Removing Pain from Your Process Scott Hoover Advanced Solutions, Inc. AB8057 You can have highly capable teams, the latest technology and limitless resources, but if your processes are not optimized, you are leaving money on the table. The AEC industry faces severe challenges, with productivity issues topping the list. A study by the National Research Council (NRC) cites "improved efficiency through more effective interfacing of people, processes, materials, equipment and information" as a key area that is ripe for change and improvement. During this class, learn how industrial engineering methods traditionally applied towards manufacturing environments are being used to improve efficiency, eliminate waste, increase profit and improve competitive advantage for AEC firm workflows. Learning Objectives At the end of this class, you will be able to: Understand the process of performing a workflow simulation to identify and resolve bottlenecks in a firm s current processes Learn to collect data, build models, design alternative processes, validate assumptions and simulate "what if" scenarios Explore how to apply workflow simulation processes Review a workflow simulation case study for a regional, multi-discipline AEC firm About the Speaker Scott Hoover, Ph.D. is an Industrial Engineer for Advanced Solutions, Inc. Scott s primary role is to work with firms to help analyze and optimize design in their current workflows. In this role he develops workflow models and workflow simulations that help clients address issues and achieve business goals. He also assists clients in developing plans to implement workflow change on actual building, road, and product design projects. Scott received his Master s and Doctorate in Industrial Engineering from the University of Louisville. Scott has researched multi-industry processes and procedures with analysis, quantification, and creation using computer simulation models to improve methods and procedures. Email: shoover@advancedsolutions.com www.advancedsolutions.com
What do Industrial Engineers Do? According to the Bureau of Labor Statistics, Industrial Engineers find ways to eliminate wastefulness in production processes. They devise efficient ways to use workers, machines, materials information, and energy to make a product or provide a service. Another way to think of it is, Engineers make things, Industrial Engineers make things better. Techniques of an Industrial Engineer If you were to go to school to study to be an Industrial Engineer, you would quickly learn there are many techniques and tools Industrial Engineer have that they are able to apply in the workspace. Traditionally, Industrial Engineers were used solely in the Manufacturing industry. Recently, people have begun to realize the usefulness of Industrial Engineers in other industries such as Healthcare, Service, and AEC. The following is a list of techniques and tools used by Industrial Engineers: Lean Process of identifying the least wasteful way to provide value to the customer Forecasting Process of making statements about future events. Typically used in production management to determine the most economical number of items to produce for a given time period Kaizen The utilization of all resources (employees) to continually improve all functions of the workplace Operations Research The application of advanced mathematical models to help make better decisions
Six Sigma Set of techniques and tools used for process improvement with the goal of identifying and removing the causes of defects and minimizing variability within processes 7 Wastes (Muda) One of the three types of variation according to the Toyota Production System. The 7 wastes are transportation, inventory, motion, waiting, overprocessing, over-production, and defects Statistical Process Control (SPC) A method of quality control that uses control charts and statistical trends to monitor and control processes Ergonomics & Human Factors The practice of designing systems, processes, or products to fit the need of the people who use or work with the system, process, or product (i.e. fitting a job to a person) Inventory Control The supervision of supply, storage, and accessibility of items in order to ensure an adequate supply without excessive oversupply Work Measurement Application of techniques, which establish the time for an average worker to carry out a specified task at a defined level of performance 5s A workplace organization method that uses a list of 5 Japanese words (Seiri = Sort, Seiton = Straighten, Seiso = Shine, Seiketsu = Standardize, and Shitsuke = Sustain), which describe how to organize a workspace for efficiency and effectiveness Decision Analysis Discipline comprising the philosophy, theory, methodology, and professional practice necessary to address important decisions in a formal matter Value Stream Mapping Method for analyzing the current state and designing a future state for the series of events that take a product or service from its beginning through to the customer Return on Investment (ROI) Performance measure used to evaluate the efficiency of an investment or to compare the efficiency of a number of different investments Continuous Improvement Ongoing effort to improve products, services, or processes Simulation The imitation of the operation of a real-world process or system over time What is Simulation? The problem with the term simulation is that it can mean different things to different people. If you ask a healthcare worker about simulation, they will probably respond with, simulation is the act and/or practice of a lifesaving technique, such as CPR.
