Research And Implementation Of Database Storage Engine For Fiber Bragg Grating Demodulation System

The development of FBG demodulation technology is rapidly.It's widely used in the fields of construction,machinery,transportation,aerospace,ect.In the practical application,the sensor network is generally composed of multi-channel and multi-path FBG sensor,and for the convenice of networking and managing data,the open source database MySQL is selected to store data.MySQL achieve data persistence by storage engine module through IO,which support for custom storage engine.In some scenes of multi-channel and multi-path sensor,the data increased dramatically,requiring a large amount of storage space for continuous and high-speed storage.At present,we have to sample the storage data at the sacrifice of high speed performance,which greatly limit the demodulation system applying to the high speed situation.To solve the problem above,the article research the database storage engine for high-speed FBG demodulation system.This paper analysis the internal common interface standard of the storage engine,and study the rapidly caching and real-time compression scheme of the FBG sensor data.The main research work is as follows:(1)The design of Data buffer.The analysis of the FBG sensor data storage and application characteristics shows that in the process of data storage only a small amount of current hot data have to be read rapidly,and the real-time demand for expiration history cold data is not high.So the project Group design the cycle double buffer structure whose function can be converted to each other in a data buffer cycle.(2)The research of FBG sensor data compression scheme.Figure out a large number of FBG reflected wavelength data,and get the data properties and changing rules.Study and compare the principle of commonly used lossless compression algorithm and its application scenarios.Add the Differential operation for FBG sensing data during the compression process to remove the static redundancy.Propose the differential data compression coding scheme which combine RLE and LZW to adaptive output the shortest compression coding.(3)The implementation of storage engine interface.Studies the internal structure of storage engine,then overwrite corresponding interface to realizes the table processing and the function of query,insert,index,search,ect.Design the parallel structure of the data buffer and compression thread,and integrate it into the reading and writting data interface of the storage engine.(4)Performance testing of storage engine.The paper test the storage engine in data writing,real-time data query performance and data compression.After the analysis and comparision,the paper showed that the improved data compression algorithm achieved excellent characteristics in the FBG sensing data.The performance experiment show that custom storage engine can response rapidly to Bulk data to be written in batches.The response time to query fixed-length interval of real-time data increases steadily with the growth of the amount of data.Moreover,storing the static and the dynamic FBG sensor data respectively,with the compression rate 26.19% and 15.4%,which is better than other compression schemes.By the test of FBG sensor vibration experiment,the stability and practicability of the storage engine meet the design requirements under different environment.