Proper automation means that software doesn’t need to be completely replaced when hardware changes every few years. It remains up-to-date, even if production lines are updated, upgraded, or need to be renewed. This means lower costs and less disruption of business operations.
Richard BIESSEN • Head of Product Management
Marcel TEUWENS • Project Manager
Jean-Paul LUBSE • Head of Service
KSE Proven Process Technology
Automating feed production can improve product quality and production capacity, and reduce operating costs. It needs to be prepared thoroughly, and in a structured manner. For a start, you’ll want to retain as much of your existing hardware and software as possible, so the automation needs to be hardware independent. It must work with your ERP (Enterprise Resource Planning) and other software. And different parts of the factory may have their own integrated automation supplier, which need not be a problem as long as the interfaces have been specified and its clear who needs to be contacted.
The system must be easy to upgrade so you can easily add new machines to make it future-proof. Object oriented software is modular, easy to reuse, and easy to modify. You can add new equipment easily without rewriting the complete code, and you’re not tied to a few people with specialized knowledge. This can of course connect to the PLCs that are already part of the factory.
Proper automation means that software doesn’t need to be completely replaced when hardware changes every few years. It remains up-to-date, even if production lines are updated, upgraded, or need to be renewed. This means lower costs and less disruption of business operations. The exact approach will depend on whether you are designing a new plant, upgrading a manually operated plant, or replacing an end-of-life or underperforming automation system. You can start small to gain experience before scaling up to complete factory automation, or do the whole process in one go. It’s not just today’s products that need consideration, but any planned products, too. The system must be versatile enough to handle future requirements.
All automation suppliers approach the process differently. This article covers why it’s critical to proper design. Then the different processes that will need automating (in four major blocks that are often treated together), and the type of information an automation system will produce. Then deal with system simulation and test and finally servicing which, although dealt with it last, it also needs to be considered right from the beginning.
THE DESIGN DOCUMENT
MUST BE ACCURATE – AND COMPLETE
A customer’s first contact of course is normally with an automation suppliers’ (technical) sales people, who can do enough research and analysis for a ballpark estimate of what is needed and how much it will cost. You’ll need much more in-depth research and analysis of hardware, software and procedures, though, and that forms the design document. Any automation supplier quoting for the work needs to know exactly what hardware and software the customer has, and what can be retained after automation.
We start building the plan and project milestones together with the customer. Understandably, you often don’t want to change the way you do things, so negotiate to get the best system. Make sure to reuse hardware and software wherever possible. It’s recommended to use off-the-shelf software as much as possible, and develop any special software needed to tie the system together.
The design document is key to the whole process. This can be prepared by you, a consultant, the automation company itself or a combination of the three. It’s rarely straightforward because you often know what you want but automation companies know what is possible: the design document should mesh the two. The perfect design document would answer all the important questions. It should include all the hard- and software, PLCs and I/Os (sensors and relays) installed at your premises. Thorough preparation is key to prevent discovering half way through automating the plant that a server or I/O or other hardware that should be there isn’t, or should be working correctly and isn’t (such issues are often left over from past maintenance workarounds).
Two major diagrams are central: P&ID (Piping & Identification Diagram) and a Functional Diagram.
THE BASIC DESIGN DOCUMENTS
The P&ID is the basic document for developing the automation. It shows an overview of the plant with all the equipment and instrumentation. Many pieces of equipment, control panels, sensors and relays will be reusable (a typical plant in Europe will have 4000 to 5000 I/Os), but some will not. Some relays may operate at 230V and will need to be replaced by 24V versions. Others may not comply with ATEX and other safety requirements.
The P&ID is the basis of a total retrofit checklist of the hardware, cabling, sensors and other equipment, and for consistency this should use the same codes as the P&ID. Alongside the P&ID diagram you need a functional specification describing what the new plant should do – this is the basis for the software design. There are several ways to put this information on paper but best here, too, is a structured way to help develop the automation software.
To develop the functional diagram, draw up the process step by step. That means learning the factory’s procedures, and that means asking the operators since they’re the ones who know. For intake, for example, you need to know exactly what happens when a truck comes in, when it goes to a weighing bridge and to reception, and how and where the truck is unloaded.