A Pilot, military personnel, or professional driver will think of a vehicle simulator if you mention simulation to them. Depending on the type of engineer you talk with, you will probably get different definitions as well. A Mechanical Engineer will probably think of an air flow simulation in the design of a product (i.e. CFD).
If you talk to a Civil Engineer or an Architect, they are going to think of simulation in terms of airflow through a building or room (i.e. CFD). However, if you ask an Industrial Engineer what simulation is, they will tell you it is an act of modeling a process or workflow. What is a Workflow? A workflow is an orchestrated and repeatable pattern of business activity enabled by the systematic organization of resources into process that transform materials, provide services, or processes information. When a workflow is created, it needs to include the tasks, procedural steps, organizations or people involved, required input and output information, and the tools needed for each step in a business process. What is Workflow Simulation? Workflow simulation is the process of creating and analyzing a digital prototype (model) of a physical process (people, materials, energy, etc.) to predict its performance in the real world. Brief History of Simulation Before the creation of computers, people would perform simulations by hand. As you can imagine, this was a long and time consuming process. The first recorded simulation study was performed in 1777 by Georges-Louis Leclerc, Comte de Buffon. In this experiment, Leclerc proposed that he could estimate the value of π by dropping a needle onto a table that was divided with equally spaced parallel lines. In the 1950 s, with the creation of computers, manufacturing facilities began to write computer programs to perform simulations of more complicated systems, like assembly lines. However, these simulations were still very difficult to produce given that they had to be coded from scratch for each simulation.
In the 1970 s and 1980 s, with the advances in computer technology, simulation technology also advanced. This caused other industries to begin to see the benefit of simulation, like service industries. Beginning in the 1990 s through the present, with the continued growth of computer technology, methods of creating simulations changed from hard coding to drag and drop user interfaces. This made the creation of simulation models even easier, allowing one to create a model in a day instead of weeks. Also, instead of a 2D version of the model, models can now be created in 3D. Why Simulate? There are a number of different reasons why one might perform a simulation study on a process. 1. The real system does not exist yet. This scenario normally happens when a company is attempting to implement a new facility, or a new service offering, or producing a new product. Typically in this situation, through the course of a number of planning meetings, the new workflow is developed. At this point, a model could be produced to determine the feasibility of the new process or identify some unknown variables that could only be discovered once the workflow were implemented or the simulation was created. 2. The system exists, but it is expensive, disruptive, or dangerous to change. This scenario is normally seen in large production lines or service industries. An example of this would be a hospital setting. If a hospital is attempting to make a change to a process that is utilized to care for patients, they need to be absolutely sure that the change to the process will not have any adverse effects to the quality of care of the patients. 3. Experimentation may destroy the system. This scenario is not as likely as the other two, but it does exist in cases were a change to the system might result in it not being able to be returned to its original state, should the new change not be favorable. The Process of a Simulation Study At a high level, there are 7 steps to performing a simulation study: 1. Selection of Process to Model 2. Identify Key Performance Indicators (KPI s) 3. Develop Flow Chart of Current Process 4. Collect Data 5. Develop Existing State (As-Is) Simulation Model 6. Alternative Workflow Design (To-Be) 7. Implement & Monitor
1. Selection of Process to Model When selecting a process for a study, you need to have a goal in mind of what you hope to accomplish from this study. Just picking a process for the sake of trying to improve it will not work. You need to have a goal in mind. However, just defining a goal is not as effective as defining a SMART Goal. Specific A specific goal has a much greater chance of being accomplished than a general goal Measurable Keeping track of process gives incentives to keep going Achievable Goals are attainable. Realistic Give yourself the opportunity to success by setting goals you will actually be able to accomplish Timely A time bound goal is intended to establish a sense of urgency and prevent goals from being overtaken by the day-to-day crises that invariably arise Examples: We wish to reduce the time spent on Change Orders by 5% by the end of this year. We wish to reduce the number of Change Orders by 20% by Jan 1, 2016. We would like to respond to 20 more RFP s this year without increasing the team size. 2. Identify Key