Battery Life for IoT Sensors

Battery Life for IoT Sensors

Battery life for IoT sensors determines whether a deployment is operational and cost‑effective. This article explains how device chemistry, firmware patterns, radio features (PSM/eDRX, ADR), and field‑proven energy budgets combine to deliver predictable multi‑year lifetimes for municipal and industrial sensors, and how to test and validate vendor claims.

Designing for IoT sensor battery life

Designing predictable battery life is an engineering exercise that ties hardware, firmware, radio, and site reality together. The high‑level pattern is always the same: prioritize race‑to‑sleep, minimize wake events, and verify per‑event energy with lab traces and season‑long field logs.

• Start at the hardware level: choose regulators and PMICs with low quiescent current and use power gating to remove entire subcircuits when idle. See practical patterns in low‑power IoT design and consider OTA firmware update strategies that let you iterate energy improvements in the field.
• Pick the chemistry to match the duty: coin cells (CR2032) are good for tiny indoor beacons; for outdoor pit or vault sensors, prefer primary Li‑SOCl2 in AA/D formats or a rechargeable 18650 with solar energy for frequent uplinks. See battery life for IoT sensors guidance.
• Choose the right radio/profile: for long sleep with rare uplinks use NB‑IoT or LTE‑M and exploit PSM/eDRX; for low‑duty, long‑range private networks pick LoRaWAN and use ADR and short payloads to reduce time‑on‑air.
• Mind the network: PSM/eDRX configuration and the MNO’s profile directly affect months/years of life — GSMA guidance on Mobile IoT (NB‑IoT/LTE‑M) explains recommended PSM/eDRX usage and trade‑offs. (gsma.com)
• When adding intelligence on the node, measure the on‑device inference cost: TinyML reduces uplink energy only when inference mJ_cost < saved_tx_mJ; otherwise batch or push inference to the gateway. Use edge computing patterns to decide placement.

For quick budgeting, use the simple battery life formula:

Battery life (years) ≈ (Battery capacity in mAh ÷ Average current in mA) ÷ (24 × 365)

Adjust for temperature derating and self‑discharge when you move from lab to field.

Why Battery Life for IoT Sensors Matters in Smart Water Management

Battery life drives OPEX, alarm continuity, and feasibility of large fleets: a sensor that needs swapping every 12 months across thousands of remote assets quickly becomes unaffordable. Cities pursuing smart water or flood early‑warning systems need multi‑year devices so crews can focus on analysis and intervention rather than routine replaces. The EU Smart Cities evidence base emphasises scalable, low‑maintenance infrastructure as a cornerstone of resilient urban services. (smart-cities-marketplace.ec.europa.eu)

Standards and regulatory context

Complying with regional radio and safety rules limits what you can configure and therefore affects battery life. Key references to consult when writing procurement requirements:

• LoRaWAN and regional parameters — duty cycle, channel plans, and regional constraints that change time‑on‑air and legal TX behavior. (lora-alliance.org)
• Mobile IoT / PSM & eDRX best practices — how to exploit long sleep intervals for NB‑IoT/LTE‑M without losing critical reachability. (gsma.com)
• EU product and radio compliance: RED 2014/53/EU, ETSI EN 300 220 for SRD devices where applicable, and IEC 62368‑1 for battery safety.

Tuning your device inside these limits — e.g., LoRaWAN SFs, network dwell times and PSM timers — is where procurement teams save years of battery life.

Background and field evidence

Lab traces are only half the story. A multi‑node telemetry study published in Applied Sciences (2025) measured average device consumption around 11.8 mAh/day on a remote metering platform; when that field profile is used to size a primary pack, multi‑year life becomes realistic, but only if retries, GNSS fixes and cold‑weather derating are included in the model. (mdpi.com)

Note the lab‑to‑field gap: real site RSSI/RSRP, retries, and temperature cycles can multiply per‑packet energy compared with ideal lab runs — demand raw Joulescope traces and seasonal logs in procurement.

Practical implications and procurement checks

• Battery chemistry: avoid CR2032 for outdoor pit sensors because coin cells derate heavily at low temperatures; prefer Li‑SOCl2 AA/D or 18650 where recharge or solar is available. See solar energy harvesting if you can supplement primary cells.
• Protocol tradeoffs: LoRaWAN often minimizes energy per packet at short payloads and good RF conditions, while NB‑IoT/LTE‑M can beat it for built‑in coverage and roaming when PSM/eDRX is available. For operator-profile details see GSMA Mobile IoT guidance. (gsma.com)
• KPIs to insist on in tenders: average current (µA) in deep sleep, per‑event energy (mJ), p95 wakes/day, retry counts, and raw Joulescope traces replicable in the lab and in the field.
• Cold weather planning: use a −25°C winter profile and add a 20–40% margin for coin cells; Li‑SOCl2 primary cells derate less but are still affected by electrochemistry.
• Analytics: implement SoH estimation and battery degradation models in your IoT platform to schedule swaps and reduce emergency truck rolls.

