Control-level design challenges in smart factory automation systems

The Industrial Internet of Things (IIoT) is often used to address new concepts that apply to multiple industries that share requirements beyond the Internet of Things (IoT) such as real-time communication and controls as well as increased safety, security and reliability. A term used as a further subset of the IIoT is Industry 4.0, which applies IIoT concepts to manufacturing and factory automation. While these movements are broad and future-looking, a few key concepts from these movements are starting to be implemented. Concepts like smart maintenance based on predictive analytics and distributed intelligent systems bring many benefits such as higher flexibility improved efficiency and reduced down time due to maintenance. However, the implementation of these concepts expands existing design challenges for factory automation equipment.

Factory automation systems designed for Industry 4.0 typically contain three main levels of equipment that drive real-time communications and control:

  1. The field level where I/O modules, actuators, and drives are engaged in the physical operation of the factory,
  2. The control level where PLCs or CNCs collect information from and issue commands to the field level,
  3. The operator level where human machine interface (HMI) devices communicate information with operators and where operators can issue commands.


Figure 1. Three levels of equipment (Source: Texas Instruments)

Each of these levels has its own set of taxing design challenges that require optimized hardware and software solutions. The challenges related to the control level are particularly difficult to meet due to a number of factors, such as increasing manufacturing efficiency and the enablement of Industry 4.0 concepts. Pressures to increase automation and decrease the price of solutions drive up the number of nodes each controller is expected to communicate with, but real-time requirements remain.

At the same time, designs become more complex as manufacturers begin incorporating new concepts that are coming out of the Industry 4.0 movement. For example, supporting cloud connectivity to report data used for predictive analytics requires a mix of real-time and non-real-time functions to be supported from a single device, and the real-time performance must increase from generation-to-generation to support additional field level nodes. Control level equipment faces the design challenges that apply to all industrial automation designs- power consumption, long supply lifetimes and reliability requirements.

In addition to these concerns designers face certain challenges that arise from the increasing number of nodes that one controller can support. A larger number of supported nodes means that fewer controllers are needed in the overall factory solution to create a less expensive automation solution or that more nodes can be supported in the factory for higher levels of automation. However, with the increase in the number of supported nodes, the processor performance must also increase while still consuming low enough power as to not force an increase in the size of the enclosure.  Additionally, most PLCs are designed to not require a fan so power dissipation is a key design aspect.

With PLCs and CNCs concurrently controlling a high number of nodes or functions in the factory, the real-time aspects of their operation is critical.  Two components are needed for a solution to the tight timing requirements: real-time operating systems (RTOSes) and flexible time-aware peripherals for industrial communications. Real-time operating systems are used in these types of equipment for their determinism and ability to control latency in order to meet the critical timing needs. Commercial RTOSes have been widely used in industrial control applications for many years, and there is also a growing interest in RT Linux solutions which have all the advantages of the large open source community around Linux while adding the time-awareness and determinism required for industrial automation applications.

Next Page >>

For the communications peripherals portion of the real-time solution, the main need is to support industrial fieldbus protocols in a way that allows for low latencies and reduced protocol cycle times  even when there is the demand for an increased number of nodes. It becomes an even more complex challenge when it is considered that multiple fieldbus standards must be supported from a single design. Multi-protocol support is necessary in order to make the end products compatible with the multiple standards that may be found already in use in a factory- such as EtherCAT, PROFINET, Ethernet/IP and more. Support for multiple protocols is complex to solve with hardware (ASICs), as different board designs will be needed for each supported fieldbus since each protocol may require its own specific ASIC. It is a smaller challenge if a programmable approach is taken where a software or firmware change is all that is needed to change the fieldbus protocol.

To facilitate this real-time communication solution, controllers need a large number of peripheral interfaces since they communicate on multiple levels with the fieldbus networks in the factory, with the backplane for the connected I/Os, actuators, drives, or with other controllers. Due to the implementation of Industry 4.0 concepts, they are also communicating with data collection servers for predictive analytics using protocols such as OPC UA. All of this together drives a need to have a large number of peripheral interfaces, especially Ethernet and to also have a flexible, programmable communications solution.

Industry 4.0 implementations require systems to efficiently support real-time and non-real-time tasks can also pose challenges. There needs to be separation of the tasks so that an error in the non-real-time task—transferring analytics data to a server— will not cause a failure or delay in the real-time task of communicating with field nodes. One solution would be to put the tasks on completely separate processors; however this adds cost and increases the power consumption and size of the design. Multicore processors provide a better solution on the hardware side, and software virtualization allows partitioning of real-time tasks onto one core or cluster which runs a high-level OS like Linux, and non-real-time tasks onto a separate core or cluster running an RTOS. Designers must pay attention to resources required for each task, but hypervisor solutions that provide a framework like this are available from the popular RTOS vendors for the industrial space.

Another effect of Industry 4.0 is that wireless connectivity technology is beginning to find its place in the industrial space. This is driven by the substantial and measurable cost savings in engineering, installation and logistics that wireless communication promises. The increased data throughput, security, reliability, interoperability and flexibility of Wi-Fi networks make them an increasingly viable option for industrial applications. Wireless networks are up to 10x cheaper than the wired alterative and engineering costs are dramatically reduced as extensive planning is no longer required to route wires. Wireless connectivity solutions carry the advantage of easy provisioning, meaning that these networks can be easily re-configured without the need of re-wiring or changing hardware setup. To save power, wireless system on chips (SoCs) are designed to enter extremely low-power states, consuming only few milliwatts in connected idle state. Wireless SoCs can be easily integrated with most microprocessing units (MPUs) or even microcontrollers supporting SoC-specific bus interfaces such as SDIO. While real-time industrial Ethernet communication is not moving to wireless connectivity in the near term, applications like maintenance through over-the-air upgrades and safe flexible data gathering are primary drivers for the integration of wireless technology into factory automation equipment. The TI WiLink Wi-Fi and Bluetooth/ Bluetooth low energy combo connectivity modules make a great wireless connectivity choice for the Sitara line of processors with a complete solution available today.

While Industry 4.0 and more broadly the IIoT continues to grow, so will the challenges in factory automation systems, especially those related to communication. A flexible, programmable communication solution like the PRU-ICSS on TI’s Sitara ARM processors, gives developers the ability to create more effective factory automation solutions by adapting to changing standards and upgrading to new communication protocols as they are introduced to the market. The PRU-ICSS is a fully-deterministic real-time communications solution that can be programmed by developers or used with firmware provided by TI for EtherCAT, Profinet, Ethernet/IP, and more industrial standards.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.