Wind Tunnel In-Memory Cache for High-Speed Data Streaming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wind tunnel data collection systems face limitations in sampling rate, data resolution, and latency, leading to incomplete or missed critical data points, which can result in adverse consequences during testing, such as aircraft instability or engine stalling, due to insufficient real-time data access and storage.
Innovation Solution
An instrumented wind tunnel with an instrumentation system featuring at least 10 sensors connected to an In-Memory Networked Cache (IMNC) that communicates at a transmission rate of at least 10 Gb/s, allowing for high-speed data collection and real-time processing, enabling sampling rates of up to 200 kHz and live streaming with virtually instantaneous data access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is collected at typical sampling rates of 10 kHz to 20 kHz, then data collection is manageable with current systems, but significant data points and entire phenomena of interest are compressed or missed
Solution Approach 1:
The patent replaces traditional mechanical/data storage systems with an in-memory data structure store that operates in RAM. This substitution enables sampling rates up to 200 kHz by eliminating the bottlenecks of disk I/O and traditional database operations, thereby capturing complete phenomena without compressing or missing critical data points.
Solution Approach 2:
The patent creates a copy of the data storage system in memory, using an in-memory data structure store that mirrors traditional database functionality but operates at memory speeds. This allows simultaneous high-rate data collection and processing without the limitations of persistent storage systems, enabling 200 kHz sampling while maintaining data integrity.
2Productivity
If Fast Fourier Transform analysis is used for data processing, then single channel processing is achieved, but the desired sampling rate and resolution cannot be provided
Solution Approach 1:
The patent segments the data processing architecture by separating data collection, storage, and analysis functions. The in-memory data structure store handles high-speed data collection and storage independently, while multiple analysis channels can process different aspects of the data simultaneously. This segmentation enables both high processing speed and high measurement precision across multiple channels.
Solution Approach 2:
The patent adds a new dimension to data processing by implementing an in-memory storage layer between data acquisition and analysis. This intermediate layer enables parallel processing of multiple channels at high sampling rates, transforming the traditional single-channel sequential processing model into a multi-channel parallel architecture that maintains both speed and resolution.
3Reliability
If external drives are used for data storage, then data can be recorded, but communication speed is insufficient to determine disruptions during testing in real time
Solution Approach 1:
The patent replaces external drive storage with an in-memory data structure store that resides in RAM. This substitution eliminates the speed limitations of disk I/O and provides real-time data access at memory speeds, enabling immediate detection and response to test disruptions without the communication bottlenecks of external storage devices.
4Speed
If data is docked to RAM and published to a network, then fast access is achieved, but data points are missed during publication
Solution Approach 1:
The in-memory data structure store is self-service in that it manages its own memory allocation, data structures, and persistence without requiring external intervention. This self-managing architecture eliminates the coordination overhead and synchronization issues that cause data loss during network publication, allowing continuous high-speed data access and publication without missing points.
Data Source
AI summary
An instrumented wind tunnel. The wind tunnel receives a test subject which can be a model of an object to be tested with dynamic airflow. The test subject is fitted with at least one sensor to deliver data to a chassis, which in turn uses a network having at least 10 Gb/s capability to deliver the data to a RAM cache. The RAM cache can push the data to multiple viewers for real time monitoring of wind tunnel tests. The operator can then make live adjustments of the test conditions.