Coexisting Modbus meters, BACnet plant controllers, KNX lighting and MQTT sensors are the norm in buildings, and the multi-protocol gateway is the key to heterogeneous integration. This article breaks down the four real jobs of a gateway (southbound polling engine, protocol translation and data-model mapping, northbound multi-interface serving, store-and-forward and time base), provides a five-dimension selection scoring table (southbound driver nativeness / northbound interface match / point capacity & performance / edge capability / engineering tooling) and three integration architectures (rip-and-replace / gateway bridging / edge aggregation + unified platform), plus five pre-acceptance details: measured polling, point-table engineering, time sync, store-and-forward boundaries, and the security baseline.

Modbus TCP and RTU share one application layer; only the transport differs. This article clarifies RS-485 capabilities and limits (load, distance, single-master constraint), the three structural advantages of Modbus TCP (concurrent multi-client access, bandwidth decoupled from distance, reuse of IT infrastructure), provides a five-question field decision table, a three-layer "485 at the edge, TCP on the trunk" mixed architecture, five deep pitfalls of mixing (Unit ID misconfiguration, multi-master tunnel conflicts, one-size-fits-all timeouts), plus a scan-cycle estimate for 32 meters at 9,600 bps.

KNX is the only open building-control protocol adopted as both an international standard (ISO/IEC 14543-3) and a Chinese national standard (GB/T 20965), widely used in hotel guestroom control, luxury residences and lighting retrofits. This article unpacks KNX's three core concepts (group objects, group addresses, four media), the hard 64-devices-per-line topology limit, its layered complementarity with Modbus and BACnet, and three field-proven pitfalls: bus power budgeting, group address planning, and ETS source file handover.

Energy efficiency requires more than installing meters. This handbook covers the complete closed loop: sub-metering, EUI benchmarking, operational optimization, and M&V verification — five practical steps to get your loop running before heating season.

Spec sheets are easy to read, but field failures don't care about datasheets. The real dividing line in sensor selection is scenario understanding, not accuracy figures. This pillar guide walks through a five-step selection method across five building scenarios — HVAC plant rooms, indoor air quality, power distribution and sub-metering, ducts and hydronics, and special applications — closing with a parameter-based quick-reference table and three field-proven pitfalls.

Protocol selection is the first fork in any smart building project. This pillar guide compresses our Q3 protocol series into one comparison table covering Modbus, BACnet, MQTT, and OPC UA, plus a three-question decision tree (device layer, data consumer, network conditions), three proven implementation routes, three engineering pitfalls, and Zhiyinghui's full-stack product mapping from sensing to platform.

Many energy monitoring projects shine at acceptance and are ignored three months later—the root cause is an unanswered question on day one: who is it for? This article breaks the build-out into six stages: tripartite-signed requirements, site survey and frozen end-use point tables, sensor and acquisition deployment, data integration with unified clocks, dashboard and alarm design, and a 30-day trial with checklist-based acceptance—plus six frequent pitfalls and the engineering discipline behind each.

In industrial parks and large public buildings, SCADA handles power and utilities while BA runs HVAC and lighting—two systems, two screens. This article presents an integration architecture: compare the DNA of both systems, choose one of three integration patterns by project scale, design tiered point uplinks, collection cycles, time sync and alarm tiering, with an interface checklist and acceptance points.

When an energy monitoring system goes live but nobody looks at the data, the problem is usually the dashboard, not the data. This article presents four design principles—audience-first metrics, four-layer drill-down from overview to device, chart selection with baselines, and proactive anomaly reporting—compares custom front-end, commercial BI and BMS-native tool routes, and includes a 90-day rollout plan plus five common pitfalls.