The Cen­ter for Dy­nam­ic Data

Stor­age and dis­tri­bu­tion of data via IFTA DataHub

The IFTA DataHub soft­ware col­lects all data on a Win­dows based PC to ar­chive, store and dis­trib­ute it on­line in a net­work for vi­su­al­iza­tion pur­pos­es. The sources of the data is not only the IFTA  Sig­nalMin­er with dy­nam­ic data, but also data from OPC, DataSock­et or other pro­pri­etary in­ter­faces that record the op­er­at­ing con­di­tions of the ma­chine or en­vi­ron­men­tal in­flu­ences. Merg­ing all these data al­lows the dy­nam­ic data to be viewed in a con­text. In ad­di­tion, com­pact over­view files are cre­at­ed by merg­ing data chrono­log­i­cal­ly and log­i­cal­ly. Thus, it is pos­si­ble to pro­vide a daily over­view of the ma­chine be­hav­ior, which re­quires only a few MB of data and is also eas­i­ly re­triev­able by re­mote ac­cess. The func­tions of stor­age, trig­ger­ing, fil­ter­ing and sum­ma­ry can be freely parametrized.

 

Long-Term Data Record­ing

If the IFTA DataHub runs on our data serv­er so­lu­tions, such as the IFTA SlotPC and the IFTA PanelPC this en­ables in­tel­li­gent long-term data record­ing over months and years with­out the need for ad­di­tion­al ex­ter­nal de­vices.

Long-term Anal­y­sis of Dy­nam­ic Mea­sure­ment and Oper­at­ing Data

Find­ing the nee­dle in a haystack

The new DataHub soft­­ware col­lects data from a wide range of sources and en­ables com­bined ana­lys­is. For ex­am­ple, dy­­nam­ic meas­ure­­ment data of more than 100 GB/day from IFTA meas­ure­­ment sys­tems can be ana­­lyzed to­geth­­er with op­er­at­ing data. This data is stored 24/7 in ring buf­fers. The search for rel­ev­ant events in these large amounts of data cor­res­ponds to the lit­er­al search for the "nee­dle in a hay­s­tack".

Here, the DataHub sup­­ports the user by cre­at­ing small, easy to han­dle over­view files that allow a quick check and find­­ing an­om­alies. In ad­di­­tion, the pro­vid­ed on­­line con­nec­­tion al­lows real-time in­­spec­­tion of all data streams and data stream­ing into the cloud.

 

Data Ag­gre­ga­tion for Ef­fi­cient Anal­y­sis of Large Data Sets

Dy­nam­ic Vi­bra­tion Mea­sure­ments Gen­er­ate large amounts of data

Dig­i­tal record­ing of dy­nam­ic vi­bra­tion data gen­er­ates large amounts of data very quick­ly. For ex­am­ple, ac­qui­si­tion of vi­bra­tions up to 5 kHz re­quires a data rate of at least 10 000 sam­ples per sec­ond (S/s). As­sum­ing that each sin­gle sam­ple re­quires 4 bytes of mem­o­ry, this re­sults in a data rate of 40 kByte/s, which are 144 MByte per hour or al­most 3,5 GB per day. Since usu­al­ly more than one sen­sor is used, mon­i­tor­ing a com­plete ma­chine gen­er­ates a much larg­er amount of data than this.

 

 

How does data ag­gre­ga­tion work?

This flood of data brings along a va­ri­ety of chal­lenges. Many of them can be ad­dressed with a tech­nique called data ag­gre­ga­tion. Here, a "data con­den­sate" is pro­duced from raw data in real-time which needs sig­nif­i­cant­ly less stor­age space and yet con­tains all es­sen­tial in­for­ma­tion. This con­den­sate al­lows quick in­sights into large data sets and serves as a guide and in­di­ca­tor dur­ing anal­y­sis.

The IFTA DataHub sup­ports ag­gre­ga­tion over time, which is il­lus­trat­ed in Fig­ure 1. For a given se­ries of sub­se­quent time frames, a high-res­o­lu­tion sig­nal is ag­gre­gat­ed to three data points per time frame: The first point equals the min­i­mum value of the sig­nal with­in a time frame, the sec­ond the max­i­mum, and the third the av­er­age value. Ac­cord­ing­ly, three new low-res­o­lu­tion sig­nals are gen­er­at­ed, which char­ac­ter­ize the ini­tial sig­nal. Their time res­o­lu­tion - and hence their mem­o­ry re­quire­ments - de­pends on the se­lect­ed win­dow width, which is cho­sen by the user.

