Evaluation of bottlenecks and process flow

From LabAutopedia

<< Back to Project Management
The ultimate goal of implementing laboratory automation is to improve the productivity of the laboratory and the enterprise as a whole. To achieve the best possible productivity gain, it is important that automation be focused on the parts of the enterprise where it can have the most impact. Time, resources and money are often wasted implementing automation applications that have little strategic impact on the organization, while other opportunities for real productivity gain go unaddressed^[1]. The process flow of an organization must be studied and understood to identify the points where automation implementation could have a measurable impact. This requires an enterprise-level view of the organization and the use of process evaluation techniques.

1 Approaches and theories
2 Resources
3 Tools
4 References

Approaches and theories

Detailed Articles: Lean Sigma in the Lab, Improving your lab with simulation

There are many approaches and theories about process analysis and optimization. A very popular process optimization approach, especially in production environments is the Lean Manufacturing approach, which is itself an adaptation of the Toyota Production System (TPS). The "Lean" philosophy considers the expenditure of resources for any means other than the creation of value for the presumed customer to be wasteful, and thus a target for elimination. Some have adapted this approach to create the concept of The Lean Laboratory. Most laboratory examples have focused on testing laboratories^[2]^[3], which tend to have more in common with manufacturing environments. Some applications to the R&D laboratory arena has been reported, although the fit of the Lean approach in the R&D environment is a matter of debate.^[4]^[5]^[6]. Another popular approach is Six Sigma, a business management strategy, originally developed by Motorola to identify and remove the causes of defects and errors in manufacturing and business processes.^[7] It uses a set of quality management methods, including statistical methods, and creates a special infrastructure of people within the organization who are experts in these methods. The elimination of defects and errors in a laboratory process might seem at first unrelated to evaluating the opportunities for automation, but in fact it can be highly related. A given laboratory process may be holding back the enterprise effort if it is prone to errors and rework is often necessary. Laboratory automation can be a tool for minimizing human errors in some cases.

In recent years, Six Sigma has sometimes been combined with Lean Manufacturing to yield a methodology named Lean Six Sigma, or simply Lean Sigma. The debate over when, how or if to employ such methodology is endless.^[8]

Regardless of the approach used, the key points to understand with regard to process flow and laboratory automation are:

The goal is enterprise optimization. Lab automation should be implemented where it can have the best impact on the enterprise. Creating a highly automated laboratory that does not in some way impact the enterprise critcal path is of questionable value. Thus automation opportunities should be evaluated from both an enterprise view and an individual laboratory view. The impact of technology on the balance and flow of the entire enterprise process must be understood, planned and managed.
Automation can enable good science, not create it! Simply applying more technology horsepower will not create good science. If the process in question is not scientifically sound or well understood, it should not be automated. The ability to generate more data, faster cannot replace and must integrate with the process of asking and answering the proper scientific questions. If a process is scientifically sound, but sensitive to human errors, then automation may provide a solution.

Resources

Often organizations lack the data to support a systematic analysis of all their process flow. Annual financial planning exercises usually are inaccurate because the focus is to provide next year's budget forecast and not on examining process flow. In many cases, broad assumptions are made resulting in functional groups that are not appropriately sized, equipped or managed to optimize the bottlenecks. Specialists do exist who are trained in process analysis, and organizations should consider making use of such resources. The field is referred to as Operations Research or Industrial Engineering. While often thought of as being focused on the manufacturing sector, operations research/industrial engineering now focuses on many different fields. The typical Bachelor of Science in Industrial Engineering (BSIE) curriculum includes introductory chemistry, physics, economics, mathematics, statistics, properties of materials, intermediate coursework in mechanical engineering, computer science, and sometimes electrical engineering, and specialized courses focused on methods and types of process analysis. Graduate programs typically offer a diversified program across industries. The usual postgraduate degree earned is the Master of Science in Industrial Engineering/Industrial Engineering & Management/Industrial Engineering & Operations Research.

Tools

Capacity analysis

An organizations ability to consistently meet throughput requirements is evaluated via capacity analysis. This approach identifies workflow bottlenecks and capacity-constrained resources. An enterprise should be as efficient and bottle-neck free as possible prior to spending time and money on laboratory automation. Capacity-modeling tools fall under two broad categories, static and dynamic. Static modeling is the most common and the easiest to program. Static modeling can be very effective as a first-pass analysis tool where future requirements are still wildly variable. The assumptions on a variety of planning metrics have a time-independent perspective. For example, with static modeling, one uses a fixed time (e.g.monthly) demand to calculate the labor and equipment needed to support the required work volume. Dynamic (simulation) modeling, although more complicated to build and use, provides a more realistic tool for planning an on-going process. Dynamic models, by definition, are time dependent. They analyze how systems or areas will react to changes over time. Instead of examining equipment and resources on a fixed time (e.g.monthly) basis, the model simulates work moving through the process on a realistic schedule. This model can then provide a picture of operations during a fixed time, including achievable cycle times, sources of delays, inadequate staffing levels, and shifting bottlenecks.

Simulation models offer many benefits, including the ability to conduct time-sensitive analyses, provide for variability in workload fluctuations, and help with bottleneck identification (including the ability to analyze shifting bottlenecks). If an organization is facing decisions that may benefit from such detailed information, then it should seriously consider using a simulation model. However, if a company is simply conducting a first-pass capacity analysis in a straightforward environment where the implications of inadequate resources are not severe, a static model may be sufficient. Another factor to consider is the nature of the process. Simulation requires that clear rules be in place for process routing, resource usage, processing times, and other parameters. If the process you are modeling has too many open-ended choices (regarding what equipment is used in the process or when certain actions are performed, for example), you should consider using a simpler model. Finally, as with any model, the availability of data is critical to its success. A good simulation requires extensive and accurate data for process flow parameters. Since a main benefit of simulation is its ability to analyze the effects of real-life variability on resource requirements, work output, and other performance aspects, it is important that reliable data be available to determine the averages and variability for each parameter.

