Aggregation and calculation at the edge
SimtoLink local data processing does edge computing: aggregation, filtering, calculation, and alarms, without sending every raw sample. Bandwidth and storage both drop.
A rule engine is built in, for formulas, time-series aggregation, a sliding window, difference, and integration. FFT (fast Fourier transform) and a wavelet transform are there for heavier analysis. It runs on an edge gateway or an edge computing box and works with the SimtoLink cloud. Power, cement, and steel sites already use it. Uplink volume falls 70%～90%, and the indicators that matter stay visible in real time.
Modules and what they can do
Builds differ by the kind of calculation and by how much is stored, for edge preprocessing on different industrial jobs. The table is a typical reference:
| Function | Typical capability | Typical use |
|---|---|---|
| Computation | 100+ arithmetic functions, sliding window, difference, integration | Power, power factor, harmonic calculation |
| Alarm | Threshold alarm, rule combination and cascade | Anomaly detection, fault warning |
| Advanced analysis | FFT, wavelet transform | Vibration spectrum, fault feature extraction |
| Storage and cache | 7～90 days on the device, cached automatically while offline | Post-analysis, offline recovery |
The aggregation period can be set from 1 min～1 h. Python or JavaScript scripts can be added. A visual rule screen builds complex logic without writing code. What is switched on depends on the project. The product specification decides.
Typical deployments and what they delivered
From a power distribution board to cement vibration and the health of steel equipment, local processing covers high sample rates where bandwidth and edge intelligence both matter. These are the jobs and the results:
Edge calculation in distribution cabinets
One utility reads current from more than 1,000 distribution cabinets, calculates power, power factor, and harmonics at the edge, and uploads only the results. Bandwidth falls 80%.
Vibration FFT at a cement plant
One cement plant runs FFT on vibration at the edge and uploads only the fault features. Cloud compute drops, and equipment state stays visible in real time.
Rolling-mill health at a steel plant
One steel plant aggregates mill current and temperature at the edge and looks for anomalies. Equipment health is assessed in real time, and fault warnings are accurate more than 95% of the time.
Pump-station compression for a water utility
One water utility aggregates and compresses flow and pressure from many points at the edge, and uploads only statistics and alarms. 4G data cost falls 70%.
Configuration and selection
- Where the math runs: Decide which calculations stay at the edge and which go up. Set the aggregation period so freshness and volume stay in balance.
- Alarms: Fire the rule at the edge first, then sync it to the cloud. Rules can be combined and cascaded.
- What is kept: Important raw data can stay on the device for a long time, for later analysis and for tracing a fault.
- Changes: Prove a rule change in a test environment before it goes live. If the aggregation period changes, judge the effect on the operation.
- Help with the job: SimtoLink supplies configuration templates, practice notes, and rule-tuning advice, including custom Python or JavaScript.
How the job is run, and how it is kept
Move a local-processing job in this order:
SimtoLink can tune the rules so edge calculation and the cloud upload stay in balance, for both bandwidth and freshness.
Support
SimtoLink configures and tunes local-processing rules, including formulas and alarm rules, and can supply configuration templates and practice notes. A product manual and a proposal quote are available.