They say “Experience is the best teacher.” Well, that would certainly be the case for Matthew (Matt) Beier, who is now senior process engineer at CRB. Matt has two-and-a-half decades experience in food processing—everything from corn wet milling to pet food—and with process design and engineering. When his last workplace facility folded a while ago, he decided it was time to hop over the fence and go to work for an engineering/system integration/design/build company where he could use his past experience to help other food manufacturers with process design.
I interviewed Beier, who now heads up the food group at CRB, for an informal, short discussion about approaches to integrating any existing islands of automation in a food/beverage plant, which you can listen to here. If you have 20-40 year-old equipment in your facility that keeps on running, but its controls are aging or ancient, there is hope that you can keep what works and modernize it by pulling worthwhile information from it with the appropriate changes to instrumentation and controls.
A more formal and detailed Q&A follows directly below, where Matt takes a deep dive into integrating old and aging equipment and its islands of automation—more than we could cover in a 25-minute audio interview.
Matt Beier, CRB senior process engineer, has integration experience at several food companies before coming to CRB. Photo courtesy CRB
Wayne Labs, Food Engineering (FE): Matt, tell us about your previous work experiences and your shift to CRB.
Matthew Beier (MB): I graduated from Kansas State University with a BS in Feed Milling and Management.
Over the past 25 years I have worked for a variety of companies in feed milling, pet food manufacturing, soy processing, corn wet milling and sorbitol production facilities, and now I am with CRB. The facilities that I worked in ranged from having virtually no automation, dependent on hard wired board control to extremely large automated systems with several landing in between.
I held a variety of positions increasing in management level for floor supervisor, area process engineer, start-up engineer, construction engineer, capital design engineer, plant engineering manager and now senior process engineer at CRB.
At CRB I work with clients across the food industry. Having worked on the owner’s side for 20+ years gives me a unique perspective as I shift to working for a design-build firm. We see clients in all stages of automation and they often look to us to help them evaluate it for their facilities.
FE: Assuming that most food companies have equipment with local, manual controls or have purchased newer equipment with built in automation with Ethernet connections, what should be some of the early steps taken toward integrating/automating a facility? Would these early steps involve calling in an engineering house? What about goal setting? Discussing budget? Do you have a step-by-step process to be followed to discuss a potential integration project with a client?
MB: The very first thing that a company needs to do is set a very clear goal for its project.
Some common goals we see include:
- Tying some islands of automation together
- Updating specific processing machinery that is worn out to newer models that come with automated functionality
- Taking advantage of technology to reduce work force
Some more complex or economic-driven goals would be:
- Automating processes to allow the use of big data to drive efficiency and profitability
- Automation to tie into predictive maintenance programs
- Automation to allow at-distance monitoring, operation and third-party, at-distance (remote monitoring) services to analyze and optimize the functionality of the process
This process can be very intimidating to companies that do not have the support staff to drive, identify and implement such goals. In many cases a company would be far wiser to get an engineering or technical group involved to help them frame the goal, define it and develop the path forward. Part of any engineering or technical group’s effort would be to help determine a rough cost for the initial goals. Then the cooperative effort would be to determine if that cost falls within the financial framework the company can support. Engineering can and should help the company determine return on investment (ROI), overall equipment efficiency (OEE) and simple payback. The two groups should work together to determine if future growth and new product production will be included in the model.
Each integration project is unique. The steps needed for each company to be successful on the path to automation vary as they must be specific to the current situation at the facility. Now, there are common steps that should be taken in each situation and many of them we’ve already covered, but the model or path is not a one-size-fits-all and must be flexible to fit each situation. When we approach a client’s integration path we take the time to understand the client’s needs, help define goals and budget and then use our industry knowledge and experience to provide the client with the best path forward. The journey is a team effort that must involve both parties with open minds and equal voices but, that is guided by experience.
FE: Of course, it would be nice to have an unlimited budget, but that’s not often reality—and if it were, a food company would probably build a new greenfield facility. So, with a mix of manual, semi-automatic and automated equipment accumulated over the last 30 years or more, what is practical, judicious, cost-effective and expedient in beginning an integration project that could extend over a few years?
MB: Every company has its own capital spending budget that they set annually. Within that capital spend budget there is normally an amount that is corporate driven and also a portion that would be manufacturing plant driven.
Past Experience: Upgrading for “big data” and efficiency
In a past work experience the corporation determined they wanted to upgrade the controls and data collection to allow the company to take advantage of big data to feed accurate information to their Lean Six Sigma (LSS) program to help drive efficiency and profitability. The need for the accurate data was identified by the LSS program, and that was a necessary first step in this process.
