If you've no account register here first time
User Name :
User Email :
Password :

Login Now

Drinking from a Firehose: Data Collection Part II

Recently I attended the IMTS in Chicago also known as the “greatest (manufacturing) show on earth” to paraphrase Barnum and Bailey. And it was a bit of a circus. Instead of rings there were several large halls chock full of the latest manufacturing technology (hardware and software) and every vendor who is anyone was there showing their stuff. Lots of noise (machine and human), lots of people, it was great.

So, what does this have to do with our firehouse analogy?

Let me elaborate. The “hidden” theme of this show was energy and resource consumption. The concern about energy monitoring, display and decision-making was pervasive. Not in the banner over the booth, but in the displays on the floor. Specially for the large control and motor/driver manufacturers like Fanuc, Siemens, and Bosch-Rexroth. They were  all showing technologies for measuring and displaying energy data on controller or dashboards on computers.

Other companies were pushing the application of their machines and solutions to the growing alternate energy market – for example MAG was pushing production of wind and solar components, large and small, and OKUMA had a banner proclaiming “Solutions for Energy.” Everyone is seeing the push to reduce and the potential for market share in creating the solutions.

And why? Demand from customers, growing business opportunities and/or push back from people using their systems in production.

One of the people I have interesting discussions with about the trends of manufacturing and what’s hot and what’s not is a principal in a large high precision manufacturing company in the Midwest. They have a range of clients from medical device to aerospace and the US Navy. To see their facility is to observe parts being made of tiny medical devices on a “Swiss” rotary transfer machine all the way to cowling components for surrounding the jet engines on the Airbus A380 giant airplane.

They also do work for companies like Johnson & Johnson and when I asked my friend if they are getting any serious push from their customers on energy he gave me a resounding YES!

Johnson & Johnson has a statement on their website, amongst a list of their expectations for the company’s environmental performance, that their goal for External Manufacturing (ie my friend’s company) is “100 percent of external manufacturers in conformance with Johnson & Johnson Standards for Responsible External Manufacturing by 2010.” To date J&J states that they have “shared our Standards and/or integrated these standards into formal contracts with more than  80 percent  of our external manufacturers by year-end 2007.” Performance on the environment in the contract with their external manufacturers!

This means data…data on energy consumption of manufacturing…which means data from machines on performance cross linked to parts…meaning energy data linked to steps in the production of the part including on a line by line basis for the program code driving the machine tool in the case of material removal processes. This adds up to a lot of data – the subject of the last posting.

Recall that we had estimated that sampling energy data values for a “medium sized facility” for a day (here meaning 25 CNC machines, 10 programmable logic controlled machines and assorted other handling and line equipment with 8 data sources per machine at a sample rate of 5 hertz) would yield a data stream of 86,400,000 data points each day. And that if we added the other sources, we’d likely end up with 100 million data values a day to deal with.

So, let’s continue our discussion from last time. Data can be related to events and information associated with those events. Thus, data can be understood as something that occurred either at a specific time or over a range of time. In manufacturing systems, events can be a numerical value (for example, the instantaneous power consumption at a specific time) or can be a type of annotation (for example, the alarm state of the machine tool over an interval). Complex events are abstractions of events that are created by combining simple events. For example, based on simple events pertaining to the tool position, the instantaneous power consumption, and the machine tool’s program in machining a part, we can create maps linking power and stages of part production.

The paper I referred to in the last posting describes what is called “events stream processing techniques” that include rules engines (RE) and complex event processing (CEP). These techniques can be used to create higher level abstract events and reason on them by pattern matching and identification. The figure below is an example of software architecture for temporal analysis. This spans multiple data

inputs from several devices, standardized data bus (e.g. MTConnect), and use of rules and complex event processing to create these “maps linking power and production.”

My friend can use this to answer J&J’s concerns about how much energy they are using to create the products they make. And, we can extend this to water, other resources, or whatever the customer wants tracked. And, knowing consumption is the first step to reduction.

The paper from part one of this posting went on to show the results of a case study applied to an energy
monitoring and analysis framework using energy consumption and process parameter profiles from machining experiments.

But at the show, Dr. Vijayaraghavan (the coauthor on the paper we were discussing in the last posting on data handling) and his company System Insights had a neat demo in the Mazak booth showing the real time implementation of this. On a website you can see, for a number of Mazak machine tools of varying sizes, the instantaneous power consumption. If you click on one of the machine icons you go to a “Mazak Energy Dashboard“) for the machine (see below) and get the data, over time periods of

whatever you like for the operation of the machine. You can see total energy use (in kWh), energy cost (for the location you choose – US, Japan, Germany or, in the US, state by state), and “savings” relative to a benchmark machine test in the categories of energy, money (based on cost of energy), Co2 emission equivalent (based on the energy to CO2 conversion for the locality’s energy mix) as well as that equivalent in terms of Al cans saved, miles of auto driving or use of compact fluorescent lamps. And, it has in the lower right hand corner an cool real-time power meter readout.

A further chart from that machine window shows real time power plot over time and summary info, shown below for the Integrex i200S Mazak machine tool. The summary numbers are a bit

different in the two figures as I accessed the data on the website at different times as I was preparing this posting. If we dig deeper, as in the figure last posting September 6th on examples of analysis across temporal scales, we can see the ability to correlate power with specific machine motions. That is next on the dashboard.

This starts to convert our firehouse of data into rather manageable mouthfuls!

I visited the Bosch-Rexroth booth and they were showing similar information albeit, in this case, from a specific set of servos driving a machine simulator.

It’s happening. Data flows will increase. Are you thirsty?

David Dornfeld is the Will C. Hall Family Chair in Engineering in Mechanical Engineering at University of California Berkeley. He leads the Laboratory for Manufacturing and Sustainability (LMAS), and he writes the Green Manufacturing blog.

David Dornfeld
David Dornfeld Director, Laboratory for Manufacturing and Sustainability University of California, Berkeley55
Practical Guide to Transforming Energy Data into Better Buildings
Sponsored By: Lucid

10 Tactics of Successful Energy Managers
Sponsored By: EnergyCap, Inc.

Packaging LED & Advanced Rooftop Unit Control (ARC) Retrofits for Maximum Performance
Sponsored By: Transformative Wave

Leveraging EHS Software in Support of Culture Changes
Sponsored By: VelocityEHS


2 thoughts on “Drinking from a Firehose: Data Collection Part II

  1. Interesting to note that all this data reporting, collecting, and analyzing means more infrastructure and data centers. Not that this is bad as the potential for future savings only come from understanding where one is today. However someone needs to account for the energy, carbon, water, natural resources associated with all of the new data centers required to handle this information.

Leave a Comment