Analyze consumption, demand, and energy costs

Use the right quantity

Energy and power answer different questions:

• Consumption energy, measured in kWh, is the amount used over a period.
• Load or demand, measured in kW, is the rate of use at a point or interval.
• Net energy or power is consumption minus production for the selected measurement scope.

Use energy to compare totals and costs. Use power to find peaks, low-load periods, and operational demand patterns. Before beginning, confirm the component, measurement direction, units, and coverage.

Step 1: Establish a consumption total

1. Start a new chat.
2. Insert the intended site or component.
3. Choose a complete period.
4. Ask for consumption in kWh and request the resolved dates and component list.

For example:

For @Main Warehouse / Grid Meter, calculate total consumption for August 2026. State the exact start and end timestamps, unit, included component, and any missing data.

If the result covers a whole site, confirm that parent, child, and calculated components were not counted twice.

Step 2: Reveal the daily or hourly pattern

Ask Vega to prepare a breakdown at a granularity suited to the question:

• hourly for operating schedules and overnight use,
• daily for weekday and weekend differences,
• weekly for short-term trends,
• monthly for seasonality and long-term reporting.

Example:

Show daily consumption for the same component and period as a table. Include the total for each day and identify incomplete days.

Then ask for a chart if the shape of the pattern matters. A line chart is usually helpful for hourly data; bars are often clearer for daily or monthly totals.

Step 3: Find peak demand

Power data can be aggregated within each time bucket using maximum, minimum, or average values. State which one you need:

Show the maximum load in kW for each day in August 2026 for @Main Warehouse / Grid Meter, then list the five highest daily peaks with timestamps.

Do not call a monthly energy total “peak demand.” Confirm that the answer uses power in kW and a maximum statistic. To distinguish a short spike from sustained demand, follow up with:

For each of those peaks, show the surrounding hourly values and explain whether the load was brief or sustained.

Step 4: Estimate baseload carefully

Vega does not know the site’s true unavoidable baseload without operational context. Use a measurable proxy and label it as such:

1. Define non-operating hours and relevant days.
2. Ask for minimum and average load during those windows.
3. Compare several complete weeks.
4. Check whether scheduled equipment, charging, heating, or production affected the period.

Example:

For the same meter, show average and minimum hourly load from 00:00 to 05:00 on complete weekdays in August. Treat this as a baseload proxy, not a confirmed avoidable load.

Step 5: Analyze energy cost from invoices

Use imported invoice data for actual billed cost rather than multiplying measurements by an assumed price.

For @Main Warehouse / Grid Meter, summarize invoices overlapping January through June 2026 by billing period. Show consumption, net and gross cost, cost per kWh, currency, and any partial-overlap warning.

Vega includes a partially overlapping invoice in full; it does not prorate it. Keep currencies separate and use invoice reconciliation when billed and measured quantities must be checked.

To investigate a cost change, ask Vega to separate volume from unit-cost effects:

Compare the latest invoice with prior comparable invoices. Show changes in consumption, cost per kWh, total cost, and cost per billing day. State the historical median and number of baseline invoices.

Step 6: Turn findings into checks

Ask Vega to rank observations by evidence and potential impact, without presenting estimated savings as guaranteed:

Based only on the measured pattern and imported invoices, list the three strongest cost-reduction opportunities. For each, show the supporting values, operational assumption, additional check required, and whether savings can be quantified from available data.

Typical investigations include reducing out-of-hours load, shifting flexible activity away from coincident peaks, correcting abnormal equipment schedules, and resolving measurement or billing issues.

Validate the result

Before acting:

• compare equivalent complete periods,
• check kW versus kWh,
• verify local timestamps and daylight-saving boundaries,
• confirm coverage and overlapping intervals,
• keep hypotheses separate from observed values, and
• follow Verify Vega’s answers before acting.