The next most important item was to establish a director-level leadership group that would support the initiative. This group identified the facilities and systems that would have the greatest ROI and then the corporation had to rank each project so there was an order to how they were approached since their funds were not unlimited. The corporation determined that they would allocate X amount of capital spend for each of the following 5 years to specifically address these projects. It should be noted that in some cases a plant could have none, one, or many of the top projects in line to be funded. The corporate leadership placed a single person in charge of running the program.
The plant level leadership and engineering staff then had to develop the projects to include cost, implementation plan, hardware and software to be utilized and phasing-in of equipment. The plants were not limited to internal staff. They also had the ability to engage outside expert firms to assist in identifying the best technology and assist in the design and implementation of the projects.
There were many cases where projects didn’t make it in the priority list for the first few years of the program, but they were good ROI projects for the particular plants. In many cases they would be bringing a single island of automation into the controls system or a couple islands. So, the plants would use their own capital budgets to attack these projects earlier and allow the plant to take advantage of the automation and then focus on larger higher dollar integration projects that became more viable with the islands of automation that the plant attached on their own.
Each year the plants had to update the list of integration projects and update the ROI’s. In turn the corporation would adjust the overall project plan. Projects did get shuffled during the entire campaign as energy prices, product demand, outside economics and the like changed and impacted the ROI’s and company direction.
FE: With older manual or semi-automatic equipment that’s in perfect working condition, would it be practical to place sensors in or on the equipment to monitor product quality/throughput/efficiency (for example, a tunnel oven or blender, packaging line, etc.)? What about adding sensors for maintenance (e.g., predictive) purposes? In sum: How can you integrate older process equipment into newer plant-wide systems with HMIs, historians, etc.?
MB: There is plenty of practicality in modifying existing equipment to bring it to a more automated functionality. This is the focus for many companies’ area process engineers. They identify process improvement projects that include updating equipment through the installation of instrumentation to provide critical control information that may have previously been noted on paper by an operator or not at all. One of the first places to start is to contact the equipment manufacturer to determine if the equipment has a formal upgrade package available or if they have helped update controls and functionality of their equipment with other customers.
Past Experience: Add instrumentation to older equipment
I have performed similar tasks in the past that varied in complexity and scope.
I had one project in particular where our goal was to automate a very large centrifuge. There were some basic readings from instruments that came on the unit 15 years prior, but we still had to manually take samples and adjust process flows and valves. Some of the valves were fully manual while others had setpoints changed in the distributive control system (DCS). Normally only one to two adjustments were made per hour using this mostly manual system. The purity of the discharge streams had a large amount of variability and that was lost revenue opportunity.
We worked with the machine manufacturing representative and an instrument company to add technologies to the unit such as online density measurements, machine flushing and product temperature equipment. Some of these instruments required modifications and new installation methods to perform the functionality required.
After six months of installation and testing, the end product was a centrifuge that was fully automated and integrated into the DCS, allowing for multiple adjustments per hour. This improved stream quality, increased yields on higher-return products and very quickly paid off the project. The results of this project justified the upgrade of more machines at that location and then at other plants in the company.
FE: With equipment or machines with built-in PLCs that are not externally connected to a network, is it practical to add in network functionality? What kind of useful data could be derived from this equipment? What are some guidelines connecting previously unconnected machines to a plant network?
MB: It can be practical to tie the remote PLCs into an overarching DCS. The benefits of the DCS give the owner several important options that are value added. One extremely important function that is added is the ability to have a data historian for a system. The historian opens all kinds of avenues for the plant as it brings trending capability with big data to the table.
For example, if a machine was experiencing large variability in pH, the combination of a historian with trending data would allow you to determine that a control valve was causing the issue because it was sized improperly. The applications for product tracking, traceability and systems trouble shooting are nearly endless.
Another benefit a DCS brings to the table is that it allows the plant to use cascade loops to control entire production systems. The loops use the data to produce live trend charts of different portions of a system or the whole process. This allows an operator to predict where the system is going and make adjustments early. During my manufacturing days we would say this was the best situation because the operator is running the process and not the process running the operator. Typically, in this situation operations groups are looking forward at least twelve hours, allowing them to make small changes and causing minimal impact on the system.
Another consideration is that programming remote PLCs is logistically more challenging in the plant atmosphere versus programming and updating a DCS. The DCS providers have advanced to a point where programming a system, tuning an instrument, or building a human machine interface (HMI) faceplate is all Windows-based, efficient and can be done remotely. This allows a company to centralize their controls group to one location or even utilize a third party to handle those tasks. When a plant is located in potentially remote or inhospitable areas of the country it can be challenging to find staff for these types of positions.