 

 

Ben­e­fits OF TIME AGGREGATION

Tem­po­ral ag­gre­ga­tion of data al­lows the ef­fi­cient anal­y­sis of large time pe­ri­ods. Fig­ure 2 il­lus­trates this ad­van­tage. The left side of the fig­ure shows spec­tro­grams of 100 MByte of sen­sor data for two ag­gre­ga­tion lev­els in­clud­ing raw data. Where­as with­out ag­gre­ga­tion only a very short time pe­ri­od is cov­ered (bot­tom), the high­est ag­gre­ga­tion level (60 s) pro­vides an over­view of a very wide time range. If an event is de­tect­ed here, it can be zoomed into the data by suc­ces­sive­ly load­ing the next finer ag­gre­ga­tion level (5 s or raw data). The right side of the fig­ure il­lus­trates this for the time do­main ac­cord­ing­ly. In doing so, the data to be pro­cessed and thus the re­quired mem­o­ry and com­put­ing re­sources are min­i­mized, which en­ables ef­fec­tive work­ing even for large data sets.

     

    Ben­e­fits at a glance

    • Fast vi­su­al­iza­tion and anal­y­sis of large time pe­ri­ods
    • Pin­point ac­cu­rate zoom­ing into raw data
    • Ag­gre­gat­ed data is avail­able im­me­di­ate­ly as data ag­gre­ga­tion is per­formed con­ti­nous­ly dur­ing data record­ing

    Overview Func­tions of the IFTA Soft­ware DataHub

    Basic Struc­ture of the IFTA DataHub Soft­ware

    The IFTA DataHub soft­ware con­sists of two com­po­nents: a back­ground ser­vice that han­dles the ac­tu­al tasks and a web-based in­ter­face for users. Thanks to this sep­a­ra­tion high­est re­li­a­bil­i­ty and ro­bust­ness is given.

    • The back­round ser­vice col­lects data from all sources and bun­dles, fil­ters, ag­gre­gates and packs them into files for life vi­su­al­iza­tion and anal­y­sis in the IFTA TrendView­er soft­ware
    • The web-based in­ter­face al­lows users to com­fort­ably con­fig­ure, check sta­tus and in­ter­act via brows­er, on the IFTA sys­tem as well as re­mote­ly from their own lap­top
    • A user man­age­ment can be used to de­ter­mine who can carry out which ac­tions

    Col­lect­ing and Bundling of Mea­sure­ment Data

    In the set­tings of the data sources it is con­fig­ured from where the DataHub should col­lect data. The IFTA Sig­nalMin­er firmware on the IFTA DSP is the main source for mea­sure­ment data - here any dy­nam­ic data are ac­quired.

    In ad­di­tion to dy­nam­ic data typ­i­cal­ly fur­ther mea­sure­ment val­ues are re­quired (op­er­at­ing data, en­vi­ron­men­tal con­di­tions, etc.), which can be ac­quired over the fol­low­ing in­ter­faces:

    Upon re­quest pro­pri­etary in­ter­faces can be con­nect­ed.

    Bundling, Fil­ter­ing and Trig­ger­ing Mea­sure­ment Data Over Time

    The "Ag­gre­ga­tion Set­tings" bun­dle con­fig­ured data sources into new data streams.

    Ad­justable func­tions:

    • Which sources of data streams should be bun­dled
    • Which sources of data streams should be fil­tered, as only rel­e­vant sig­nals should be se­lect­ed in order to re­duce the amount of data
    • Set time ag­gre­ga­tion in order to re­duce the data rate and hence the amount of data
    • Con­fig­ure trig­gers with a pre- and post-trig­ger time span

    Record­ing Mea­sure­ment Data for the Of­fline Anal­y­sis

    In the "ADF Writ­er Set­tings" sav­ing of mea­sure­ment val­ues in files can be con­fig­ured so that they can be used for the off­line anal­y­sis later, for ex­am­ple, in the anal­y­sis soft­ware IFTA TrendView­er. In order to pre­vent the stor­age medi­um from over­flow­ing, the old­est files are delet­ed when the con­fig­ured stor­age ca­pac­i­ty has been reached in order to make room for new cur­rent files.

