The introduction of cloud (specifically Software as a Service or “SaaS”) has had a profound impact on the methodology that organizations use to roll out software. The speed and scale that SaaS moves across a company is unprecedented and so much so that it is difficult to even compare to previous non-SaaS projects. So the question is, how can paper producers and converters also take advantage of this software paradigm shift without introducing risk to the stable operation of their plants?

Cloud has some obvious advantages when compared to traditional software, most notably it provides these value drivers which are all built into the SaaS pricing as part of the software design paradigm:

• Experts (not on your payroll) that maintain and manage the hardware, disaster recovery, and networking infrastructure

• Automatic software updates and maintenance of the platform and application(s)

• Centralized configuration and management of applications

• Ability to securely access the environment from all devices without punching a hole in the corporate firewall

• Options to federate out applications and processes that are defined and managed in cloud but run locally at the edge (aka fog) layer

These value drivers are important because in the past the deployment of new software meant that there was a large amount of overhead to not only manage the initial setup, but also to control the variations and configuration nuances that occurred over time. These expenditures are the hidden cost of software that are often spread across multiple departmental budgets in order to provide full coverage and are often overlooked when talking about legacy compared to SaaS applications.

Can this now be transferred to the production processes in paper manufacturing? And how should you proceed? To understand this, we first have to take a closer look at what are some of the areas where software generates the most overhead for organizations to manage over time. Since the answer varies across each manufacturing location and the heritage of how it matured over time, you can expect that there is homework to be done first. These questions are a good start for what to tackle up front when discussing business investment priorities and target locations for the investment into updating processes, methodologies and the technology that supports them:

1. How many manufacturing locations have software over 5 years old? Of these, how many are managing critical tasks?
2. How much technical debt has been incurred over time for both homegrown and commercial software? What are the supporting technologies required for them to operate and are they also separately licensed?
3. What kind of staffing is required at each site to simply keep the lights on for hardware and infrastructure management?
4. What is the value of the current system and how long will it keep its value for the manufacturing site?
5. Does it have a role specific for the site or does it influence the broader organization and how?
6. What is the level of effort required to report across multiple locations and to maintain these reports over time?
7. How well coordinated are planning, logistics, operations, quality, and maintenance teams?

With these answers fresh in your mind there is likely some growing interest in how exactly this would work to fit for the paper industry and your needs. 

Success Stories of Cloud in Manufacturing Plants

While this may still sound like a dream of the future for the paper industry, there are already experiences with cloud-based MES solutions in other industries, for example in the automotive industry. This company, Smart Press Shop, used a cloud-first development strategy for building core ERP and manufacturing execution systems (MES) to digitalize production from start to finish – and run entirely in the cloud using 100% green energy.

It makes sense to take a closer look at the details to be able to better assess further developments for the paper industry, especially with regard to issues such as feasibility and reliability. What has become clear is that even for cloud-based systems, MES options for expansion and for customer-specific adaptations are absolutely necessary. Every company has its own individual manufacturing processes and precise ideas of how, for example, the user interfaces must be designed to provide optimum support for the workers. Accordingly, the solutions must be flexible and adaptable through the use of APIs and extendable by additional industry- und customer-specific functionalities.

The same applies to integration with existing automation systems and databases such as historical process and energy data. Many companies have been able to derive great benefits directly from this. For example Arpa, a manufacturer of surface materials, wanted to lower resource consumption and waste while keeping quality and time to market high by optimizing the product’s complex production process. They designed and built a new factory from scratch built on SAP software. SAP Manufacturing applications automate production and operations are optimized through the collection and analysis of data from each system, subsystem, and sensor. AI and machine learning technologies help Arpa constantly improve performance and sustainability. Smart, automated fine-tuning of the manufacturing process reduces energy and water use by 80% while machine learning algorithms slash scrap waste to near zero. Productivity increased 6-fold with a 24x7 production cycle powered by autonomous, laser-guided vehicles (LGVs). And €750k in production costs savings were realized during the factory’s first year of operation.

The analysis capabilities and the application of intelligent technologies such as Big-data, ML/AI and predictive analytics in the MES and ERP systems are clearly superior to those at the operational level in SCADA and automation systems. In the combination of operational and business data, there are completely new ways to find the parameters that influence product quality or optimize energy consumption.

