EnergyEfficiencyEvents¤

Energy efficiency against production and machine state: output per kWh over time, energy used while the machine is idle, specific energy consumption (kWh per unit), and a comparison across shifts.

ts_shape.events.energy.efficiency_tracking · Engineering guide · API reference

Example¤

import numpy as np
import pandas as pd
from ts_shape.events.energy.efficiency_tracking import EnergyEfficiencyEvents

t = pd.date_range("2024-01-01", periods=7 * 96, freq="15min")
running = (t.hour >= 6) & (t.hour < 22)
df = pd.concat([
    pd.DataFrame({"systime": t, "uuid": "meter:main", "value_double": np.where(running, 12.0, 3.0)}),  # kWh deltas
    pd.DataFrame({"systime": t, "uuid": "counter:parts", "value_integer": np.cumsum(running * 30)}),
    pd.DataFrame({"systime": t, "uuid": "machine:running", "value_bool": running}),
], ignore_index=True)

ee = EnergyEfficiencyEvents(df)
waste = ee.idle_energy_waste("meter:main", "machine:running", window="1h")
print(f"Idle energy: {waste['waste_energy'].sum():.1f} kWh")
sec = ee.specific_energy_consumption("meter:main", "counter:parts", window="1D")

Methods¤

Method	Returns
`efficiency_trend(meter_uuid, counter_uuid, window='1h', trend_window=24)`	per window: `energy`, `units`, `efficiency` (units per kWh), `rolling_avg_efficiency`, `trend_direction`
`idle_energy_waste(meter_uuid, run_state_uuid, window='15min', idle_threshold=0.0)`	per window: `energy_consumed`, `machine_running_pct`, `machine_running_ratio`, `is_idle_waste`, `waste_energy`
`specific_energy_consumption(meter_uuid, counter_uuid, window='1D')`	per window: `total_energy`, `total_output`, `sec` (kWh per unit), `sec_trend`
`efficiency_comparison(meter_uuid, counter_uuid, shift_definitions=None)`	per shift: `avg_energy`, `avg_output`, `avg_efficiency`, `total_energy`, `total_output`

Notes¤

The meter must carry kWh deltas (see PowerIntegrator for kW signals). The state signal is boolean, True = running; counters are reset-aware.
idle_energy_waste books a window's energy as waste when the machine ran less than 10 % of it and it consumed more than idle_threshold. The run state is held until the next change (time_weighted=True); time_weighted=False averages the samples.
rolling_avg_efficiency is the trailing trend_window windows taken as one period (Σunits / Σenergy), and sec_trend compares each window's SEC with the trailing 7 windows' Σenergy / Σoutput (ISO 50006), not with the mean of the per-window ratios, which over-weights low-output windows.
efficiency_comparison defaults to three 8-hour shifts from 06:00.