Performance Indicators (KPI s) Your KPI s can follow the SMART criteria. This means the metric has a Specific purpose for the business, is Measureable to really get a value of the KPI, the defined norms have to be Achievable, the improvement of a KPI has to be Relevant to the success of the organization, and finally it must be Time phased, which means the value or outcomes are shown for a predefined and relevant period. KPI Examples: Resource Utilization Process Time Throughput Wait Time Project Cost 3. Develop Flow Chart of Current Process We have already discussed what a workflow is. At this point in the simulation study process, we now want to take the identified process we are evaluating and build a visualization of the process. One of the greatest benefits of developing a flow chart is it provides awareness of all activities, tasks, resources, and materials involved in the process. This process is also sometimes called Value Stream Mapping. 4. Collect Data The fourth step in a simulation study process is the data collection phase. This step can be a very time consuming and tedious task. This part of the process is different for every simulation
study simply because of the available data for the given workflow. In some instances, firms have very good data collection systems or have workflows already developed of the process being studied. Other times, very little data is available at the very beginning which then requires different techniques to collect the needed data. There are a number of different methods of collecting data: Time Tracking Project Management Systems Workflow Diagrams Interviews Observations Surveys Expert Knowledge Each method has its pros and cons. For instance, time tracking systems can provide a large amount of information very quickly, but the accuracy of some of the data can be brought into question. Observations of the process in motion is the best way to gain a complete understanding of the process. However, it is a time consuming task to perform. The trick to performing a simulation study, and more specifically collecting the data, is to find a happy median of methods to collect data in the most efficient manner. 5. Develop Existing State (As-Is) Simulation Model Once the data has been collected, we are ready to begin building the model. Depending on the process being modeled, this step could take a couple of days or a couple of weeks. Once the model has been created, the next step is to review the model with the client to verify and validate. The difference between the two are: Verification = Did we build the model right? Validation = Did we build the right model? Once the model has been verified and validated, we can analyze the current system. If it is a case where the client does not know where their bottlenecks or problems lie, we will take a look and identify the problem areas with the process. If there is something specific that needs to be looked at, we will focus on those specific process steps. Either way, the goal is to discuss the KPIs that were identified in Step 2: Bottlenecks, Throughputs, etc. 6. Alternative Workflow Design (To-Be) Once the As-Is state has been analyzed, it is time to ask yourself, the What If... questions. This is where all experimentation of the new process is performed without interfering with current system. This part of the process is typically a collaborative process involving an individual from the firm and the simulation expert. Once a change has been decided upon, the simulation workflow is updated and the model is rerun and analyzed. We want to be sure how this change is going to effect the workflow and in turn, how it will impact the business. If the change is not favorable, then you repeat this process.
7. Implement & Monitor Finally, the last step in performing a simulation study is to implement the new change to the workflow and monitor. Once the new the workflow has been implemented, the firm needs to monitor it to be sure it is matching the expectations that were shown through the simulation. After the new workflow has become stable and the desired results have been reach, look to continuously improve this process or choose a new process to improve upon. Demo For the demo, we are going to follow the 7 steps of performing a simulation to provide a real life example of how to perform a simple simulation study. 1. Selection of Process to Model The process we are going to model is a four step packaging line that creates a bag full of different color marbles.
The goal of our simulation study of this packaging line is to try and see if we could determine a way that we can improve our throughput by 10%. 2. Identify Key Performance Indicators (KPI s) For this experiment, it is obvious that one of our KPI s is going to be throughput, since our goal is to increase that metric. In addition, I am interested in looking at the Resource Utilization of our 4 workers and the potential wait time that is incurred at each step of the process. These metrics can show me mathematically if there are any bottlenecks within our process. 3. Develop Flow Chart of Current Process As previously stated, our process if a 4 step packaging line. Step 1 is to open the bag and place 2 red marbles in it. Step 2 is to place 5 black marbles and 7 clear marbles into the bag.
