating an optimization approach.
Anatomy of a Commercial TOU Rate Schedule
A commercial TOU rate schedule typically defines four or five distinct pricing tiers: super off-peak, off-peak, partial-peak, peak, and sometimes a critical peak pricing (CPP) overlay. The exact nomenclature varies by utility, but the structure is consistent.
Under Portland General Electric's Schedule 32 (Large General Service) as an example, the billing period is divided into summer (June–September) and non-summer (October–May) rate seasons, each with separate peak and off-peak windows. The weekday on-peak window in summer runs from 7 AM to 10 PM. Off-peak applies all other hours and weekends. The energy charge differential between peak and off-peak periods in summer can be substantial — often a factor of 2x to 4x depending on the rate schedule version.
Pacific Power, serving eastern Oregon and parts of Washington, structures its commercial TOU rates differently. Its Schedule 47 for medium general service breaks the day into three periods — on-peak, mid-peak, and off-peak — with different windows on weekdays versus weekends, and different rate seasons than PGE's. A fleet operator with depots in both PGE and Pacific Power territory is literally working with two different rate calendars that cannot be managed identically.
Clark Public Utilities in Washington runs a different rate methodology entirely, based on flat residential-adjacent commercial rates with seasonal adjustments rather than discrete TOU windows. This is worth noting because it illustrates that "TOU optimization" means something different for each utility territory, and the assumption that every commercial customer is on a meaningful TOU rate is incorrect.
Reading Your Actual Rate Schedule
The first practical task is obtaining and parsing your actual rate schedule — not a utility marketing summary, but the tariff document filed with the relevant public utility commission. In Oregon, these are filed with the Oregon PUC. In Washington, with the Washington Utilities and Transportation Commission. The documents are public record and available through each commission's tariff database.
What you're looking for in the tariff document:
- Rate code and version. Confirm you're reading the schedule currently applied to your account. Your utility bill's rate code should match. Schedule versions change; the document you find online may be different from the version currently active on your account.
- Season definitions. Exact calendar dates, not "approximately June." Some schedules use billing cycle dates rather than calendar dates, which can shift the effective season boundary by days depending on your meter read date.
- Peak period definitions. Hours are typically in local time. Confirm whether the schedule uses Pacific Standard Time or Pacific Time (which adjusts for daylight saving).
- Holiday provisions. Most schedules treat holidays as off-peak days or weekend days. The list of qualifying holidays varies. Some schedules specify exact dates; others reference federal or state holiday designations that can themselves shift.
- Demand charge structure. TOU rates often carry both an energy component ($/kWh by period) and a demand charge component ($/kW measured monthly or by peak period). These interact. Optimizing purely for energy cost can inadvertently ignore the demand component and vice versa.
The practical problem with reading tariff documents is that they're written by utility engineers and regulatory counsel, not for the audience of fleet operators. The definitions are precise but dense. A 40-page rate schedule requires careful reading, and the critical details — like an exception clause that modifies peak hours for metering intervals measured at distribution vs. transmission voltage — are easy to miss.
The Optimization Window
For most fleet charging operations in the Pacific Northwest, the primary optimization target is the overnight off-peak window. Under a typical PGE commercial schedule, this runs from roughly 10 PM to 7 AM on weekdays (hours vary by season). Energy costs during this window are materially lower than peak hours — often by $0.05 to $0.12 per kWh depending on the schedule — and demand charges are typically lower as well.
The optimization problem is: given a set of vehicles that returned to depot between 3:00 PM and 5:00 PM with varying states of charge, dispatch them to reach their target SOC by 6:00 AM departure, while minimizing total energy cost and keeping peak demand below a defined threshold.
