liveBook · Manning

Sampling rate

As explained in Digital Twins in Action

Sampling rate is the number of measurements acquired by a sensor per unit of time.

Sampling rate is the number of measurements a sensor acquires per unit of time. It sets the temporal resolution of a digital twin’s observations: frequent readings can capture fast changes, while slow-changing conditions may need only occasional readings. Choose it from the behavior you need to observe, not from the sensor’s highest available setting. section 3.1.2 section 3.2.2

Start with the monitoring objective. Identify the physical parameter that indicates the condition of interest, then estimate its operating range and rate of change. Soil moisture changing over hours can be sampled four times daily; electrical consumption that spikes when an appliance turns on needs more frequent measurements. The question is whether the chosen interval preserves the changes needed for a decision. section 3.2.1

Match the rate to the decision

An hourly power reading from a pool pump is enough to detect whether it runs and calculate daily energy use. Vibration monitoring for bearing defects is different: high-frequency signatures can require 10,000 samples per second, and slower sampling may miss the indicator. These rates serve different goals; neither is a universal default. section 3.2.2

Do not treat sampling rate as accuracy or response latency. Accuracy concerns how close a reading is to the true value; latency concerns how soon a response is needed. A slowly changing irrigation condition can tolerate delay, while leak detection calls for a subminute response. Battery-powered deployments add another constraint: unnecessary sampling and frequent wireless transmissions can shorten battery life and increase maintenance. section 3.1.2 section 3.2.2

Table 3.2 Situations in which higher performance is required across the dimensions of accuracy, sampling rate, and latency, and situations in which more moderate performance will do
Dimension
High performance required
Moderate performance sufficient
Accuracy
Leak detection: ±0.01 L/min. Small flow changes indicate leaks; false positives waste investigation time, and false negatives allow damage.
HVAC control: ±0.5°C. The system can’t control more precisely; ±0.1°C accuracy would waste money on unused precision.
Sampling rate
Vibration monitoring: 10,000 Hz. Bearing defects create high-frequency signatures; slower sampling misses critical failure indicators.
Energy monitoring: 1 sample/hour. Sufficient to detect when equipment runs and to calculate daily usage; faster sampling drains batteries without enabling better decisions.
Latency
Pressure relief: <1 second. A safety-critical response to prevent equipment damage or injury requires immediate action.
Soil moisture: minutes to hours. Changes occur slowly; irrigation decisions tolerate significant delays without consequence.

Compare cases where high sampling performance is necessary with cases where a moderate rate supports the required decision. section 3.2.2

How Digital Twins in Action covers it

How deep it goes
Defines temporal resolution and works through selection by system dynamics, stopping short of a calculation procedure.
Examples
Contrasts four daily soil-moisture readings, hourly pool-pump power readings, and 10,000 Hz vibration monitoring.
What it assumes
Starts from a monitoring objective and physical parameter; introduces temporal resolution without assuming signal-processing theory.
What you can do after
Operate/configure it by choosing a rate that captures relevant changes without imposing unnecessary battery and maintenance costs.
How it connects
Links sampling rate to sensor selection, temporal resolution, rate of change, power consumption, and response latency.

Also explained in

Compare how 2 books explain sampling rate →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Digital Twins in Action. The section links point to the parts of the book it draws on.