What is the "right" level of automation?

G

Thread Starter

Goldberg, Jeff

Hello,

My name is Jeff Goldberg, and I am a graduate student at MIT. I have a philosophical question for the list.

How do you determine the "right" level of automation?

In other words, when do you automate a system?
When do you leave it with manual control?
What are the drivers for these decisions?

Thanks for your help,
Jeff Goldberg
[email protected]
[email protected]
508-271-3886
 
Jeff Goldberg asks :

>I have a philosophical question for the list.
>How do you determine the "right" level of automation?
>In other words, when do you automate a system?
>When do you leave it with manual control?
>What are the drivers for these decisions?

Jim Pinto responds :

Automation - and the right level of it - is indeed a philosophical question - akin to the discussion we had on "intelligence".

My brief answers (which I hope will raise some responses) :

1/ Automation was originally intended to eliminate the need for dirty and boring jobs for humans. Automation was intended not only for the elimination of "manual" involvement, but the promotion of human involvement to the thinking and planning level.

2/ Automation quickly became an economical justification : lower cost, higher production rates, better quality, lower waste, etc. So, the justification became - automate wherever possible - and find something better for humans to do.

3/ When automation includes intelligence, the dividing line blurs, and the question becomes philosophical (review the past Automation list discussion - "Where Intelligence begins".

4/ When "synthetic intelligence" improves to a level of "self-organization" - you begin to have a whole new level of discussion, primarily philosophical......

Cheers:
jim
----------/
Jim Pinto
email : [email protected]
web: www.JimPinto.com
San Diego, CA., USA
----------/
 
D

Daniel Boudreault

Hello Jeff,

> I have a philosophical question for the list.

I was thinking about this statement... I don't think upper management use philosophy these days...

> How do you determine the "right" level of automation?
> In other words, when do you automate a system?
> What are the drivers for these decisions?

One of the last things that is brought up in a sales pitch is R.O.I.
If we can show our customer that his return is only 18 months, he would probably jump on it. But if you tell him that it is 50 months, they won't bite. At least that is my experience with the management point of view.

I would automate tasks that are dangerous to make the place safer to work.
I would automate repeative tasks.
I would automate tasks that require precision.

> When do you leave it with manual control?

I always leave manual controls. It helps a lot with debugging and setup.

Dan Boudreault
 
R

Ralph G. McDonald, P.E.

Hi Jeff

Speaking from experience as a consultant and programmer you need to take the following into consideration:

1. Safety: Will the proposed level of automation make the process or equipment more or less safe then the alternative. This includes operator safety as well as envirometal safety.

2. Quality: Will the proposed level of automation make the process or equipment produce a higher quality ( or more consistant ) product then the alternative.

3. Cost vs Savings: Total cost of ownership ( capital, cost of money, maintenance, etc. ) vs Savings ( Quality, energy savings, etc. )

4. Complience with regulations: Will the proposed level of automation allow let the process or facility comply with present and proposed govermental regulations.

5. Cost to train or hire workers to operate and maintain the system.

6. The technical "comfort level" of the management and operating personal.

The most complex or costly solution is not alway the "best solution". I can not give you a "cook book" method of deciding. We set down with the client and go through alternatives, sometimes including a complete study & report phase, sometimes carrying serveral alternatives through the preliminary design phase. Maybe some others will have answer your questions with a more structured approach.

Ralph
 
E
I would suggest automation if I could answer yes to one or more of these questions:

1. Can you remove some danger from the job? Will you increase the safety of the worker or reduce risk to the environment?

2. Can you make a better product, but still make it at a cost that is
competitive?

3. Can you make the same product, but make it faster and/or for less cost?

Ed Mulligan

Speaking for me, not for Starbucks. . .
 
R
> How do you determine the "right" level of automation?
>
> In other words, when do you automate a system?

Usually it is financially driven. Meaning if the net cost of the automation is covered in 3-5 years, automate.

