Posted by Alex Miggiani on Wed, Mar 03, 2010
The manual way? Good Luck - that is as useful as snow in Dallas!
Speaking of snow in Dallas, I'm still in the Dallas airport after an installation at a leading fresh vegetable packaging company where we solved their manual metal detector data collection issues.
All food industry companies have metal detectors on all of their lines. Most are required by their customers to log all metal detection instances and QA testing. Most do it manually. Our customer had a need to more accurately track their metal detection to offer their customers a higher level of QA checks.
A real-time performance system was installed to track the occurrences, duration, and reasons for ALL metal detector instances. This system also tracks uptime, production, and efficiencies.
With the automated tracking of ALL metal detector occurrences the customer can get a better handle on quality and showcase the data to their customers. The reports track every metal detector occurrence and the associated reason. The reporting shows the chronological distribution of each occurrence and a pareto chart of occurrences vs. reasons. This data is stored in online reports for 2 years.
Now my only problem is getting home from Dallas when there's 6 inches of snow on the ground. Only two de-icing trucks for one of the largest airport in the US. I'm gonna be here for a while...
Posted by Ashok Sivanand on Wed, Feb 24, 2010
I finally picked up a copy of Freakonomics, from the series by Steven Levitt and Stephen J. Dubner. The economist-journalist combo manage to objectively derive using statistical evidence, conclusions to debatable subjects. Examples include the effect of good parenting on education and how sumo wrestlers and teachers both cheat. Their most noted finding is perhaps how they correlate the decline in crime rates in 1992 to the legalization of abortion in 1973.
My biggest take-away from the book would be that interesting (and not otherwise obvious) questions can be asked regarding a subject (and answered) if one has access to a comprehensive data set on that subject.
Sadly, very few production managers or continuous improvement champions in the manufacturing industry today have an accurate, comprehensive data-set with which they can ask similar questions pertaining to their manufacturing floor. This could handicap their continuous improvement efforts when they are invested in the wrong places and not utilized to their full potential. Imagine Superman trying to save Metropolis without X-Ray vision.
Real time performance management systems have given manufacturers Superhero-like insight by accurately highlighting and quantifying opportunities for improving performance on their shop floor. This not only allows them to minimize losses by responding to problems immediately but also provides a reliable, accurate data-set which they can analyze like Levitt and Dubner have: Manu-freakonomics.
Watch this space for stories on how a bakery solved a mystery regarding a seemingly profitable product and how a food packaging company cut changeover times in half.
Posted by Tony Singh on Fri, Jan 29, 2010
One of the greatest benefits of my job is the opportunity to be in front of the customer the very first time they look at the Plantnode reports. It's the expressions on their faces - sometimes shock, sometimes concern - but mostly amazement. Before implementing some type of performance monitoring solution it is not uncommon for folks to think they're running as efficiently as possible, more so, that they're running a tight ship. They may have some downtime periodically but the assumption is they are already at the limits of their production capacity...
...then they see the Machine Status report.
Imagine you have the tools to monitor daily production on a minute to minute basis. You can see how often your assets are running and how often they're not. You are able to respond to issues immediately, as they occur. If you could see how much downtime truly occurs in a full day and how much it greatly affects your productivity, all in real-time, you'd be amazed. Maybe the issue is start up time, maybe it's break time creep, or maybe it's maintenance response time. Whatever the case, a few minutes of lost time over the course of a full day can certainly add up. How much money could you potentially save, or how much more production could you gain in a week, a month, or even a year?
To answer that question I need to share an experience I recently had with a well known pharmaceutical company. A month after Plantnode was installed, I had the opportunity to present the monitoring results for two tablet-packaging machines.
Now - this customer was aware they had inefficiencies in their process, but despite a looming recession they were still running at what they considered full capacity. I distinctly remember their faces when they saw all the idle time. It didn't make sense to them. How could they be running at capacity when they were struggling with at least ten to twelve minutes of downtime an hour? Some hours their machines hadn't run at all - what exactly constitutes "capacity" at that point? Comparing the Machine Status and Production reports our customer could clearly see how all that idle machine time directly affected their output and how overlooking "a little bit of downtime" could quickly turn into huge losses in productivity.
