Plant and shop floor control systems are all undergoing dramatic change. There are many reasons for this, but most significant is the reducing cost of technology. Andrew Ward explains
Rapidly decreasing prices, standardisation and high-performance processing platforms, together with enhanced communications and virtually universal operating systems, are fundamentally changing what can be achieved by everything from simple automatic loop controllers to complex software simulations of entire factories.
“The megatrend has been the shift to software-based control”, explains Peter Thorne, director of consulting at industry-watcher Cambashi. “Over the years, control systems designers have figured out how to replace relays, switches and so on with software. Now, the software-based system is the expected implementation of control systems – partly because of better price/performance ratios, but perhaps more importantly the greater flexibility, and the scope it gives manufacturers to connect and integrate different parts of the organisation.”
With devices at all levels having much greater capabilities, boundaries are becoming blurred. It’s no longer possible to categorically state what each family of device does, and where it should be used – there’s too much overlap. The buying decision process has, unfortunately, become more complex, but the good news is that solutions that were once only contemplated by the largest multi-nationals are now within reach of the SME (small to medium size enterprises).
At the same time, external business pressures and the forces of e-commerce are driving the trend for more and better integration, all the way from shop floor control systems up to enterprise commercial applications such as enterprise resource planning (ERP) systems.
However, not all implementations in the past have produced the expected business benefits. In some cases, companies have embarked on complex major projects – such as MES (manufacturing execution systems) and ERP – without even clearly stating the business goals before they start. So the most important aspect of any implementation of factory automation systems must be a stated set of objectives, expressed in business terms. It’s not unusual to interview IT people about new technology and find they have no concept of what, if any, business benefits have been achieved, several million pounds down the line.
Guy Washer, managing director at market research firm Benchmark, believes strongly that manufacturers should be focussed on lowering production costs, cutting inventory and reducing logistics costs. “A strong currency and the skills shortage are outside their control – instead, they should be worrying about their internal business problems, identifying the benefits that shop floor integration can bring and then getting on with it.”
“Above all, automation helps the guy on the shop floor be more efficient and effective,” says Rod Blackwell, manufacturing industry marketing manager for Microsoft UK. One of the impacts of this, and a major driver for automation, is greater productivity, says Kevin Prouty, senior researcher at analyst AMR Research. “The next big driver is greater consistency and reliability, leading to improved [product] quality. Last but by no means least is getting data visibility – so analysis leading to action can be taken on that data.”
Automatic loop controllers
Relatively simple feedback devices that monitor sensors and then alter the setting of a valve or pump for example, automatic loop controllers save on manpower, but even they haven’t been immune to the changes sweeping over the factory floor. “Intelligent sensors and valves can make the loop controller redundant,” says Prouty. “Furthermore, these new devices embed chips that are like a very tiny PLC (programmable logic controller), and what’s important about them is that some now include Web servers, providing visibility of data by other systems across the network. The demand for data, and the need to run and operate your plant better in real time, will add enough cost benefit to loop controllers to justify these new features.”
Programmable logic controllers are devices that replace electro-mechanical systems to automate and control industrial plant. Being programmable, they offer more flexibility than the devices they replace, as well as the ability to integrate with other systems via a network. Thus, although they can be used in a stand-alone fashion, they frequently interface to a higher level of plant control, such as a SCADA (supervisory control and data acquisition) system.
Traditionally, PLCs have been based on proprietary hardware architectures and operating systems, but there is a trend for these to be replaced by the industry standard PC architecture and commercial operating environments. This isn’t without its drawbacks, as Michael Juniper, divisional manager at controls firm Omron, explains. “Open controllers give you a higher degree of freedom to implement your system, but also responsibility to make sure it only does what it’s supposed to do. You don’t want to get into the nightmares of cost of ownership associated with desktop PCs. A PLC has all the flexibility you need, but by being constrained, is also robust enough for the industrial environment.”
And PLCs are yet another class of devices where Web servers are intruding. “We have put Web server software on our ONC – Open Network Controller – so it can talk to factory automation networks and allow people to use Web browsers to access it,” says Juniper. “We have gone for a very simple concept that will easily allow you to view data on the shop floor.”
Ultimately, the ubiquity and low cost of the PC platform and standard operating systems will bring irresistible pressure to bear. “I am seeing more and more commercial technology that you would consider PC-based turning up in both PLCs and DCSs,” (distributed control systems) says Prouty. “PLCs are slowly going to transform themselves into black-box PCs. There was user and vendor perception that the commercial technology wasn’t robust enough, but of course it is now used in so many critical applications that it has proved itself.”