It is also a safety issue. Some activities are rather boring (leading to inattention and possible injury) or inherently dangerous. These activities are best automated if at all possible.

> When do you leave it with manual control?

1. When it is difficult to automate it effectively. many processes do not lend themselves to extensive automation, although some parts may be automatable.
2. When its a short term thing. Its rarely cost effective to automate something that will only be used for a few months, or to make a few parts.
I.e. - don't use extensive automation on pilot plants.

> What are the drivers for these decisions?

1. $$$$$
2. Availablility of operators with the skill level to manually run the system.
3. Whether the automation can be introduced with a minimal amount of production loss.
4. Availability of sensors to monitor the process. Some processes do not lend themselves well to monitoring by instruments. These processes will need operator attention regardless of the level of automation.

In the real world, almost all processes are a mixture of manual, semi-automatic, and automatic. The key is to find the best combination.
 
L
Jeff,

I believe the "right" level of automation, like any capital investment decision, is an economic question, not a philosophical one. Do the analysis. Calculate the ROI. Do a risk assessment. Compare that with alternative investment opportunities.

Sounds easy, but there are some costs and benefits that are difficult to quantify in meaningful terms. And some data will be subject to important confidence limits. That's why experienced automation engineers get paid the big bucks! ;)

There is a large laundry list of potential costs and benefits. But each opportunity will have it's own application specific list. Some things to consider include:

raw material savings
utility savings
labor savings
increased capacity
reduction of off-spec and rework
increased flexibility/ market responsiveness
reduction of lower valued by-products and waste products
avoidance of other capital expenditure
project execution costs (hardware, software, engineering & services)
maintenance costs before/after
sustainability of benefits
training costs
safety risks before/after
environmental risks before/after
project risks

Manufacturers will likely account for other factors like their business strategy, competitive situation, future issues like pending regulatory, labor, etc issues. This is not a complete list, but should give you a starting point.

Regards,

Lou Heavner
course X
 
J

Jeffrey W. Eggenberger

The primary factor in Ford Motor Company is Cost Savings. In order to automate a process, there has to be a "Material" savings pay-back within x
years of x percent.

Note that Ford Motor Company only uses savings of material or energy, and not manpower cost savings.

Jeffrey W. Eggenberger
Electrician: Industrial, Commercial, and Residential
 
Automation is a business decision, first and foremost. Philosophy is for in the bar at the airport. Do you think the motivation for the
cotton gin was to keep people from breaking their backs? You might put an automatic screw driver in a system because of the large number of screws that have to be driven. You might say that a reason for doing this is to keep the operator from having to perform so much repetition. Nevertheless, you won't do it unless it is cost justified. Hey, and what about getting engineering to get rid of some of those screws? If engineering did their job we wouldn't need automation - the product would assemble itself.
 
G

gerald beaudoin

Seems like most of the criteria for going down the automation road have already been thrown out onto the floor......repetitive work, danger and safety considerations, etc. As for manual control.....I alway try to leave the option of performing any step manually, usually with a "Man..Off..Auto" switch. If the automated system
fails, then at least you can limp through on manual. The manual function also allows for easy troubleshooting and maintenance operations such as sanitation without powering up the entire
system.


Gerald Beaudoin
Leahy Orchards Inc.
[email protected]
 
P

Pierre Desrochers

MONEY the only decision...

Of course personal injury costs
Of course repetitive action cause injury
Polution costs
Public outrage too !
Danger ... HA! have you ever seen workers in other places of this world ...
Danger is a joke .

The US is in the most automated industrial countrys because the workers made the capitalist pigs pay... or they closed and just reopened a plant in Mexico... and farther if needed ...

Automation seems the capitalist answer to greed and its casualitys

Don't take me wrong I am one of the club members... and not all industrials are this bad but necessity is the mother of ... who was that ?

It is my experience that the best automation engineer will never be able to compete agains the best RIO report writter ... and this is a fact .

