Adaptive Data Sampling for Control Valves Using Confidence Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Industrial plant networks face inefficiencies in data collection due to limited bandwidth, where only one system device can be sampled at a time at a low sample rate, often resulting in substantial time spent collecting non-useful data from idle or inactive devices.
Innovation Solution
A device management computer system selects and samples data from operable system devices, analyzing a small initial dataset to determine its usefulness before deciding whether to continue sampling or switch to another device, optimizing data collection by focusing on devices providing valuable data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the monitoring system collects data from multiple system devices in turn at a fixed sample rate, then all devices are monitored systematically, but a substantial period of time is expended collecting non-useful data from idle or inactive devices
Solution Approach 1:
The system performs preliminary assessment by collecting a small initial dataset from each device before committing to extended data collection. This preliminary action allows the system to evaluate data quality and device activity status early, avoiding waste of time on idle devices while maintaining systematic monitoring coverage across all devices.
Solution Approach 2:
Instead of collecting a fixed large number of samples from every device, the system applies partial action by collecting only a small initial subset of data. Based on this partial dataset, the system determines whether to continue collecting more data or move to the next device, thereby reducing time spent on non-useful data while preserving reliable monitoring of active devices.
2Device complexity
If the monitoring system collects a fixed number of data samples from each system device, then data collection is simplified and systematic, but opportunities for collecting useful data are lost when idle devices are sampled
Solution Approach 1:
The system transitions from a static fixed-sample approach to a dynamic adaptive approach. The number of samples collected from each device is no longer fixed but varies based on the device's activity status and data quality. Active devices receive more sampling opportunities while idle devices are quickly moved past, optimizing productivity without significantly increasing system complexity.
Solution Approach 2:
The system incorporates feedback mechanisms by evaluating the initial dataset from each device to determine whether to continue collecting data or switch to another device. This feedback loop allows the system to adapt data collection strategy based on actual data quality and device status, improving useful data collection efficiency while maintaining manageable process complexity through automated decision-making.
3Loss of energy
If the monitoring system samples data from one system device at a time at a low sample rate, then bandwidth limitations are respected, but the efficiency of data gathering is reduced
Solution Approach 1:
The system performs preliminary evaluation by collecting a small initial dataset from each device before committing to extended data collection from that device. This preliminary action allows the system to identify and focus bandwidth resources on active, useful data sources while quickly moving past idle devices, thereby respecting bandwidth limitations while improving overall data gathering efficiency.
Solution Approach 2:
The system dynamically changes the sampling rate and number of samples collected from each device based on device activity status and data quality. Active devices receive higher sampling rates and more samples, while idle devices are sampled minimally or skipped entirely. This parameter adaptation optimizes bandwidth utilization efficiency while significantly improving data gathering productivity.
Data Source
AI summary
A computer-implemented method for monitoring characteristic data includes selecting a first operable system device and receiving a first plurality of data from the first operable system device. The first plurality of data represents at least one characteristic of the first operable system device at a first plurality of points in time. The method also includes determining whether the first plurality of data is useful. If the data is useful, the method also includes receiving a second plurality of data from the first operable system device, the second plurality of data represents at least one characteristic of the first operable system device at a second plurality of points in time, wherein the second plurality of points in time is substantially larger than the first plurality of points in time. If the data is not useful, the method further includes selecting a second operable system device.