Processes have often evolved over time and are not always on paper. Trying to discover them often leads to interesting discussions. “So, where do you put the pallet from the fork lift truck?” “In one of the bays there.” “How do you know which one?” “We just know – from how we’ve done it before, it always goes there.”
Because they need to be so detailed, these sessions can sometimes last more than a week discussing how people do things – and particularly why people do things. Even small details can be important, like who does the loading/unloading. If truck drivers do it, they will all need to be trained in the new procedures, so it is often better for a single customer operator to do it all.
By using ISA88, the batch process control standard which is a design philosophy for describing equipment and procedures. Use it to split the process into different blocks. First you split it into separate process lines: receiving, pelleting, blending and so on. Then split each line into units: blender, transporter and so on.
ISA88 covers the design of the physical process, machinery and process control. Alongside that is ISA95, which helps in developing an automated interface between enterprise and control systems. It effectively gives you the process mapping: a flowchart of logical process steps, describing all the manual steps in the order that they happen, from when a truck driver comes in. ISA88 and ISA95 give you a design that can also be understood by non-programmers.
Contamination can be an important design consideration. It can appear anywhere in the animal feed and premix production process, and will need to be prevented, particularly for premix plants. From intake to outloading, there are many possibilities for components to mix. It can even be an issue with hand tipping, and it is worth reserving dedicated hand tipping stations for different contamination groups and to reject any extracted dust. Hand tipping station(s) should be placed at the last possible point in the production process so that any sensitive material doesn’t need to be handled and transported through the whole production line. This can be in the mixer, or just before packaging.
Small manufacturers will want at least some help with the functional design, while large manufacturers (particularly multinationals) will do the work themselves. Many are now indeed carrying out the exercise without planning automation. They are rationalising their operations, studying how and why process procedures in different countries and plants differ to eventually streamline and optimize their process.
Once there is a good map of workflow, a story-board of each step is the next thing in line. Add photos and used documents of a truck coming into the factory, for example, then entering reception to make all the steps in the flow visible, so everybody knows what you’re talking about.
Mark on the diagram where the new automation take/send information from/to ERP systems, and this interface with the current software is critical to good operation. Exactly what gets interfaced and how depends totally on how you want to proceed, and how much is already automated. It can be approached in many different ways and at different levels, and this greatly influences the scope and the price of final automation.
All this is the basis for designing the automation system, and is actually all needed before an automation company can make any sort of realistic quote for the new system. It really needs to be done before quotes are invited, otherwise automation companies will be duplicating work that only one can charge for. All will be quoting for different things making it difficult to compare, since you are comparing pears with apples.
The structured hardware and software development approach is much better than the old method of preparing a 200-odd page document that is hardly looked at after a design document has been accepted. The work is really part of the automation project itself. One of the automation companies is sometimes asked to prepare the pre-engineering work, but if it is unpaid then they will be unlikely to agree to sharing the work with other companies tendering.
The next decision is what to automate and when. Even when a complete factory is being automated, the work is often done/commissioned/tested in four stages: intake/transport, dosing/weighing/grinding/blending, pelleting/extrusion, and finally bulk outloading.
AUTOMATING THE FOUR
MAJOR PRODUCTION STAGES
You need a good grip on your raw material stock for proper intake and transport. Accurate records, even in complex transports with multiple destinations, ensure effective silo occupation and prevent unexpected empty dosing silos. You’ll need a clear picture of the status of all shipments.
Operators and drivers need to be guided step by step through the entire intake process, from delivery and weighbridge to handling at the intake pit or blow intake. RFID can minimize the chance of errors, and help give the detailed product traceability that is essential to track down root causes if there is a problem with an order.
The next stage is dosing, weighing grinding and blending. For accurate and consistent dosing and weighing, you need to optimize the dose parameters during and after each dose. Raw materials can have varying flow properties, but this control is down to machine control – the individual machines must all work correctly within the automated system.
Most important in grinding and crushing is to maintain product quality by giving the right material structure. Close behind that, though, is achieving the highest possible capacity. Flow needs to be automatically matched to the speed of the mill, and blending operations should produce a consistent product for the least processing time. Monitoring the mixing times prevents demixing, with start/end detection at a filled mixer. If complete automation is not possible or too expensive, liquids and other ingredients can be added by hand.