I truly believe that if computers where not around... Zimbabwe would be building cars.

-Pierre
 
A

Anthony Kerstens

The way I've always looked at automation is that you just design how the functions would be done with manual controls (if the operator were a superhuman :), and then apply logic to that framework for automatic functions.

The right level of automation depends on how much button pushing and switch pulling your operators can afford to do.

Anthony Kerstens P.Eng.
 
R

Ralph Mackiewicz

> MONEY the only decision...
>
> Of course personal injury costs
> Of course repetitive action cause injury
> Polution costs
> Public outrage too !

There is truth to the above. There are costs for all these items and money/cost justification should be the basis for all projects, automation or otherwise.

> Danger ... HA! have you ever seen workers in other places of this
> world ... Danger is a joke .

Danger is no joke. The reason that worker safety is not a consideration in those "other" parts of the world is because the tyrants who run those countries and their factories don't place any
value on the lives of their workers, unlike the typical capitalist.

> The US is in the most automated industrial countrys because the
> workers made the capitalist pigs pay... or they closed and just
> reopened a plant in Mexico... and farther if needed ...
>
> Automation seems the capitalist answer to greed and its casualitys

There is little truth to the above junk. What a bunch of BS. Capitalism has done more to improve the lives of the ordinary working person than any socialistic/communistic society (or more accurately nightmarish society) ever did. Just look around you and try to explain why the capitalistic societies are the ones where the workers prosper. Karl Marx was an idiot. Adam Smith and F.A.Hayek rule!

> Don't take me wrong I am one of the club members... and not all
> industrials are this bad but necessity is the mother of ... who was
> that ?
>
> It is my experience that the best automation engineer will never be
> able to compete agains the best RIO report writter ... and this is a
> fact .

In most effective organizations, the people that work out the return on investments work in partnership with the engineers, not competitively. In many cases, the engineer themselves calculate the ROI.

> I truly beleive that if computers where not around... Zimbabwe would
> be building cars.

This is a pipedream. Cars are built in the US, Europe, Japan, Brazil, Thailand, China, etc. etc. etc. because people there buy cars. The
economics of building cars in Zimbabwe and delivering them to buyers in the US, Europe, Japan, Brazil, Thailand, China, etc. etc. etc.
would bankrupt any company stupid enough to try this with or without automation. Let alone having to deal with the murdering socialist thug who runs Zimbabwe.


Regards,
Ralph Mackiewicz
SISCO, Inc.

The opinions given above should be attributed to me and not my company because my company did not write them...I did.
 
B

Boudreaux, M (Mike)

Jeff,

Recently, in a project I am working on, we were faced with the question of whether to automate a set of valves or not. Of course, to not automate them will reduce some up front capital cost, and cause the project cost to be reduced. But, to not automate them will require more manual control by an operator in the field.

The choice was easy. We just evaluated each valve and determined how often the valve would be manipulated based on the number of times per batch and the number of batches per day. If a valve was manipulated "often", then we
automated it. Otherwise, it wasn't. For each dollar you spend, there has to be some sort of return involved. In the case of automation, the return comes back in reduced labor and better quality/operability.

Most of our automation decisions are generated through our HAZOP process, during which we evaluate risks and ways to reduce them through automation.

Regards,

Mike
 
Re: the right level of automation. The perfect system would automate all repetetive and predictable tasks, eliminating needless human errors. The perfect system would gather, compile and present all necessary information, leaving judgement calls to the error prone humans who are infinitely more capable of judgement than computers. RonR
 
hi,

i'm a student from the netherlands and right now i'm researching this topic for an assignement. i think the level of automation is solely based on economic incentives. In the Netherlands labour is very expensive and thus the level of automation in the netherlands is pretty high. An exception to the rule might be dangerous activities.(than again... dangerous work is usually better rewarded and automation then becomes even more cost effictive) I'm sorry to say there's nothing filosophical about this topic
 
Top