--- title: "LoRaWAN Water Meter Field Test for Readings and Leak Alarms" canonical: https://wavect.io/blog/lorawan-water-meter-field-acceptance-test/ language: en description: "Field-test a LoRaWAN water meter before ordering. Compare the dial with the dashboard, try a difficult installation site and ask for a separate leak alarm test." image: "https://wavect.io/img/blog/headers/header_lorawan-water-meter-field-acceptance-test.png" author: ["Kevin Riedl"] published: "2026-11-12" updated: "2026-11-12" reviewed: "2026-11-12" --- # How can a water utility field-test a LoRaWAN meter reading and leak alarm? [**Back**](/blog/overview/) [![Kevin Riedl](/img/team/kevin.webp)](/team/kevin-riedl/) [Kevin Riedl](/team/kevin-riedl/) https://linkedin.com/in/wsdt 4 min read · 12 Nov 2026 Last reviewed November 12, 2026 [**Next**](/blog/lorawan-nbiot-sigfox-pilot-cost/) TL;DR Before ordering more meters, ask the supplier to show the same reading on the meter and in the dashboard at a typical site and one with difficult reception. Test the leak alarm separately using the method specified for that model. Record what the operator actually saw and which parts of the delivery path still need a field test. A technician at the meter can read out the figure an operator should see at the next scheduled update on screen. A leak alarm takes more work to verify. What makes this model declare a leak, when does it report one, and can the person on duty identify the meter? Before ordering a batch, I would ask the supplier to demonstrate both the reading and the alarm on the same sample meter. ## What did the Innsbruck field trial involve? Innsbruck’s municipal utility, IKB, was already leading the project when we joined. A meter reading could still mean a trip, even where staff could collect it from a passing vehicle. The utility wanted its citywide LoRaWAN network to spare those trips and give staff more recent readings. In an [interview published by IKB](https://www.ikb.at/karriere/einblicke/coole-smarties), a member of its telecommunications team describes responding earlier to leaks that customers had yet to notice. We tested water meters in an ordinary installation location and in a spot where we expected weaker reception. Cargo-bike trackers were among the other devices we worked with. We developed dashboards with IKB to help staff see where something might be wrong. Some messages arrived on a schedule. Others appeared only when an event occurred. The datasheet often matched what the device sent. With some devices, we only discovered a difference when we could compare it with a real packet. We kept real packets so we could replay them offline. Our [IKB case study](/case-studies/ikb/) describes the fieldwork. ## What should the supplier demonstrate before a volume order? For an acceptance test, I would ask operations to choose a typical installation and a representative location with difficult reception. Use the intended meter settings at both. Ask the supplier for the reporting interval, the condition that triggers a leak alarm and the point at which the meter transmits it. Have the utility agree how old a reading can be and how long the duty team can wait for an alarm. | Test | What to do | What to keep | | --- | --- | --- | | Meter reading | Read the register at a known measurement time and find the corresponding reading on screen. Repeat over the agreed reporting intervals at both sites. | Meter ID, value and unit at both ends, measurement time and time received. Record missed reports separately. | | Leak alarm | Use the manufacturer’s documented procedure on the test meter or test rig. Have the duty team watch the agreed alarm channel. | Trigger time, arrival time, meter ID and displayed alarm state. Record whether the operator could identify the installation to attend. | | Return to service | End the test and follow the meter’s documented procedure for clearing the alarm. | Time and final state on screen. Record who will follow up if the alarm remains open. | A test mode may let the meter send an alarm without exercising the sensor’s detection logic. That demonstration covers the reporting path from the meter onward. Ask the manufacturer what bench test would also establish detection. ## Where do you look when the dashboard stays quiet? First check when the report was due. A meter may carry an alarm flag in its next routine transmission. Once the expected time is clear, check whether the application received any packet from that meter around then. [IKB’s account of its LoRaWAN network](https://www.ikb.at/themenwelten/lorawan) describes the part that sensors and gateways play in delivery. If a packet arrived, compare the relevant field in the datasheet, the state you observed at the meter and the value your decoder produced. [The Things Stack documentation on payload formatters](https://www.thethingsindustries.com/docs/integrations/payload-formatters/) explains how transmitted bytes become readable fields. Our [LoRaWAN payload codec tester](/tools/lorawan-payload-codec-tester/) lets you run the same recorded packet through a decoder again. You can check a correction without waiting for the device to report another leak. Return to the installation to verify reception there. ## Which results justify placing the order? For each location, the buyer should be able to see how many readings were due, how many arrived and whether the duty team could identify the alarm within the agreed time. If readings arrive from both locations but the alarm arrives from only one, keep the second location open for a repeat test. A proposal to add a gateway needs two checks before the buyer accepts the quote. Can the operator secure permission to install it and get access later for maintenance? For teams still choosing a radio technology, our [pilot cost comparison](/blog/lorawan-nbiot-sigfox-pilot-cost/) covers that decision. Once the network is chosen, the [guide to scoping a pilot](/software-development-guide/what-is-a-discovery-phase/) can help you agree on the evidence you need before placing the order. To [arrange a field test](/contact/?intent=smart-city-iot) with our [municipal IoT team](/services/smart-city-iot/), bring a sample meter, its datasheet and two candidate sites. ## Questions from procurement and operations ### Does a fresh reading establish that the leak alarm works? It confirms that one reading made it through and was processed. To verify an alarm, record when the test triggered it and watch for the corresponding state on the operator’s screen. ### How long should we run a meter field test? Use the actual reporting interval, the meter’s detection period and the utility’s required response window. Agree beforehand how many scheduled transmissions to observe at each site. ### What can a saved packet tell us about a rare event? The team can run one recorded alarm through a decoder repeatedly and compare the result with the expected state. A field run then checks reception at the installation. ### Who should witness acceptance for the utility? The metering team checks identity, value and unit. The operator responsible for incidents demonstrates how to find the alarm, identify the installation and initiate the next step. ### Can we accept a meter after a bench demonstration? Record what the bench demonstrated and mark reception at the installation as untested. Take the same sample to the intended sites and repeat the reading and alarm checks before signing off on coverage. QA and production readiness ## Read next Testing, audits, maintenance and hardening practices for reliable production software. [Start here**QA for AI-Generated Code**](/blog/qa-for-ai-generated-code/) - [Both software vendors say they’re done. What should you test before signing off?](/blog/multi-vendor-software-acceptance-testing/) - [README ZIP malware: review the actual AI tool install](/blog/readme-zip-malware-install-review/) - [AI-generated Odoo modules: audit the business contract](/blog/ai-generated-odoo-module-audit/) - [Odoo migration acceptance: reconcile records and balances](/blog/odoo-migration-acceptance-reconciliation/) - [Odoo AI agent tests: validate saved activities](/blog/odoo-ai-agent-business-state-tests/) [Next article**Both software vendors say they’re done. What should you test before signing off?**](/blog/multi-vendor-software-acceptance-testing/) [**Back**](/blog/overview/) [![Kevin Riedl](/img/team/kevin.webp)](/team/kevin-riedl/) [Kevin Riedl](/team/kevin-riedl/) https://linkedin.com/in/wsdt 4 min read · 12 Nov 2026 Last reviewed November 12, 2026 [**Next**](/blog/lorawan-nbiot-sigfox-pilot-cost/)