    Ad­justable func­tions:

    • Def­i­ni­tion of the stor­age lo­ca­tion and the ring buf­fer stor­age lo­ca­tion to be used
    • Def­i­ni­tion of the file name for­mat
    • Record­ing of data streams ac­cord­ing to time slices
    • Defin­tion of the data streams to be saved

    Pro­vid­ing Mea­sure­ment Data for On­line Visu­al­iza­tion

    In the "Set­tings for the On­line Serv­er" one con­fig­ures the stream­ing of the mea­sure­ment data to on­line clients such as the anal­y­sis soft­ware IFTA TrendView­er. For con­tin­u­ous mon­i­tor­ing by users or when per­form­ing spe­cif­ic mea­sure­ments, the user can use this to vi­su­al­ize and an­a­lyze the mea­sure­ment data live, e.g. using TrendView­er.

    Defin­ing Sig­nals by Cal­cu­la­tions

    In the "Cal­cu­la­tions Set­tings" sig­nals de­rived from ex­ist­ing data sources by cal­cu­la­tions can be gen­er­at­ed.

    The main ap­pli­ca­tions for this are:

    • User de­fined trig­ger con­di­tions
    • User de­fined stor­age con­di­tions

    Pro­vid­ing User Input as Sig­­nal

    In the "User Input Set­tings" fur­ther sig­nals can be de­fined as data source di­rect­ly by user input. After def­i­ni­tion of the sig­nal the user can set val­ues di­rect­ly that are then saved along with mea­sure­ment data or the user can con­trol other ac­tions such as record­ing and trig­ger­ing.

    The main ap­­pli­­ca­­tions for this are:

    • Man­u­al con­trol of record­ing
    • Man­u­al ac­ti­va­tion of a trig­ger
    • Man­u­al ac­qui­si­tion of fur­ther in­for­ma­tion that is not avail­able over a sen­sor or other data source e.g. mea­sure­ment point num­ber, en­vi­ron­men­tal con­di­tions, states

    1st Place for the IFTA DataHub

    Pre­sen­ta­tion of the messtec + sen­sor mas­ters award 2020

    The IFTA DataHub soft­ware was award­ed 1st place win­ner of the messtec + sen­sor mas­ters award 2020 on Septem­ber 22, 2020.

    Dr. Jakob Her­mann, Gen­er­al Man­ag­er at IFTA, is par­tic­u­lar­ly pleased about this award: "The DataHub soft­ware is the heart of our vi­bra­tion mea­sure­ment sys­tems. It col­lects dy­nam­ic mea­sure­ment and op­er­at­ing data from var­i­ous sources and pro­vides smart 24/7 long-term stor­age for later anal­y­sis. In 2016, our anal­y­sis and vi­su­al­iza­tion soft­ware TrendView­er was also award­ed first place. To­geth­er, they are a pow­er­ful soft­ware pack­age for the ef­fi­cient root cause anal­y­sis of vi­bra­tions in ma­chines and plants that has been "award­ed" by the read­ers of MD-Au­toma­tion.“

     

    The Next Step: Data Anal­y­sis with IFTA TrendView­er

    Once data has been col­lect­ed and ag­gre­gat­ed via the DataHub soft­ware, anal­y­sis and eval­u­a­tion of the data can be done with the IFTA TrendView­er soft­ware. In this way, for ex­am­ple, er­rors can be found, data can be com­pared, and op­ti­miza­tion de­ci­sions can be made.

     

    The bet­ter you know your ma­chines, the more op­­tim­iz­a­­tions are pos­si­ble.

    Prod­ucts with In­te­grat­ed DataHub Soft­ware

    Ar­gusOMDS

    The proven mon­i­tor­ing & pro­tec­tion sys­tem for sta­tion­ary long-term use.

    Dy­naMaster

    Di­ag­nos­tic tool for high-speed anal­y­sis & in­tel­li­gent vi­su­al­iza­tion

    AIC - Ac­tive Damp­ing

    The damp­ing for your com­bus­tion dy­nam­ics.