Because so many variables influence pelleting and extrusion, finding the optimum conditioning must normally be evaluated case by case. It depends on the formulation, hygiene requirement (such as pasteurisation), process control and cost. Ensuring that the conditions are consistently repeated regardless of who is operating the machine is a challenge.
Conditioning is perhaps the most important factor that influences the pellet quality, and must be evaluated case by case. It is probably the least understood process by operators, plant managers and equipment suppliers. Conditioning improves both the nutritional and the physical quality of the end product, and the efficiency of the process. For example, it increases the pellet durability index (PDI), pasteurizes the mash, reduces anti nutritional factors (ANFs), improves feed conversion ratio (FCR), decreases processing costs, increases production rate and increases gelatinisation of starch.
Conditioning is mostly based on steam (a combination of temperature and moisture), and the key parameters are retention time, moisture content, temperature and pressure. It is virtually impossible to set a standard for conditioning different raw materials, and all of these need to be known and taken into consideration in the process control.
The objective is to provide consistent pellet quality under different conditions and varying ingredients in the formulas. Pellet quality can be determined by the amount of fines in the product, so not generating excessive fines is also important. You also need to consider type of animal feed (chicken, dairy feed, pig feed, etc.), formula or recipe, required pellet hardness and size, type and quality of the raw materials, coarseness of the feed, desired production capacity in tons per hour, whether heat treatment is required or not and the required quality level of the end product.
All in all, it’s a real challenge to control all the variables and get the same results time after time.
Bulk outloading, too, has a wide diversity of executions. You can work with direct bulk outloading, directly with weighing and indirectly with weighing. It may involve transport belts, mobile and stationary bunkers and counter set loading, to name a few.
REAL-TIME REPORTING – WHO SHOULD GET WHAT?
Accurate data enables you to monitor progress, check Key Performance Indicators (KPIs), and take preventive action on a downward trend. But you need more than data alone; you need to be able to select the most relevant information, validate it and present to the people who need to use it. And all this has to be done quickly, so you’ll need a versatile and customizable reporting tool.
The system should generate tailor-made graphical reports and then automatically distribute them to the people who need them. The reports can be customized for specific end users, for example, weekly summaries for management or detailed KPI analyses for specific lines. Without a reporting tool the data has to be manually collected, processed presented and distributed. This is time consuming and potentially costly because any delay in receiving data may lead to a delay in taking appropriate action.
Automation systems need to generate the essential top-level information for CEOs to make their strategic decisions. Information needs to be tailored to avoid large amounts of paper that are never studied. Decide on the overall KPIs that will confirm that that the factory is working at its maximum capacity and quality for lowest running costs, and that orders are being dispatched on time.
Often it is the Finance Manager who is most committed to starting small, ensuring that the processes are automated profitably. Again, you should produce only the essential information. Automation suppliers can work with other suppliers or 3rd party software to automate administration: orders, invoices and accounting. Plant managers need KPI charts or OEE (Overall Equipment Efficiency) measures per machine, operator or product line. They should be able to connect by phone or laptop to any plant in the world. Bottleneck analysis can suggest small changes to the process that will make a large difference to production.
Some of the greatest benefits of automation are for production planning and logistics managers to ensure you’re delivering to customers on time. When a client order comes in, you need to predict when it will be ready. You should be able to track customers and their orders, trucks, and drivers to deliver on time.
The system should show when customer orders are ready, raw material stocks are topped up or when a truck can be loaded. You need to make best use of available silos and integrate logistics planning seamlessly. Schedulers should ensure that transport schedules correspond with production schedules. Proper automation can eliminate the need to be constantly contacting production and logistics departments.
Operators need a complete system overview in simple screens. Batch status should be clearly coded. Software modules need to integrate control, visualization and alarm handling. You’ll always need to allow direct operator intervention if there are problems, though. This leads to higher quality, fewer rejects and less downtime.
Automation gives plant/maintenance engineers a Maintenance Management System (MMS) with predicted risk of breakdowns, with machines and spare parts to be maintained and replaced. Monitoring Access to I/O streams helps problem solving, and you should connect to external maintenance management software. To manually implement MMS is very labour intensive. The system gives quality managers the control information they need, too. Making KPIs available helps create a culture of optimization.
