2016 ohio craft brewer's conference - brewery automation - john blystone
TRANSCRIPT
2016 Ohio Craft Brewers Conference
Brewery AutomationGetting 80% of the way with 20% the $$$
Presented by
John Blystone, PEElectrical & Controls Supervisor
What we will cover• Why Automate?• The traditional ($$$) way• The new ($) ways• Who / What / When / How• Compare and contrast
2016 Ohio Craft Brewers Conference
Brewery AutomationGetting 80% of the way with 20% the $$$
Presented by
John Blystone, PEElectrical & Controls Supervisor
About Me
John Blystone, PEEducation
• BS Electrical EngineeringThe University of Akron, May 2005
• MBA StudentOhio University, Expected Graduation 2017
Certification• Professional Engineer, State of Ohio (PE 78408)• International Society of Automation Controls Systems Engineer
Experience• 2 years, MT Systems – automating chemical and foundry equipment• 4 years, Louis Perry & Associates – automating rubber and coal power equipment• 3 years, Midlands Millroom Supply – automating more rubber equipment• 2 years, Great Lakes Brewing Company – automating beer equipment
Personal• Like beer, but love scotch – Islay single-malts especially• TOTAL geek – think XL-sized Sheldon Cooper with marginally better social skills
Why Automate?
Speed• Maximize equipment utilization• Maximize employee skill utilization• Keep operator/brewer focused on value-added tasks
Quality / Repeatability• Same thing every time (alarm and stop if not)• Easy to log data in one form or another (future diagnostics)
Perfect your Process• With higher sample size (speed), repeatability,
and data – you WILL learn things• Allow for higher sample rates
(average density/temp/pH over an entire process)
The two extremes
We are going to look at the extremes, then touch on in-betweens
Performance / Capability / Uptime
Co
st /
Co
mp
lexi
ty
The test case
Had to pick something for apples-to-apples comparison
Digital Inputs 48
Digital Outputs 32
Analog Inputs 16
Analog Outputs 8
Temperature Inputs 8
Assuming:
24 VDC Digital InputsRelay Outputs
4-20 mA Analog I/OT/C Temperature inputs (no RTD)
The traditional way
The wonderful, reliable, expensive PLC!
Current brewhouse program using 840 kB of 2 MB (using 1756-L71 processor, SLC I/O)
Scanning all code once every 6.5 milliseconds, or about 153 times per second.
Type # # of Cards Cost per Card Sub Total
Digital Inputs 48 3 (16 per) $382 $1,146.00
Digital Outputs 32 2 (16 per) $682 $1,364.00
Analog Inputs 16 2 (8 per) $1160 $2,320.00
Analog Outputs 8 1 (8 per) $2130 $2,130.00
Temperature Inputs 8 1 (8 per) $2250 $2,250.00
Total $9,210.00
The traditional way
BUT WAIT, THERE’S MORE!!!
That was only I/O cards!
We need a rack, power supply, Ethernet card, and the processor itself!
Description Cost
Previous Page Total $9,210
13-slot Rack $791
Rack Power Supply $1,220
Ethernet Card $2,370
1x Processor $5,790
Total $25,381
The traditional way
Typically programmed with Ladder Logic – unique to industrial automation!
The traditional way
Most common AB HMI option – PV+…at ~$12k!!!
The traditional way
Who programs these things?• Hire an engineering or automation consultant• Typically $100 to $140 an hour (consulting)• All relatively easy for another engineer to pick
up and modify in the future
Why would you go this way?• You expect 3-shift 365 operation (minimal downtime)• You need online editing (no downtime for software changes)• You want or need a high level of complexity• You want or need a large number of I/O devices• You foresee any of the above relatively soon
The new way
Arduino Introduction
Arduino (company) started in Italy in 2005
Goal to make affordable microprocessors
Whole range of processors and add-on cards
Additional I/O available thanks to I2C network
The new way
Intel GalileoArduino compatible
10/100 Ethernet built in
512 kB available for program
SD card offers 32 GB storage
Powered via 5 VDC “wall wort”
The new way
What about I/O?