FE: If you start out new with a single vendor’s equipment in a line (e.g., Tetra Pak, GEA, etc.), the integration process becomes easier, right? What are the benefits/tradeoffs of this approach?
MB: When working with a single vendor for a production line it often simplifies the installation and coordination between the equipment in the line. Now one of the challenges can be if the vendor can provide you with instruments, programmable logic controls (PLCs) and controls hardware that matches or can communicate with the existing plant equipment and/or DCS. Special attention needs to be taken to ensure that their equipment provided can communicate to the existing systems and to what degree you have control with that existing DCS.
Further some vendors have proprietary programming and software that is completely locked down. If you need to make changes to the system you have to use that vendor’s instrument and controls group even in emergency failure situations. It is also very important to understand what equipment they do not manufacture and are buying out in the process. Down the road, it can get complicated if there are issues with that third party equipment service and possible warranty work if a warranty.
Another consideration would be that in dealing with only one vendor for a whole line of equipment, there is only one group responsible. Your communication channels can be very efficient with only one point of contact. It should be noted that you should always request a single project manager from the vendor since large equipment vendors have multiple manufacturing sites and as an owner you do not want to have to deal with each site for the equipment they make.
One final thought is that your plant or company may not have a large procurement department. Purchasing many different pieces of equipment from multiple vendors, and each with its own terms and communication channels, can easily bog down a small procurement department.
FE: I once visited a yogurt facility where it seemed the only remaining step in an otherwise totally automated plant was to automate removing the cups out of the cartons they came in and placing them in the filler. How do you help clients integrate/automate that one last step?
MB: For the example you described, the approach is not any different than most other design projects. You need to first agree on a charter that defines exactly the goal of the project, in this case, automating container delivery. The charter should also define the budget, any client-required design aspect, project completion date and capacity requirements.
Next the group will need to identify the current control system type and the proper protocols and software that allow for communications and control of the new equipment. Vendors will need to be identified that can provide the equipment needed to perform the task and meet required build specifications.
The design engineers will work with controls experts to identify the proper specifications with the vendors design team to ensure the unit specified can be integrated properly with the current system. Quite often the reason a piece of equipment or task has not been automated is due to the cost or lack of available technology at the time of original installation. Then after a few years technology advances and reduces the cost of the device or a device is invented to perform the task. Things are constantly evolving and improving in automation.
FE: Most food companies will have some sort of ERP/business software in place. But, we haven’t said much about software. What should be in place beyond local machine control(s)? HMI? Historian? Process/batch control, MES? What would be the bare minimums required for integration?
MB: A successful fully-automated facility has an overarching DCS from one of the many capable companies in place. You should always have a historian installed as it gives you several important capabilities. Such capabilities include allowing plant personnel to trend back hours, days, weeks or even years on each specific instrument that is being historicized. Another important aspect is that it allows a more automated system for product traceability and, if needed, identifies product recall needs. Without a historian, a food plant depends on paper records for tracking and tracing products. Paper takes a lot of time and room to store, protect, label and file. If there is a potential recall event, time is of utmost importance, so a historian allows very quick data retrieval.
HMI’s are integral to field monitoring equipment and it is common for most major equipment, like palletizers, check weighers, container fillers, packaging, etc., to come with their own vendor supplied unit. It is important that the vendor unit can communicate via Ethernet to the plant’s overall system and if possible, having the same look and feel of the graphics. With a DCS you typically have centralized control rooms where there is a large bank of control monitors. When cascading systems together, it is important to have both control screens showing the actual process running with control points, but also a screen available to display charts and trends of the process. These trend monitors allow the operations group to analyze and identify changes needed that will keep the process running within specifications hours in advance. The old mantra in plants was that you only cared about how the process ran for your shift, but that’s all changed. With automation and big data trending, plants are leveling out and running in a proactive mode.
This data also lets plants and operators look back at times the plant experience issues, determine the cause and how to better approach it the next time those same circumstances arise again. Learning from these issues will make processors more efficient, and potentially more profitable, in the future.
FE: What key takeaways should processors have after listening to this conversation?
MB: The journey to automate single machines, multiple machines, entire production lines or even the entire plant should not be feared. There are many challenges in the process of making these changes: don’t be hesitant to bring in groups that specialize in this type of project. Often companies have that normal “we can figure it out” attitude that made the business successful, but this is not one of those times when that works. Manufacturers all too often underestimate the money, time and resources needed to perform these tasks. Do not follow that path. There are a multitude of technological advancements that can make a facility cost competitive, efficient, and profitable if implemented properly. Let this world be opened to you but do not start the journey alone. Rely many very qualified resources in industry that can help make it a more seamless transition. Use those experts, knowledge and experiences to your advantage.