Creating a simulation model is a complex and time-consuming process. Management must be patient and allocate time for design because reworking and redesigning the model after it is already in development can add significant time to the development process. The design process should be seen as a cross-functional exercise. All key stakeholders providing inputs to the model will be affected by it. This collaborative approach to development can help ensure that buy-in exists for the model and that the model will accurately reflect the needs of the impacted groups. It is important to keep the model as simple as possible. With input from all key stakeholders, review the process flows and create a simplified version of the process that focuses on the most critical resources and operations. Avoid going to the other extreme of over simplification by getting concurrence of all key stakeholders Use care when deciding what should and shouldn't be modeled. In many cases, a small first-pass static model can be a good starting point for the more complex simulation model.

Once the simulation model is created, it should be put through a rigorous verification process to guarantee its performance and accuracy. One good verification technique is to use historical data to test the model to compare model results with actual performance. Some simulation models are not robust. You must test the model in conditions that exceed the range of normal parameters (for instance, additional or fewer resources and higher or lower work volumes). Since a simulation model considers variability that can randomly sway outcomes, a successful model should not be expected to reflect real-life performance case-by-case. Rather, it should show an accurate trend or average performance.

Over time, the modeled operation will change, and a formal process should be put in place to update the model. A valuable technique to make this manageable is to assign a model owner in charge of tracking all desired changes. These changes are reviewed semiannually or annually in the context of future operational strategy, and a short design process is initiated to clearly define the scope of the updated model.
Additionally, process data should be collected on an ongoing basis in order to maintain model integrity. Systems should be put in place to capture actual processing characteristics (averages and standard deviations) of key model steps. These figures should be periodically reviewed, and the model should be updated accordingly.

Gap analysis

Gap analysis compares an organizations current output with a desired output. While capacity analysis may be employed in gap analysis, the primary focus of the exercise is on identifying the means to bridge the discrepancy or "gap". This goal may or may not be completely aligned with achieving ideal process efficiency. For instance, if a laboratory needs to achieve a certain level of output by a certain date and for a defined period of time, a combination of capacity analysis and critical path analysis is required. Capacity analysis can identify what process components can influence the desired "gap-bridging" and a critical path analysis will examine which of those components can be best be influenced within the desired timeframe.

Takt time

Takt time is defined as the maximum time allowed to produce a product in order to meet demand. It is derived from the German word taktzeit which translates to clock cycle. Product flow is expected to fall within a pace that is less than or equal to the takt time. In a lean manufacturing environment, the pace time is set equal to the takt time^[9].

Takt Time is defined as: T = T_a / T_d

Where:

T_a = Time Available Time to Work (e.g.minutes of work / day)
T_d = Total demand (eg. product units produced / day)
T = TAKT Time (e.g. minutes of work / unit produced)

For example, a business goal may be for a laboratory to process 500K samples every 10 days. If the samples are processed in 384 well microplates, 500K samples = 1300 microplates. If each workday offers 12 hours of available work, 10 days offers 7200 minutes of time to work. Thus the TAKT Time (minutes/plate) = 7200 minutes / 1300 microplates = 5.5 minutes/plate. This is the pace at which the processing of microplates must be completed to meet the business goal. The pace of all the individual tasks in a process must then be compared to the desired Takt time, and the following points considered.

The entire process can’t go faster than the slowest, bottleneck task. (see next bullet point)
The bottleneck task should be running at full capacity, i.e. full efficiency. If this does not meet Takt time pace, then changes must be made. This could be an automation opportunity.
Tasks that run faster than takt time are a potential waste of resources. Consider moving some resources to the bottleneck task. These are not automation opportunities unless it would allow shifting resources to the bottleneck.
Ideal balance is rarely achieved. Do not get obsessed with achieving perfection.
Removing one bottleneck will create another.
Business goal changes will change the desired Takt time.
Maintaining proper Takt balance require constant review and tuning.

Limitations of the Takt time approach

The Takt time approach works well for relatively simple processes, and less well for complex, branching or highly changable processes. In the example above, if the processing of microplates involved duplicating the plates and sending those duplicates through different processes, a simple Takt time calculation is impossible. Similarly, if the processes (e.g. assay) that samples were subjected to changed every 10 days, then Takt optimization would be much less applicable.

References

↑ Finding Greater Efficiency in Modern R&amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;D Laboratories, LaboratoryEquipment.com
↑ Clinical Results; Ortho Clinical Diagnostics
↑ Power of Lean in the Laboratory: A Clinical Application; GE Healthcare
↑ Lean + Six Sigma: Process Improvement Needs to Know Its Place; CIO.com
↑ Weller, H.; Nirschl, D.; Petrillo, E.; Poss, M.; Andres, C.; Cavallaro, C.; Echols, M.; Grant-Young, K.; Houston, J.; Miller, A.; Swann, R. Application of Lean Manufacturing Concepts to Drug Discovery: Rapid Analogue Library Synthesis. J. Comb. Chem. 2006, 8, 664-669.
↑ Lean Six Sigma Implemented at the Lab; Los Alamos National Laboratory
↑ The Inventors of Six Sigma; Motorola University
↑ Lean Won't Work Here Evolving Excellence
↑ Lean Manufacturing and Takt Time strategosinc.com