Laws and regulations require information to be recorded from the supply of raw materials to the delivery of finished products. In case of calamities, you need to identify the cause quickly, along with detailed product traceability. You need to know what went into every product, along with every activity and event for every dosing, weighing, milling, grinding and transport machine, with a time stamp. All batches are labelled for ingredients, suppliers, operator, and parameters like temperature, humidity, pressure and energy consumption. If there’s ever a problem, we know the exact suppliers and process conditions involved.
Reporting can also be valuable for business management strategies such as Six Sigma, which are widely used to improve the quality of process outputs by identifying and removing the causes of defects and minimizing variability in manufacturing and business processes.
Information Technology deeply penetrates into all aspects of production. Administration, order intake, production, packaging and delivery to the customer can all be integrated. All data should be entered only once in the automation system.
If a particular piece of equipment is not performing according to expectations or is noticeably less productive than similar machines alongside it, the reason may not always be obvious. A report that cross references detailed information takes the guesswork out of resolving this kind of issue. For example, that machine may have had to work more often with sticky or difficult ingredients than its neighbours, and could even be more productive.
TESTING THE INSTALLED SYSTEM
Of course any installed system needs thorough testing. A simulation mode greatly improves system testing, and lets you solve problems without affecting production.
Factories will have many different software versions, and any problems will show up with the simulation normally. This is very successful for making small final changes, as well as building confidence that there will be minimum downtime at changeover. The simulations can also help later, during training, to give everyone a hands-on feel of the production processes.
Then you should be ready to install the hardware and software. It is of course important to keep production downtime to a minimum during the changeover (and build up stock to ensure delivery).
Once it’s all installed, it’s time to test the hardware and I/Os, including any hardware installed by other suppliers. Then test the software separately from the hardware. When all has been shown to be fault free, a final test with software and hardware together is advised. First make a “dry” or “air” test with no materials going through to ensure everything works to specification, fooling the sensors to think all the materials are there. This also checks if the individual machine control systems are working correctly within the overall automation.
A first run with materials in the line gives a final check that all the individual machine control systems are also properly integrated. When this is successful the system is ready for commissioning. Then comes the handover from the supplier and the start of automated production.
Once the software is working, the first step is to check product quality has been maintained or (normally) improved. Once the quality is right, you increase output to its maximum reliable value.
One point that managers at the customer site need to remember throughout. People in your organization are used to doing their work in a certain way. Experience learned diplomacy when talking with them to help them through the normal resistance to change. You should clarify what will happen, though, assuring operators that they will be trained to work with the new system, or for other work in the factory.
SPECIFYING SERVICE LEVELS IN THE SLA
While most people think of automating to improve quality or boost production, there can also be real benefits in service quality. The highest priority is to keep production running with an absolute minimum of downtime. Automated systems can warn of potential problems in advance so that best use can be made of planned process downtime. Left to themselves, problems normally occur at the least convenient time – during a rush order, on a Saturday or Sunday or at the end of a shift.
The system should register how long the separate machines are running for predictive and preventive maintenance to prevent unnecessary standstill. Small companies are sometimes afraid of the costs of proper maintenance, but this is a mistake – the costs of not doing it are always much higher.
You will need to specify preventive or (preferably) predictive maintenance routines, and which parties and processes will be involved. Key is the SLA (Service Level Agreement) which can cover a few essential pieces of equipment or systems, right up to total factory automation.
The automation company will often take on the servicing, tailoring the level of service to requirements: often from 8-5 or 7-10 during the day, up to full 24/7 cover. The level will also be determined by how many spares are to be held in stock, and how fast they need to be delivered when needed.
Discussions with the IT department are critical for proper servicing. Larger companies will have their own IT environment who will want to drive the whole process, and customers will often have specialist maintenance, IT, software and ERP layer companies. This is no problem as long as everyone knows their responsibilities, where the interfaces are between the different systems, and who to call if there are specific problems.
Many customers prefer to do the servicing themselves, contacting the automation company for remote maintenance support if there are problems. For lowest costs, the first-line approach is normally remote support. Most issues can be solved by making software bridges or workarounds – so production can continue if a sensor (for example) has failed – until it is later replaced during process downtime. Remote access can alert potential system problems before you are aware of them. This enables key data to be delivered to 3rd parties like motor manufacturers, informing them of cycle times, running hours and operating parameters for predictive maintenance purposes.