When an energy monitoring project kicks off, requests like this are common: put a temperature and humidity sensor in every room, a pressure tap on every pump, and make everything accuracy class 0.5. It sounds safe, but the budget usually balloons by more than thirty percent, and less than half of the extra data is ever actually used. Sensor placement is not a case of more density being better or more expensive accuracy being more reliable; it is a triangular trade-off among density, accuracy, and cost. Too many points wastes budget; too few leaves you blind at critical moments. This article offers a quantifiable analysis method to help you spend every unit of budget where the information value is highest during the design phase.
Each additional measurement point yields progressively less new information. Take an office building as an example: the first temperature and humidity point in the lobby carries the highest value; by the tenth point, covering major floors and orientations, the marginal return is still acceptable; but going from ten to twenty, most new points fall within the correlation range of existing points, and the decisions managers can actually make barely change. A rule of thumb: when the distance from a new point to its nearest neighbor falls below the characteristic correlation scale (typically eight to fifteen meters in office buildings), its independent information contribution is under ten percent. In other words, the goal of placement design is to find the tipping point where the information gain just covers the added cost, not to blanket every room. The practice of zoned placement of temperature-humidity transmitters by functional area shows that keeping one representative point per temperature zone in ordinary office areas, verified against return-air temperature, satisfies most comfort-control needs while saving thirty to forty percent of the budget.
Each step up in accuracy roughly doubles or triples the purchase cost. Ordinary HVAC control loops typically need temperature within plus or minus one degree and humidity within plus or minus five percent; thousandth-level accuracy is reserved for laboratory calibration and custody transfer. Installing 0.5-class sensors in ordinary air handling units costs more and is not necessarily more accurate, because installation-induced errors (radiation, stagnant airflow, sensor self-heating) often exceed the sensor's intrinsic accuracy. The more economical engineering practice is to allocate accuracy by purpose: routine control points use standard accuracy, sub-metering points move up one class, and only settlement-grade points get high accuracy. After a central plant retrofit project re-allocated its point list this way, sensor procurement spending dropped about twenty-eight percent while energy accounting error stayed within two percent.
Grouping points into three tiers by decision value is the most effective cost-control method. Critical points govern equipment safety and major energy flows: few in number but high in accuracy, with redundant wiring and power. Important points serve daily operational optimization, with density set by spatial function. General points are for trend reference only and can use wireless or battery-powered solutions to cut cost. The three tiers need not all be built at once; densifying gradually based on actual data usage after go-live is more economical than full coverage from day one. The table below gives reference configurations for typical office and commercial buildings.
| Point tier | Typical locations | Density recommendation | Accuracy recommendation |
|---|---|---|---|
| Critical | Chilled-water mains in and out, main power feeders, boiler room, data room | One to two per major item of equipment, covering all operating conditions | Thousandth class or metering class |
| Important | AHU supply and return air, fresh-air units, floor mains, elevator machine rooms | One per system per operating-condition zone | 0.5 percent to 1 percent class |
| General | Ordinary offices, corridors, storage rooms | One per roughly 800 square meters, or by functional zone | 1 percent class |
| Reference | Lobby, roof, building envelope | Selected for representativeness and maintainability | Standard class, larger drift tolerated |
The purchase price of the sensor itself is only part of the total lifecycle cost. Every additional point brings cabling and tray, installation labor, system point licensing, annual calibration, data storage, and network traffic. The fully loaded cost of a wired point is often three to five times the device price, and dense cabling also disrupts occupants and damages ceilings. Wireless points eliminate cabling but require accounting for battery-replacement labor and radio coverage reliability. For projects with large point counts, doing data downsampling and aggregation at the edge before uploading valuable data to the platform significantly reduces storage and traffic overhead. The ECG100 edge data manager's local aggregation scheme for high-density point data follows exactly this idea: over a hundred points are first aggregated locally into fifteen-minute feature values, cutting uplink data volume by roughly eighty percent.
To put the economics analysis on paper, use the metric of cost per information: the total lifecycle cost per effective decision supported. Add up device, installation, calibration, and platform fees, then divide by the number of decisions the point supports in a year. When a commercial complex reviewed its points this way, it found that the roof weather station's wind-speed point, though not cheap, supported facade ventilation strategies whose annual avoided losses amortized to just a few cents per decision, making it a high-value point. Meanwhile the densely placed carbon-dioxide reference points in the underground garage were used fewer than ten times a year, pushing the amortized cost per decision extremely high; they were halved the following year and switched to periodic patrol measurement. When numbers do the talking, adding or removing points both gain a basis. Placement of wind-environment points can also reference the economical configuration of ZYH-522 wind speed transmitters at critical roof and duct sections, where one reliable point at a critical section often beats several low-quality points.
The essence of placement economics analysis is managing the sensor budget like an investment portfolio: heavy positions in critical points, light positions in general ones, and constant rebalancing as operational data accumulates. Density follows information value, accuracy follows purpose, and cost follows the full lifecycle rather than the purchase price. After completing the economics analysis, the next step is to make every point worth installing and long-lived. The IP rating of the installation environment directly determines a point's lifespan and maintenance cost; for that, see the previous article IP Protection Ratings Explained: Choosing Dustproof and Waterproof Sensors Without the Pitfalls. For point selection on pump sets and pipe networks, see optimal pressure measurement point placement for ZYH-342 pressure transmitters in central plant pump sets.