The third step is to place 4 orange marbles into the bag. Finally, the last step is to inspect that bag to ensure the appropriate number of marbles are present, and then close the bag. If the bag fails inspection, the bag is passed back down the line to be repaired. Below is a flowchart of the above described process. Start Open Bag 2 Red Marbles 5 Black Marbles 7 Clear Marbles 4 Orange Marbles Inspect Bag Close Bag Did Bag Pass Inspection? Yes End
4. Collect Data The next step in our process is to collect data that will be used in creating the simulation model. As stated above, there are a number of different types of data we need to collect, as well as different methods to obtain the data. For our experiment, we need to capture processing times for each of the four steps. In addition, we need to capture how frequently we begin a new bag, as well as what percentage of finished product gets rejected for errors. To capture the data for our experiment, we are going to use direct observation. The following is the collected data: Step 1 Processing time = approx. 6 seconds Step 2 Processing time = approx. 10 seconds Step 3 Processing time = approx. 3 seconds Step 4 Processing time = approx. 6 seconds Inter-Arrival Time = approx. 10 seconds % Failed Inspection = 5% 5. Develop Existing State (As-Is) Simulation Model Now that we have an understanding of the process we are trying to model and we have generated a workflow and collected data, we are ready to create the simulation model. To create the model, we are going to use Arena Simulation Software created by Rockwell Automations. Arena, is an easy to use, drag and drop type software where the user creates the process similarly to creating a Visio workflow. For each process step, the user inputs the required data, such as processing time and resources. You can see in the above image that the flowchart looks very similar to the Visio workflow we created in Step 3. Below shows an image of the user interface for putting in processing time and resource use for the first step of the process.
Once the model is created, it needs to be validate and verified to ensure it is an accurate representation of the system we are trying to model and that the model was created correctly. For this example, because I developed the system, I know that it is valid and verified. Now that the model is validated and verified, we can begin to analyze the system, keeping in mind our KPIs. To begin with let s look at our throughput. We are going to let our simulation simulate 2 hours of production. At the end of 2 hours, our total throughput is 672 bags. We need to look at our other KPIs to determine if there are any issues with the system that are preventing us from producing more bags than we currently are. If we take a look at the utilization of the resources in the system, we see that Resource 2, in our second process step, has a utilization of 97.4%. This indicates to me that we have a potential bottleneck at this process step.
To confirm the suspicions of a bottleneck at process Step 2, we can take a look at the wait time for each process step. Looking at the results from Arena, we can see that for the second process step, we have an average wait time of 159 seconds. This confirms our bottleneck and provides us an area that we can look at changing to improve the process. 6. Alternative Workflow Design (To-Be) At this point in the process, we want to ask ourselves different What if...? questions to see if we can develop a way to improve the process, which in turn, will improve the throughput to our desired results. What if we took the second process step and divide it into two separate steps? Let s have the current resource only put 5 black marbles in the bag, and have an additional resource put 7 clear marbles in the bag.
With a couple quick changes to the Arena model, we add the new process step and rerun the model to see what effect this change had to the model. We see that our throughput has increased from 672 to 719, a 7% increase. If we look at the utilization of the resources, we can see that we do not have one resource that is being over utilized compared to another.
However, to be sure that we do not have a bottleneck within our process we need to look at the wait time for each step. We can see from the Arena output that the highest wait time is now with the first step, with an average wait time of 4.6 seconds. While not a major bottleneck, the first process step is still considered a bottleneck. We can now continue the process of answering the What if... questions until we are able to reach our goal of increased throughput by 10%. 7. Implement & Monitor Once you have a solution to improve the process, the next step is to implement that solution. Depending on the solution that is chosen, this process could be quick or could be something that is performed over a period of time. Either way, once it has been implemented, the process needs to be monitored to be sure it is meeting the predicted outcomes of the simulation. But why stop there? One of the greatest advantages simulations can provide is the fact that now the model is created, you can continuously work at improving the process. Whether that is, making additional changes to the process that was modeled already, or if it is adding additional complexity to the model by adding additional process related to this process. For example, if we take out the marble packaging line Let s say, once we have performed a simulation study on the packaging line, we want to now investigate the marble creation process. The advantage that this simulation model has is we can create the new process (marble creation) and tie it into the marble packaging simulation. This would allow us to see, not only, how changes to the marble creation process effect that process, but also how it could affect the packaging line.