The range of text display, operator panels and touch-screen panels that provide an interface to PLCs are know collectively as HMI components, or human-machine interfaces. Gert Rohrmann, product manager at automation and drives giant Siemens, says: “These sit on the same network as the PLCs and work from the same database, so you automatically configure both HMI and PLC devices at the same time.
“HMI components are moving to [the] Windows CE [operating system], thus bringing the bells and whistles of the PC world to the HMI, along with the look and feel of Windows,” he continues. “But you must have the characteristics and deterministics of the HMI world such as instant-on, and no hard drive on a machine that is shaking about, for example.”
Supervisory control and data acquisition (SCADA) systems are used to monitor and control plant status and provide logging facilities, and usually interface to the plant via PLCs. Companies such as Siemens provide tight integration between SCADA, HMI and PLC devices. “You can call up the PLC from the SCADA – so if you have a motor flashing red on the screen you can go straight to the PLC section of the program referencing the motor, and perform diagnostics and tests all directly from the SCADA terminal”, says Rohrmann. Similarly, the SCADA and PLCs share the same data form, so there’s no need to export, convert or manually enter it, thus minimising engineering costs.
Distributed control systems
DCSs have their origins in the IT world, whereas PLCs have grown from the shop floor up. However, as the functionality, programmability and architectures develop at the PLC level, boundaries between DCSs and PLCs will continue to blur, says Chris Haines, business manager, Rockwell Software. “Traditionally, the grey area in the middle has been small – you’ve had simple, discrete operation low-cost PLCs on the one hand, and complex continuous control DCSs on the other hand. However, the grey area is now getting bigger and bigger.”
According to Haines, “DCS manufacturers are coming out with more cost-effective modular solutions to compete with PLCs, but PLC functionality is coming up, with scripts and architectures that are very similar to DCS architecture. Software integration and shared database tools on PLCs now give you similar functionality to DCSs.”
The result is sometimes confusing. “At almost any chemical plant, there is a DCS running the plant and a PLC control system running the controls,” says Juniper. “The DCS system is really only doing a supervisory role – a jumped-up SCADA system, if you like.”
Manufacturing execution systems
MESs carry out functions related to production engineering, including production planning, routing, scheduling and engineering modelling, explains Peter Fox, marketing and technical support manager, ABB Automation. “For example, MES may carry out tank farm optimisation – looking at the best way of blending oil, perhaps.”
But like everything else, the boundaries of MES are moving. “MES isn’t about control, planning at the ERP level, and it’s not about supply chain management,” Haines reminds us. “However, you do have data acquisition, plus process and batch management.”
Once again, overlap is occurring as HMI and batch management systems gain functionality and move into the MES area. “Scheduling is another area that can become a bit blurred,” says Haines. “Sometimes you get scheduling in the ERP system, but not with the same functionality that you see in the MES area.”
And standards are having their influence on MES too. Haines: “Our CX Server uses [Microsoft] OPC – OLE for Process Control – as part of our Factory Intelligent Networking, which allows you to take data from the shop floor and plug it into Windows applications using ActiveX (Microsoft software components). This is the technology that allows MES systems to communicate with our hardware.”
One of the biggest changes to MES systems is the increasing modularisation, believes AMR’s Prouty. “They were large, monolithic packages – virtually a mini-ERP system on the shop floor – that required a big plant and a big problem to justify their installation. Now, with a more componentised approach to MES, you buy just the parts you need and put them on an application server. In fact, in a lot of applications you might see MES squeezed out completely – if you have good application servers and analytical tools in the plant, you can go straight to a suitable ERP system. But with a complex manufacturing problem and a large conventional ERP system, you will still need MES.”
Factory simulation systems have been increasingly popular as the prices becomes more acceptable and obvious benefits begin to flow. “We’ve built a complete factory simulator for one customer for staff training purposes, and not only have they benefited from training staff off-site, but also they’ve identified improvements to the manufacturing processes,” says Fox.
However, Sal Spada, senior analyst with consultancy ARC Advisory Group, says: “Adoption of advanced simulation and control for discrete manufacturers will stay with companies that have fairly high production – those that combine custom manufacturing with some level of batch. In general, it doesn’t seem to be going down to the second tier of manufacturer – the cost/benefit relationship just isn’t there yet. But Ethernet is certainly going to be of help in bridging the gap between simulation systems and the factory floor.”
But costs are coming down, and simulation is penetrating into smaller plants, and even discrete industries, says Prouty. “As control systems become more inexpensive, so too are simulation systems. I think in the future a PLC will come with a simulator as standard.”
Author: Andrew Ward