ERE for I2C I/O!www.ereshop.com
Type # # of Cards Cost per Card Sub Total
Digital Inputs 48 6 (8 per) $24.81 $148.86
Digital Outputs 32 4 (8 per) $38.41 $153.64
Analog Inputs 16 4 (4 per) $37.16 $148.64
Analog Outputs 8 8 (1 per) $45.99 $367.92
Temperature Inputs 8 2 (4 per)* $37.16 + (4*$117) $1,010.32
Total $1,829.38
*NOTE – Temp Inputs using AI and Omega DRSL-TEMP signal conditioners
The new way
BUT WAIT, THERE’S MORE!!!
That was only I/O cards!
We need…NOT MUCH ELSE!!!
Description Cost
Previous Page Total $1,829.38
5 VDC power supply $10.00
1x Processor $79.95
Total $1,919.33
The new way
Programmed with FREE software
Large number of HS and college students familiar with this software RIGHT NOW!
The new way
What about an HMI?
Raspberry Pi with Processing!
Hardware ~$100, free software!
Software environment similar to Arduino
Raspberry Pi with GenLogic!
Hardware ~$100, ~$100 software
Professional looking, easier, still relatively low $
The new way
Who programs these things?• In-house talent with minimal training• Consulting (Eng. Student) ~$15/hr• Not very easy for most “average engineers” to
pick up and modify• Documentation key – check out Fritzing
Why would you go this way?• You don’t need 3-shift 365 operation• You can deal with process interruptions for program changes• You are OK with simple automation, relatively slow processes• You don’t need a large number of I/O devices
The new way(s)
MANY more similar low-cost options!
F. Dangerous Lee from Actual Brewing (Columbus) uses
Embedded Control Concepts
The new way(s)
ECC’s BCS-462 DetailsI/O
8 Temperature Inputs (NTC Thermistor only)8 Digital inputs
18 Digital Outputs (6 PWM capable)
Data LoggingBuilt-in, Web-configured – from 1.1 Hours (@1s) to 2.8 Years (@ 6h)
ProgrammingBuilt-in, free, Web-based!
Finite State Machine8 FSM’s, each with 8 states – also web-configured
COST - $279.95
The new way(s)
Who?
Engineers or Technicians
Typically consultants at first, takes a while to justify cost for full-time
College students or electronics geeks
May find people on staff with more than enough tech savvy to handle
EITHER WAY
You’ll want a qualified person (engineer?) to design and implement Emergency Stops (e-stops), possibly guide panel
design (for future upgrade from Arduino to PLC), etc.
What?
EITHER WAY
You’ll want a qualified person (engineer?) to design and implement Emergency Stops (e-stops)!
BUY QUALITY INSTRUMENTS
Your software is only as good as the information it’s getting. Instruments can last a long time, and will be the major
factor in reliability, precision, repeatability, etc.
SPEND ON DESIGN AND DOCUMENTATION
The low-cost option is meant as a gateway – plan on the future upgrade during implementation! Marginal additional cost, MUCH less headache
to keep running and upgrade later!
When?
NOW!
At least start considering now what this will look like if you ever think you’ll
want it in the future.
Every system, device, tank you put in –just a couple minutes of thought
regarding “how can I automate this” can make a big difference!
How?
How you approach this is critical!
Automation helps keep the skilled operators doing what they love by alleviating the repeat, mundane tasks and is a tool that will help them make more, better beer!
The popular narrative pits machine versus human – focus on how the machine helps us do what we want to be doing and are good at, and alleviate the things we don’t like and find boring
Final words of wisdom
Data and Knowledge lead to understanding, understanding allows for control (which leads to more data and a better understanding)
Never build a robot you can’t catch (AKA Proper e-stops are VERY important)
The internet is a great source of information – and mis-information.
Check your sources. (See right)
Final word of caution
Questions?