Plantnode provides insight. It gives our customers the ability to really see where their inefficiencies lie and how greatly these inefficiencies impact their business. In this case the customer realized exactly how much time and production they were losing and when exactly it was occurring, down to the very minute. While the initial reaction may have been complete surprise, the results of this business case were even more startling - if the customer could eliminate just 15 minutes of downtime per shift this translated to a benefit of approximately $33K in just one month. And that is just a 3% improvement in efficiency!
If you had the opportunity to gain that much insight into your process, how much would you benefit?
Posted by Ash Sivanand on Thu, Jan 28, 2010
Heard from a buddy last week who works at an automotive supplier with whom I have worked closely. They've finally put an end to the age old epic battle between the operations staff and maintenance staff.
Whenever management tried to minimize downtime losses from machine breakdowns they brick-walled. They did not know where their best opportunities for downtime improvement could be found - maintenance requests, maintenance response, actual maintenance time or operator response time?
Of course asking either group led to finger pointing and situation-exacerbation.
Then they implemented maintenance barcode scanning on their Shoplogix system to track various losses during a machine breakdown... response times were down drastically by the NEXT DAY! (Interesting, the effect of a little accountability). With nobody hiding behind the data (or lack thereof), the operators and maintenance staff now get along better, and management know where to dedicate their improvement efforts. Huzzah!
Looking forward to a wedding at Jay Peak this weekend.. Also have some unfinished business with a glade-trail by the name of Timbuktu..
Sadly, not all epic battles are resolved by barcode scanners...
Posted by Tracy Robertson on Fri, Jan 08, 2010
Mistake proofing, or poke yoke, is commonly known as a technique for ensuring an assembly process is done in the correct manner however there are other "mistakes" that happen on the shop floor with equally serious consequences.
Over production, below-rate run speed, prolonged setup times, material-handling delays, etc. are mistakes that cost a manufacturing organization time and money.
Engaging your employees through the provision of real time machine truth data drives change in the correct direction and shines the light on the real issues impacting performance and profitability, e.g. Downtime, Idle, Setup.
TAKE THE BLINKERS OFF.
Posted by Tracy Robertson on Tue, Oct 20, 2009
Variance analysis has been a staple of the accounting world for decades. Actual to budget analysis is how all organizations get to a bottom line assessment of performance. Usually this occurs at month end, after the books are closed. Actual to Standard Costs are analyzed to understand how well manufacturing performed, often at the end of a job, the end of a shift or sometimes not until the end of a month. Armed with variances, organizations typically drilled into what went wrong and put processes into place to make sure the same mistakes did not reoccur. There are inherant problems with this approach.
First is the lack of detail captured related to the causes of the variance. Second is the timeliness of the information. In a manufacturing plant after the fact data is good for visibility into how to improve a process but does nothing to help increase productivity and profitability while the job is still on the shop floor.
Getting to the root cause
Manufacturing variance analysis has dealt with data from a costing viewpoint inside of the framework of an ERP, Enterprise Resource Planning, application. Inside of any ERP application the focus is not on determining the root cause for a failure on the shop floor. It is on recording the correct distribution of effort and costs to provide accurate financial information. ERP systems do not record why something went wrong, just that more labor, more machinery time or more materials were used. These overages are compared to standard and variances calculated for booking into the General Ledger.
At no time are the real problems examined. There is no way in ERP to record speed loss, when a machine is slowing down but not stopping. ERP can't differentiate between a stoppage from a tool break versus a stoppage from a tool not being part of the set up package delivered to the work station. This inability to capture the root cause of down time limits the real impact of traditional manufacturing variance analysis. If variance analysis does not make the root cause visible how will manufacturers progress?
Real-time visibility
The second issue of timeliness is just as critical. Manufacturers have one chance to get things right, when the work is in process. What operators, supervisors and plant management needs is visibility when work is starting to trend towards a variance while it is on a machine. This is the only time that the human capital in the plant can be effectively deployed to preserve productivity and profitability of work.
As organizations work to trim costs, reduce the cost of quotations, they need to rely on the fact that while the job is running, or when the plant is an hour into a shift, the managers can have visibility into the current trends. Are we running to spec or quoted time for a particular job? Where do we need to refocus our efforts? In today's economy, most organizations, if given the right level of visibility, will make the correct decisions.