Case Study Kohrs Lonnenmann Heil (KLH) Engineers, PSC, is one of the largest engineering firms in Kentucky and Ohio offering mechanical, electrical, plumbing, communication and information technology, lighting design services, commissioning, and energy solutions. KLH strives to improve quality through process innovation and the adoption of technology. To meet this end, they are constantly looking at ways to automate busy work, so their engineers can focus on being engineers. As part of a continuous improvement strategy, KLH partnered with Advanced Solutions to perform a Workflow Simulation Analysis on their Submittal Review process. Over the course of 3 months, Advanced Solutions worked with KLH to perform interviews with key employees in the process, created the workflow simulation model, validated and verified its accuracy, and performed What If analysis to determine if the planned modifications to the process that KLH wanted to make would result in a more efficient process. The following is a high level description of the process steps that were conducted to perform this simulation analysis and the results from the analysis. When Advanced Solutions and KLH began discussing which process they wanted to analyze, KLH decided that they wanted to take a look at their Submittal Review process. This process was already being analyzed by KLH for improvement and KLH believed that the workflow Simulation analysis would help confirm that they were on the right track. KLH Receives Client Submittal Admin Processing Technical Review Processing Admin Post- Review Processing Process Complete KLH s Submittal Review process consists of KLH receiving the submittal and having it go through processing performed by the Admin. Once the Admin has processed the submittal and determined that all of the necessary components are there, the Admin will pass it on to the Engineer for the technical review processing. Once the Engineer is finished, the submittal gets passed back to the Admin for post review processing at which time it is bundled up and sent back to the client, thus completing the process. KLH has an internal goal of completing this process in 5-7 business days, however, in the process s current state, they were missing that goal nearly 20% of the time. With this process improvement, KLH wanted to add a second review process, but still maintain the goal of completing the Submittal Review process in 5-7 business days. To meet this new goal, KLH wanted to modify the process 2 ways: 1. Automate as much of the process as possible 2. Create a computer system that would make the Engineers more aware of the current submittals that needed their review Advanced Solutions had one advantage going into the project KLH already had the process well documented in a Visio flow chart. This allowed Advanced Solutions to review the workflow before ever stepping inside the KLH office. In addition, this allowed Advanced Solutions to
begin the interview process of KLH personnel, during the first on-site visit. Advanced Solutions met with 6 different Software Users, Engineers, Admin, as well as the Senior Technology Consultant. In addition to the interviews, Advanced Solutions spent time observing the Admin to understand the step-by-step, detailed process associated with her job. Once the interviews and observations were complete, Advanced Solutions had all of the necessary information needed to go and create the As-Is model of the process. Once the model was created, it was verified and validated with the Senior Technology Consultant. Below is a figure of the simulation results of the As-Is model. Once it was determined that the model was valid, creation and analysis began on the alternative solution. The model was modified to represent the new process with automation and the second review process performed by the engineers. Below shows the results of the alternative simulation model.
Comparing the results of the As-Is model to the Alternative model, we see that the Alternative model significantly decreases the number of submittals that make it past 7 days for review. In addition to the number of days a submittal is in the process, Advanced Solutions also looked at the utilization of the employees. The following table shows that with the improvements and automation made to the process, it can be predicted that there will be a decrease in the utilization of the Admin by almost 45%. Also, it can be predicted that even though a second review process was added to all of the Engineers for each submittal, there was not an increase in their utilization. Upon completion of the analysis, KLH began to implement the changes to the process. Once they had the new process up and running, after a 2 week time period, they were already see a
significant improvement. Their new average turnaround time for submittals was 2.2 business days, a 25% decrease from what they had previously been doing. KLH PDF Success Story: http://www.advancedsolutions.com/pdf/cs_klh.pdf KLH Success Video: http://youtu.be/bfzjdzpsg_0