Sambhrama BH-3 · MODEL-VALIDATED PERFORMANCE

AI Performance & Benefit Evidence

Measured performance, AI-identified opportunity, actions and verified outcomes.

CURRENT FIELD STATUS

ONLINE

Last field telemetry: 2026-09-11T15:37:21

Current field telemetry is reachable and fresh.

HISTORICAL EVIDENCE AVAILABILITY

Predictive performance history: AVAILABLE

Evidence period:: 2026-09-05 to 2026-09-11
Last field telemetry:: 2026-09-11T15:37:21
Data coverage:: 440 operating samples across 7 days
Evidence class:: MODELLED

Authoritative Pilot 2 measured actual and modelled expected history remains available independently of current edge connectivity.

PILOT 2 PREDICTIVE PERFORMANCE EVIDENCE

Expected Flow AI validation available

Measured water delivery is compared with the learned expected-flow baseline to assess pump performance and identify potential water-delivery shortfall.

This is predictive/model-validation evidence. It is not an executed AI-control test and does not by itself establish measured energy savings.

MEASURED: Actual field water delivery.
MODELLED / AUTHORITATIVE AI: Expected water delivery from Expected Flow AI.
OPPORTUNITY EVIDENCE: Measured-versus-expected deviation that may indicate a performance opportunity. No intervention executed means this is not verified AI benefit.
VERIFIED AI BENEFIT: NOT YET ESTABLISHED. Requires authoritative executed-intervention and before/after outcome evidence.

OBSERVED PERFORMANCE

182.1 LPM average measured flow

AI OPPORTUNITY IDENTIFIED

18.5 LPM modelled shortfall

VERIFIED AI BENEFIT

NOT YET ESTABLISHED

Last 10 Days — Benefit Evidence

Up to 10 recent calendar days supported by authoritative history. Missing days are not fabricated.

Evidence period:: 2026-09-05 to 2026-09-11
Days with valid evidence:: 7/10
Coverage:: 70%

Date	Operating Hours	Observed Performance	Expected Performance	AI / Test Action	Outcome	Evidence Result
11 Sept 2026	—	182.1 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	200.6 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	18.5 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 51 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 51 daily operating samples
10 Sept 2026	—	235.4 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	241.1 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	5.7 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 65 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 65 daily operating samples
9 Sept 2026	—	222.5 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	233.4 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	10.9 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 61 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 61 daily operating samples
8 Sept 2026	—	283.1 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	279.3 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	0 LPM MODELLED Measured actual compared with modelled expected Measured performance agrees with the model; this is not a saving.	OBSERVED Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 66 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 66 daily operating samples
7 Sept 2026	—	186 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	197.8 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	11.8 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 65 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 65 daily operating samples
6 Sept 2026	—	231.9 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	232.3 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	0.4 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 65 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 65 daily operating samples
5 Sept 2026	—	256.2 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	266.9 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	10.7 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 67 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 67 daily operating samples

Evidence classifications

MEASURED: Direct field sensor, VFD or RMS telemetry.
MODELLED: Calculated from an existing trained or authoritative model.
ESTIMATED: Indirectly calculated because direct instrumentation is absent.
VERIFIED: Outcome supported by an executed intervention and before/after evidence.
DATA REVIEW: Anomaly, inconsistency or insufficient data makes the result unsuitable for a benefit claim.
UNAVAILABLE: Required evidence does not exist.

Opportunity is never counted as verified benefit unless an intervention was executed and supported by before/after outcome evidence.