Variance analysis has proven a useful tool for accountants to analyze past performance but in order to be an effective manufacturing weapon it must be deployed as a real time barometer of potential problems on the shop floor.
Posted by Tracy Robertson on Fri, Oct 16, 2009
Continuous Improvement Series - Part 5 of 5
As organizations evolve through their implementation of Continuous Improvement (CI) Programs they work to create a culture of change. The approach to change is really geared to elimination of waste in the facility and process. How to do things better than we are today is the main focus of a change oriented culture.
In many CI programs initial changes involve physical adjustments to machines or changes to the layout of the plant. These are significant changes that consume time and resources. As processes mature and benefits are realized organizations start to look at operations to uncover the next level of waste.
Building on the tools described in this series of blogs gives organizations a step up in their search for better management tools. First, those tools provide an effective way to monitor the success of initial CI projects on an ongoing basis. An Accurate Data foundation combined with Alerts will ensure that plant performance does not regress. Changes brought about through CI initiatives can be complimented by monitoring systems designed to prevent a recurrence of waste.
The key is setting the next set of targets. Once large changes are completed it is then time to examine and prioritize the small changes. Once again the foundation of Accurate Data yields the insights necessary to set and prioritize targets.
Real time machine data will yield detailed information on stoppages that occur during the course of the day. Associating these stoppages with root causes and providing a downtime analytical framework will point to the next set of problems that are in the way of improved productivity.
Organizations can then select targets and use tools designed to make these visible throughout the organization. LED Board technology can focus attention throughout the plant ensuring the right employees are engaged in resolving the problems. As these problems are resolved Alerts can be created so that the systems automatically monitor the production process and report when tolerances are starting to slip.
The next series of improvements can then be implemented using the same approach. In this way focus is brought to bear on problems that show potential for impact to the organiztion while systems monitor processes to ensure there are no relapses.
It is the combination of incremental change, applied continuously using a foundation of Accurate Data that can yield truly sustainable CI programs.
Posted by Tracy Robertson on Thu, Oct 15, 2009
Continuous Improvement Series - Part 4 of 5
A key determinant of the success of any Continuous Improvement (CI) program is the acceptance and implementation of change by shop floor personnel. Companies face two major problems when creating shop floor change. The first is the buy in of the employees on the floor. The second is creation of systems that help to reinforce the implementation of change on the shop floor. Without creating the reinforcement for behavioral change organizations will not be able to see when change is not occurring.
Provide Visual Cues to Machine Operators
There are several ways to implement systems to reinforce behaviors and engage people . Ideally these sort of systems provide a choice for the employee or operator to make the right decision. Creating this sort of opportunity for input creates an environment where the employee is able to act in a manner consistent with the goals laid out in the CI program.
At the same time it is also important to give visual cues across the organization so that managers can identify where work is going well and where they should pay attention to potential problems.
To achieve both these goals some easily deployed technologies can be used with great effectiveness. The first is to use display boards, local screens or some other feedback loop to let employees and managers know the status of work. Display boards have the advantage of providing status information to all personnel at once. If installed in appropriate locations then both employees and managers can quickly review performance. Screens located at the machine are a good alternative but do lack the broadcast capabilities of display boards.
Whether display board or screen based, how data is presented can determine whether or not employees are properly engaged. Simple tools that can be interpreted with a glance are ideally suited to informing operators and managers of status. The most effective tools are a combination of color and key metrics.
Run rate of a machine is an example of a metric that can be read off a machine and displayed on a board or screen. By adding some colors to indicate trend states in the machine operators can be informed as to the performance of the machine.
Traffic lights create a good common paradigm that is understood across many different cultures. Using red, green and yellow coloring of metrics, like production rate can immediately indicate performance to standards and guide personnel to the right decision to maintain performance. Green can indicate when things are running to standard, Yellow can indicate when the operation is trending out of standard and Red to indicate problems.
Combining colors with publicly mounted display board messages creates an environment where all production personnel can immediately tell whether the work they are doing is performing to standards.
The use of these tools changes the focus on the plant floor, creating the opportunity to leverage accurate machine data in real time alerts that an operator can instantly understand so that the correct decisions can be made to ensure efficient and profitable operation.
