Skip to content

Home

Jump to bottom
Krishna Prabu D edited this page Sep 29, 2017 · 11 revisions

Umbrella SI Journey

IoT Analysis: Perform analysis with IoT elements

Short Name

Equipment Failure Prediction using IoT Sensor data

Short Description

Method for Predicting failures in Equipment using Sensor data. Sensors mounted on devices like IoT devices, Automated manufacturing like Robot arms, Process monitoring and Control equipment etc., collect and transmit data on a continuous basis which is Time stamped.

Offering Type

Emerging Tech

Introduction

IoT Equipment failure prediction involves collecting Sensor values and running Predictive algorithms to predict any impending failure. Core building blocks would include identifying the features or factors contributing to equipment failures. Then configuring a Predictive model to Train the model followed by Scoring the Test data for checking the reliability of the Predictive model. Python 2.0 software is used in this Journey with sample Sensor data loaded into the Data Science experience cloud.

Author

By Krishna Prabu D

Code

Video

Overview

This developer journey is intended for anyone who wants to experiment, learn, enhance and implement a new method for Predicting Equipment failure using IoT Sensor data. Sensors mounted on devices like IoT devices, Automated manufacturing like Robot arms, Process monitoring and Control equipment etc., collect and transmit data on a continuous basis which is Time stamped. The first step would be to identify if there is any substantial shift in the performance of the system using Time series data generated by a single IoT sensor. For a detailed flow on this topic, you can refer to the Change Point detection Developer Journey [https://developer.ibm.com/code/journey/detect-change-points-in-iot-sensor-data/]. Once, a Change point is detected in one key operating parameter of the IoT equipment, then it makes sense to follow it up with a Test to predict if this recent shift will result in a failure of an equipment. This journey is an end to end walk through of a Prediction methodology that utilizes multivariate IoT data to predict any failure of an equipment. Bivariate prediction algorithm – Logistic Regression is used to implement this Prediction. Predictive packages in Python 2.0 software is used in this Journey with sample Sensor data loaded into the Data Science experience cloud. All the intermediary steps are modularized and all code open sourced to enable developers to use / modify the modules / sub-modules as they see fit for their specific application When you have completed this journey, you will understand how to

  1. Read IoT Sensor data stored in the Data base
  2. Configure the features and target variables for Prediction model
  3. Split the multivariate data into Train and Test datasets by configuring the ratio
  4. Train the model using Logistic Regression and measure the Prediction accuracy
  5. Score the Test data and measure Prediction accuracy
  6. Evaluate the Model’s Predictive performance further by computing a Confusion matrix
  7. Rerun experiments by changing the configuration parameters

Flow

  1. User signs up for IBM Data Science experience
  2. User loads the sample IoT sensor Time series data to database
  3. A configuration file holds all the key parameters for running the IoT Time series prediction algorithm
  4. The prediction algorithm written in Python 2.0 Jupyter notebook uses the Configuration parameters and Sensor data from DB
  5. The outputs of the prediction algorithm is saved in Object storage for consumption

Included components

  • IBM Data Science Experience: Analyze data using RStudio, Jupyter, and Python in a configured, collaborative environment that includes IBM value-adds, such as managed Spark.
  • DB2 Warehouse on cloud: IBM Db2 Warehouse on Cloud is a fully-managed, enterprise-class, cloud data warehouse service. Powered by IBM BLU Acceleration.
  • Bluemix Object Storage: A Bluemix service that provides an unstructured cloud data store to build and deliver cost effective apps and services with high reliability and fast speed to market.

Featured technologies

  • Python 2.0: Python is a free software environment that has many applications in the field of Web and Internet development, Scientific and Numeric computing, Desktop GUIs, Software Development, Business Applications. It compiles and runs on a wide variety of UNIX platforms, Windows and MacOS.
  • Jupyter Notebooks: An open-source web application that allows you to create and share documents that contain live code, equations, visualizations and explanatory text.
  • Data Science: Systems and scientific methods to analyze structured and unstructured data in order to extract knowledge and insights.
  • Analytics: Analytics delivers the value of data for the enterprise.

Blog

Change point detection is used to detect any behavioural change in the performance of Time Series data. Specifically, in IoT Sensor data the applications are very wide. Traiditional Change point detection that are implemented use Rule based methods that compare 2 data points or sets of 2 time series to compare and detect if there is a significant change that had taken place. This Journey uses Statistical approach to detect such change points. This journey leverages Node-RED in IBM Blue mix and R Spark services in IBM Data science experience at its core to implement.

All components are designed to be reused either as a complete flow or as individual components. With that purpose in mind, all components are made completely configurable so that multiple experiments can be repeated by tweaking the parameters. The outputs which are statistical metrics can be used further in downstream applications. The entire Logical Architecture or flow of the Journey can be split into 2 main modules. The first module that uses IBM Blue mix collects data from a IoT Sensor source and injects into a DB2 database in cloud. The second module leverages R statistical functions written in R Spark – Jupyter Notebook in IBM Data Science Experience to read this data from DB2 and then compute the statistics to detect if any change point had occurred.

Links

  • Change Point detection: In statistical analysis, change detection or change point detection tries to identify times when the probability distribution of a stochastic process or time series changes. In general the problem concerns both detecting whether or not a change has occurred, or whether several changes might have occurred, and identifying the times of any such changes.

  • Time Series: A time series is a series of data points indexed (or listed or graphed) in time order. Most commonly, a time series is a sequence taken at successive equally spaced points in time. Thus it is a sequence of discrete-time data. Examples of time series are heights of ocean tides, counts of sunspots, and the daily closing value of the Dow Jones Industrial Average.

  • Introduction to Statistical Change point: Changepoint analysis for time series is an increasingly important aspect of statistics. Simply put, a changepoint is an instance in time where the statistical properties before and after this time point differ. With potential changes naturally occurring in data and many statistical methods assuming a "no change" setup, changepoint analysis is important in both applied and theoretical statistics.

  • A Survey of Methods for Time Series Change Point Detection: Change points are abrupt variations in time series data. Such abrupt changes may represent transitions that occur between states. Detection of change points is useful in modelling and prediction of time series and is found in application areas such as medical condition monitoring, climate change detection, speech and image analysis, and human activity analysis. This survey article enumerates, categorizes, and compares many of the methods that have been proposed to detect change points in time series.

Clone this wiki locally