This sounds tractable, and for a small homogeneous fleet with predictable return patterns, it is. The complexity grows with:
- Fleet heterogeneity — mixed vehicle types with different battery capacities, charge rates, and departure times
- Operational variability — vehicles that return at unexpected times, with unexpected SOC deficits, due to route changes or unplanned service
- Partial-peak period management — some rate schedules have a partial-peak window from 7 PM to 10 PM that sits between full peak and off-peak, requiring a more nuanced dispatch strategy than a simple binary switch
- Demand charge windows — if the rate schedule includes a separate demand charge that measures the highest demand specifically during on-peak hours (coincident peak demand charge), then any charging that spills into the evening peak window creates a disproportionate cost even at off-peak energy prices
Static Schedules vs. Dynamic Dispatch
The simplest TOU optimization approach is a static charging schedule: program all chargers to activate at 10 PM and deactivate at 6 AM. This captures off-peak energy pricing and, for fleets that return in the afternoon, avoids on-peak charging almost entirely.
Static schedules have real limitations. They don't account for vehicles with critically low SOC that may need to begin charging immediately upon return despite peak-period pricing. They don't adapt when a vehicle returns at 9 AM instead of 4 PM. They don't adjust for seasonal rate changes when peak windows shift. And they don't manage demand — 30 vehicles all activating simultaneously at 10:00 PM creates a demand spike at 10:00 PM rather than at 4:00 PM, which may still trigger demand charges depending on the rate structure.
Dynamic dispatch addresses these limitations by treating the schedule as a real-time optimization problem rather than a fixed timer. Each vehicle's charging session is dispatched based on its current SOC, required departure SOC, departure time, current tariff period, and facility-level demand headroom. The schedule adjusts as conditions change throughout the night.
We're not saying static schedules are without value — for many smaller fleets on simple rate structures, they capture the majority of available savings with minimal operational complexity. What they can't do is adapt to the variability that characterizes real fleet operations, and that adaptation is where the incremental savings on more complex schedules are found.
TOU Rate Volatility and Schedule Maintenance
A frequently underestimated operational burden is keeping the scheduling system synchronized with current tariff data. TOU rate schedules are not static. Utilities file rate cases with their PUC, rates are approved or modified, seasonal windows shift, and rate schedule versions change. In Oregon, PGE's large commercial customers have experienced multiple rate case adjustments in recent years affecting both the energy charge tiers and the demand charge structure.
A scheduling system that was configured against an accurate tariff document in January may be operating against stale data by October if rate updates aren't propagated. The error mode is often silent: the system continues to schedule against the old peak window, occasionally charging during periods that are now classified differently, without generating an alarm.
The practical answer is to automate tariff data ingestion wherever utilities provide structured data feeds, and to implement alerting when tariff data hasn't been updated within an expected interval. The reality of utility tariff APIs is that data quality and timeliness vary considerably — but for fleet operators on consequential TOU rates, the operational discipline of maintaining current tariff data is non-negotiable.
Measuring TOU Optimization Impact
Fleet operators evaluating TOU optimization performance need a baseline comparison. The relevant question isn't "how much did we pay for energy?" but "how much did we pay relative to what we would have paid under unmanaged charging?" The counterfactual requires modeling what the load profile would have looked like without scheduling intervention — typically, concurrent charging immediately after fleet return, concentrated in the on-peak window.
A useful starting metric is the off-peak charging percentage: what fraction of total charging kWh was delivered during the off-peak period? For a well-optimized system, this should consistently exceed 80% for overnight-return fleets. Anything below 60% suggests either that vehicles are returning too late in the off-peak window to complete charging, that the system is failing to defer charging from on-peak periods, or that vehicle SOC deficits are large enough to require some on-peak supplementation.
The second metric is the demand-to-energy ratio on the charging system: the peak demand recorded by the charging system relative to total energy delivered. A high ratio indicates bursty, synchronized charging; a low ratio indicates smooth, spread charging that is easier to keep below demand charge thresholds. These two metrics together give a reasonably complete picture of TOU optimization health without requiring granular per-session cost attribution.
Taken together, TOU tariff optimization for commercial fleets is an ongoing operational discipline, not a one-time configuration. The rates change, the fleet changes, and the operational patterns change. Systems that treat it as a set-and-forget function tend to degrade in performance over time in ways that are difficult to detect until the utility bill arrives.