Key takeaway from FLOPRES (flash‑flood early‑warning pilot) The FLOPRES pilot in eastern Slovakia/Poland showed field efficiency: initial installations (6 sensors) were completed in under 20 minutes per location by a two‑person team, and the programme planned expansion across dozens of villages — a reminder that mechanical design and installation workflows significantly affect TCO. (blog.meratch.com)

Key takeaway from the Danube simulated‑flood trial Twelve NB‑IoT level sensors delivered millimetre‑level measurements with hourly uplinks and a five‑year battery target in the Danube floodplain pilot, illustrating how conservative hourly cadences and good coverage can enable multi‑year life. (blog.meratch.com)

How Battery Life for IoT Sensors is implemented: step‑by‑step

1. Define service goals: required operating life per asset class (e.g., 5+ years), message cadence, and alert latency.
2. Select stack & operator: choose LoRaWAN, NB‑IoT or LTE‑M and map expected coverage to site locations.
3. Choose chemistry & package: coin cell for tiny indoor beacons; Li‑SOCl2 AA/D for outdoor low‑message sensors; 18650 for rechargeable/solar installs.
4. Architect firmware for race‑to‑sleep: deep sleep current optimization, power gating, and minimal wake branches (low‑power IoT design).
5. Configure radio: enable ADR, minimize confirmed messages on LoRa, and set long TAUs/PSM and appropriate eDRX cycles on cellular — validate the MNO profile. (gsma.com)
6. Measure per‑event energy: capture wake‑to‑sleep traces with a µA power profiler (Joulescope or similar) and compute energy‑per‑packet (mJ).
7. Build the battery calculator spreadsheet: include ambient temperature bins, self‑discharge, and reserve margins.
8. Pilot across coverage tiers: collect retries, RSSI/SNR/RSRP, and winter behavior.
9. Integrate analytics: SoH estimation, trend detection and battery degradation models to schedule swaps.
10. Lock procurement deliverables: certified regulatory reports (EN 300 220 / RED / IEC 62368‑1), reproducible Joulescope traces, and a field pilot dataset spanning one winter.

Summary

Deliverable battery life is the product of conservative engineering, sound procurement, and real field data. Exploit PSM/eDRX and ADR, choose chemistry by duty and climate, insist on raw energy traces, and run a season‑long pilot to validate lab claims — this combination converts vendor lifetimes into operational predictability. (gsma.com)

References

• FLOPRES – Flash Flood Prediction System (Malá Poľana, Svidník area). Initial phase: 6 sensors installed in 2024 with an expansion target of 60 villages (Feb 2025). Notable: two‑person installs under 20 minutes; role: local early‑warning for rural communities. (blog.meratch.com)
• Danube River Floodplain Monitoring (Slovakia, 2024): 12 NB‑IoT high‑precision water level sensors used for simulated flood management. Outcomes: millimetre accuracy, hourly automated uploads and a 5‑year battery target. (blog.meratch.com)
• Bratislava Wastewater Management (Bratislava, 2023–2024): radar‑based IoT sensors and repeaters for underground shafts that enabled data‑driven wastewater operations and real‑time alarms. (Municipal partner: BVS). See Meratch case material.
• Residential Septic Tank Monitoring (Slovakia, 2024): single radar sensor and app integration for one homeowner delivering real‑time capacity monitoring and reduced unnecessary pump‑outs. (blog.meratch.com)
• BVS Bratislava partnership (Podunajské Biskupice, Lafranconi Bridge): enabling real‑time wastewater alerts for systems handling population‑equivalent flows (case study posted on Meratch).

Frequently Asked Questions

1. How is battery life calculated and validated? Use measured wake‑to‑sleep traces (µA profiler), compute energy per event (mJ), build a capacity vs. average‑current spreadsheet, then validate with a season‑long pilot that captures retries and cold weather.

2. LoRaWAN or NB‑IoT for 15‑minute leak alarms — which wins? Choose by coverage and required latency. LoRaWAN often wins when private network range and low per‑packet airtime are achievable; NB‑IoT/LTE‑M is better where operator coverage is strong and PSM/eDRX can be exploited to save energy. Validate with per‑packet mJ and operator PSM profiles. (gsma.com)

3. Which eDRX/PSM settings give the best battery vs reachability balance? There’s no single answer — GSMA guidance shows long eDRX cycles (hours) and PSM with long TAU values maximize savings, but you must match those timers to your alert SLA and the MNO’s supported values. (gsma.com)

4. Can you use a CR2032 in outdoor pits? Generally no. Coin cells derate quickly in cold, have low capacity for frequent uplinks, and are vulnerable to self‑discharge. Prefer AA/D Li‑SOCl2 or rechargeable packs for outdoor environmental sensors.

5. What makes a credible 10‑year battery claim? Provide raw Joulescope traces, a temperature‑binned battery model, field logs across one winter, R‑site retry statistics, and independent lab certification. Without these, long‑life claims are not tender‑grade evidence.

6. When does TinyML extend battery life? On‑device inference helps only if inference energy < energy saved from avoided transmissions. Benchmark both sides with representative workloads and use edge computing to decide where inference should run.

---

Ready to convert claims into verifiable numbers? Meratch helps model battery life, design pilots, and produce procurement pack deliverables (Joulescope traces, lab reports, and field datasets). Contact Meratch early to cut OPEX and de‑risk rollouts.

Author Bio

Ing. Peter Kovács, Technical Freelance writer

Ing. Peter Kovács is a senior technical writer specialising for smart‑city infrastructure. He writes for water management engineers, city IoT integrators and procurement teams evaluating large tenders. Peter combines field test protocols, procurement best practices and datasheet analysis to produce practical glossary articles and vendor evaluation templates.