Posted by Tracy Robertson on Tue, Oct 13, 2009
Continuous Improvement Series - Part 3 of 5
Once a foundation of data has been created, organizations will be able to leverage visibility into production to create an understanding of performance. A key element to measuring and understanding manufacturing performance is separating analysis from actionable activities. Both are important, but each is designed to do something different for today's manufacturer.
Analytical Measures of Performance: Using OEE
A good example of analytical evaluation of performance is Overall Equipment Effectiveness (OEE). OEE is a mathematical formula that can be used to assess the performance of any machine independent of its age or characteristics. Production, Availability and Quality are combined to create a ranking of machine performance. With OEE organizations can adopt an analytical framework that allows them to measure each machine's performance.
OEE on it's own is a great machine level analytical metric. More advanced organizations look to leverage OEE to assess how shifts or the plant performs overall. Some organizations also look to assess individual units of work, like jobs, using OEE to understand how efficient their quoting or engineering activities are.
Downtime analysis, understanding what happened when a machine was idle is also a critical analytical tool. What was the root cause of a stoppage, is that cause repeating? Is this a characteristic of a particular shift or is it something that is occurring across similar machines?
Organizations need flexible tools that allow them to aggregate data and identify trends that indicate problems. This analytical view of the plant floor is invaluable in identifying how to move forward on a Continuous Improvement (CI) program.
Using Alerts to Solve Problems as They Occur
Analytics alone are not enough though. As companies create CI programs they establish shop floor behaviors that can become active cues to focus attention while work is being performed. Creating Alerts when a particular machine or process is out of synch is an ideal way to start to actively manage the shop floor.
Alerts can be used to broadcast problem conditions across easily deployed technology. Blackberry or other paging devices and email form very standard tools that can be used to draw attention in the office or the maintenance team to a problem that needs attention. By creating Alerts in real time companies can start to react to problems while they can still be resolved.
There are two ways to manage performance in any facility once a foundation of accurate data exists. Historical analysis can be used to understand trends and impact of particular problems across machines and shifts. Alerts can be used to bring attention to bear on problems as they occur.
Posted by Tracy Robertson on Thu, Oct 08, 2009
Continuous Improvement Series - Part 2 of 5
The foundation for integration of Continuous Improvement programs into day to day management activities is accurate and unbiased data. Unbiased data is not subject to interpretation but objectively represents the actual activities on the shop floor. Accurate data captures not just what went on but creates an understanding of what problems occurred and when so organizations can understand how they can improve.
You Cannot Fix What You Cannot See
One undeniable truth in manufacturing is that you can't fix what you can't see. With unbiased accurate data organizations can shine a light on the plant floor and create a common understanding of what is truly happening across all employees in the facility. How organizations implement this is driven largely by where there is benefit. Ideally organizations would monitor all machines and all steps in the process at the same time. In reality there may be critical process points where the first pass at collecting accurate data can yield the most benefits.
Start with the Key Bottlenecks
Organizations should be examining how to implement accurate data collection programs to maximize the return to the organization. To start pick the key bottlenecks in the facility and use the knowledge gained from accurate, unbiased performance data at these points to create a road map of where next to go on the floor.
What constitutes accurate data is also important. There are basic production data elements that are necessary such as idle time, production rate, quantity produced quantity scrapped and speed loss where machine run rate is below standard. These basic elements are often assembled into analytical tools such as Overall Equipment Effectiveness, OEE. To provide a decision making foundation companies need to look to more detail than these high level measures of productivity.
Accurate and unbiased data means tracking the source of idle time, identifying problems and the scope of problems that impair production efficiency. This analysis provides the road map for an organization, illustrating which problems will yield maximum benefit when they are eliminated. Once those are resolved accurate data will point the way to the next series of benefits the shop floor could realize by improving their processes.
Environmental Factors
It is also important to recognize that many different environmental factors may also be influencing production. Temperature, humidity, vibration could all be critical factors that should be monitored as an indicator of production health.
Employees too can be an influence on problems. Being able to examine data is important, in particular to see shift over shift performance. Getting visibility into the type of problems occurring by shift can point manufacturers to training and environmental problems.
Accurate data does create a foundation to expose production processes and problems and create a road map for any continuous improvement program.