using future data to predict your process - future data to predict... · using future data to...

Download Using Future Data to Predict your process - Future Data to Predict... · Using Future Data to Predict…

Post on 07-Sep-2018

212 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

  • 2015 OSIsoft TechCon

    Using Future Data

    to Predict your Process

  • 2015 TechCon Session

    2 | P a g e

    OSIsoft, LLC

    777 Davis St., Suite 250

    San Leandro, CA 94577 USA

    Tel: (01) 510-297-5800

    Web: http://www.osisoft.com

    2015 by OSIsoft, LLC. All rights reserved.

    No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or

    by any means, mechanical, photocopying, recording, or otherwise, without the prior written permission

    of OSIsoft, LLC.

    OSIsoft, the OSIsoft logo and logotype, PI Analytics, PI ProcessBook, PI DataLink, ProcessPoint, PI Asset

    Framework (PI AF), IT Monitor, MCN Health Monitor, PI System, PI ActiveView, PI ACE, PI AlarmView, PI

    BatchView, PI Coresight, PI Data Services, PI Event Frames, PI Manual Logger, PI ProfileView, PI

    WebParts, ProTRAQ, RLINK, RtAnalytics, RtBaseline, RtPortal, RtPM, RtReports and RtWebParts are all

    trademarks of OSIsoft, LLC. All other trademarks or trade names used herein are the property of their

    respective owners.

    U.S. GOVERNMENT RIGHTS

    Use, duplication or disclosure by the U.S. Government is subject to restrictions set forth in the OSIsoft,

    LLC license agreement and as provided in DFARS 227.7202, DFARS 252.227-7013, FAR 12.212, FAR

    52.227, as applicable. OSIsoft, LLC.

    Published: May 8, 2015

  • Using Future Data to Predict your Process

    3 | P a g e

    Contents Objectives ................................................................................................................................................. 4

    Prerequisites ............................................................................................................................................. 4

    Materials ................................................................................................................................................... 4

    Instructions ............................................................................................................................................... 4

    1. Introduction ...................................................................................................................................... 5

    2. Challenge ........................................................................................................................................... 7

    3. Use PI Coresight to visualize raw data .............................................................................................. 9

    4. Integrate future data in PI AF.......................................................................................................... 11

    4.1. Associate future tags with PI AF Attributes ............................................................................ 12

    4.2. Create asset-based analytics to leverage future data in PI AF................................................ 15

    5. Create an element relative PI Coresight display ............................................................................. 25

    6. Create a PI DataLink Report to Schedule Maintenance Teams (Time allowing) ............................ 26

    7. Create a PI ProcessBook Display to show anticipated production from solar panels (Time

    allowing) .................................................................................................................................................. 33

    Annex A: List of PI AF substation parameters ......................................................................................... 42

    Annex B: PI AF Categories ....................................................................................................................... 44

    OSIsoft Virtual Learning Environment ........................................................................................................ 46

  • 2015 TechCon Session

    4 | P a g e

    Objectives

    o Define future data in the PI System o List the ways future data can be injected into the PI System o Describe why PI AF is the best container for metadata o Configure AF templates that reference these future PI points o Create AF elements based on those templates o Create AF analyses that that shows delta between forecast and actual o Create PI Event Frames to capture the metadata around the results o Create a PI Coresight Display to show your data o Create a PI DataLink Report and a PI ProcessBook Display to show your data (optional time

    permitting)

    Prerequisites

    o Some familiarity with PI Data Archive and PI Asset Framework o Some familiarity with the PI Visualization clients (PI Coresight, PI ProcessBook, PI DataLink)

    Materials

    Provided for you:

    o PI Data Archive server with PI Points and data o PI Asset Framework with a base hierarchy o PI Clients installed: PI Coresight 2015, PI ProcessBook 2015, PI DataLink 2014 SP1

    Instructions

    This lab consist of 5 exercises (2 optional). The exercises can be performed independently by using the CaliforniaWeather_Completed

    database for sections 5,6, and 7

    For each exercise, general instructions are provided followed by step-by-step instructions. If you feel comfortable with the general instructions, go ahead and try to do the exercises on

    your own. Dont hesitate to ask questions to the lab instructor if needed.

    If you choose to follow the step-by-step instructions, complete each action marked by a red diamond symbol.

  • Using Future Data to Predict your Process

    5 | P a g e

    1. Introduction In this session, we will describe one way to use future data in the PI System. Lets start with a

    common definition, in this case from Wikipedia:

    Forecasting is the process of making statements about events whose actual outcomes (typically) have

    not yet been observed. A commonplace example might be estimation of some variable of interest at

    some specified future date. Prediction is a similar, but more general term. Both might refer to formal

    statistical methods employing time series, cross-sectional or longitudinal data, or alternatively to less

    formal judgmental methods. Usage can differ between areas of application: for example, in hydrology,

    the terms "forecast" and "forecasting" are sometimes reserved for estimates of values at certain specific

    future times, while the term "prediction" is used for more general estimates, such as the number of times

    floods will occur over a long period.1

    And in the PI System, we define future data as time-series data that isn't measured and collected in

    real-time from machine sensors, and will typically have a timestamp in the future. From its primary

    function as a data historian, the PI Data Archive is designed to preserve all time-series data, which

    includes future data. As a result, future or "non-measured" data will exist in the future but also in

    the past (e.g., important forecasts kept for subsequent analysis, optimization, or auditing).

    Future data is now handled natively by PI System 2015 including storage, analytics, data access,

    reporting, and visualization and is most useful when compared or combined with real-time or

    historical data.

    Although the internals of future data in the PI System and the configuration and

    administration tasks associated to it are not in the scope of this lab, lets take a quick look

    at how future data can be stored into the PI System.

    In PI System 2015, a new concept has been introduced: future PI Points. Those can be

    created by setting the new future attribute to 1 at point creation. This attribute can be

    set using the PI SMT Point Builder or the Excel PI Builder add-in and cannot be modified

    after the point has been created.

    Tags with the attribute set to 0 referred to as historical tags are identical to the tags in all previous

    versions and will reject data with timestamps more than 10 minutes into the future.

    1 http://en.wikipedia.org/wiki/Forecasting

  • 2015 TechCon Session

    6 | P a g e

    The picture below shows a couple ways to create future tags

    Generally speaking, future tags should be used when storing data that is not collected sequentially in

    chronological order. For example, process or operational data should be kept in historical tags because it

    is measured and collected in real time. On the other hand, forecasts or any form of predictive data over

    an arbitrary time range are perfectly suited for future tags. Data for all future tags is stored in

    completely separate archive files. Future data never moves to the historical archive files even after the

    future data ages into the past. The archive files for future tags will be created automatically as data is

    written to the tags. All future archive files will be created in the directory specified during installation,

    and this location can be changed any time afterwards using the new Archives plug-in included with PI

    SMT 2015.

    Writing data to future tags is just like writing data to historical tags just with different timestamps.

    Thus, in general, you should be able to use all your favorite programs and tools for writing data. The

    allowed timestamp range is January 1970 to January 2038. Some common options are the following:

    o Standard OSIsoft interfaces like the PI UFL, PI RDBMS, or PI HTML Interfaces o Configured analyses and calculations with Asset Analytics, starting with PI Ser