Another important point is that although cloud-based solutions have developed rapidly in terms of reliability and security, there are often still good reasons for local implementation at individual sites. Among other things, the Internet connection in remote locations, high data rates and security can play a role here. Accordingly, a hybrid approach is required here, which, where necessary, also allows an edge, i.e. local, implementation in addition to the purely cloud-based setup. However, all of this should take place within a uniform architecture to enable uniform setup and evaluation.

For more information on SAP Digital Manufacturing Cloud go here.

Click if you are interested in a demo.

We covered Level 0 (instrumentation) in the last 2 articles. We now have a good foundation in the field for a 4th industrial era (I4) system. We now have to connect the field to a network, digitize, create automation logic, and make data available to the rest of the system.

The device that is responsible for these capabilities is generally referred to as a “controller”. Sometimes we are referring to a programmable logic controller (PLC), sometimes it is a controller in a distributed control system (DCS), and sometimes it is simply a data acquisition device. In all cases, a “controller” connects to the instrumentation and puts it in digital form in the memory of the device. This is what we mean when we use the term “digitize”; we are taking an analog signal in the form of electrical current or voltage and turning it into a value represented in computer memory with bits and bytes.  

The memory location may be an address, a tag, or a parameter on a tag that has a complex data type. In older generation PLCs, an address would be a data structure indicating the type of data and the physical location of the wiring. An example might be I:3/4, which means the data type is an integer, it is on IO card slot number 3, and it is the 4^th channel (pair of terminal screws) on that IO card. This is not very helpful if we have no idea what this integer represents. That is why controllers have tags. A tag might be “34TI1021”, where process area 34 might be the Kamyr digester, TI means it is a temperature input, and 1021 might be the loop number for liquor temperature. Modern controllers can have tags with complex data types. If the liquor temperature has a PID control loop, we can have tag 34TIC1021 with parameters for the process value (PV), controller output (OP), setpoint (SP), and many other parameters for configuration and tuning. The analog input value would end up in the PV parameter, which makes it convenient for finding the value and implementing the PID loop.

The ”digitize” functionality is not new in I4. We have had this capability in industry starting with I3. This often gets confused, because we see a lot of literature talking about I4 and digitization as if this is new. It is only new at higher levels in the control system hierarchy (specifically Level 3, which will be addressed later).

When we connect to instrumentation, we are not always connecting to an analog signal. Today we have digital bus protocols (Fieldbus, ProfiBus, CANOpen, DeviceNet, etc.) and wireless gateway protocols (WirelessHART, ISA100, etc). In these cases, the digitization happens in the instrumentation and the protocol is digital. This does not change the representation in the controller memory.  

In some cases, the controller we are referring to is a data acquisition device. This device is intended to perform the digitization, but does not have additional capability. Those capabilities will be the subject of next month’s article.

Thus far we have described I3 capabilities. I4 pertains to connectivity through the internet. In order to connect to tags in a controller, capabilities such as OPC and MQTT have been developed. OPC has been around and is fundamental in most I3 control systems. MQTT is a recent capability in modern controllers that facilitates connectivity to tags throughout the system. The goal in I4 is to have a unified name space; this means that when we are connecting to tags, those tags should be the same regardless of what level we are in the automation stack. This needs to be considered at Level 1. If we do not take this into consideration at this stage, it will be very difficult to have an I4 system that is manageable. (For more detail, refer to the Industree 4.0 column “MQTT” in August 2020).

Next month, we will explore the additional capabilities in these controllers that enable I4.

There are many topics grabbing the headlines these days but Artificial Intelligence (AI) is doing a good job of pushing more urgent tasks out of the way. Or is it the most urgent tasks?

It is hard to tell. Some learned people think AI may be on the verge of becoming smarter than humans. From the destruction of jobs to taking over the world, one can find articles and comments of nearly any flavor one desires.

I am of the camp of "let's go slowly" and make sure we have all the safeguards in place that we may need.

Straddling the fence, I also see a lot of promise out of further AI advancements.

We have already come a long way, in both speed of hardware as well as software.

I remember in the late 1970s, appropriate scientists were telling us we had gotten very very good at predicting the weather. How did they know? Due to the slow speed on the hardware at that time, three days after a particular weather event occurred, the prediction that had made before it occurred came spitting out of the computer. Since then, processor speeds have multiplied many fold and we now get the forecasts as a predication, not as a historical fact.

I see the day coming when our manufacturing plants can enjoy such prognostications, not just a few days in advance, but a month or more at a time...raw material supplies, equipment efficiencies, downtime, finished goods and market demand.

At that point, our work today in Industry 4.0 will look soooo old fashioned.

Prototyping has certainly come a long way since I started my career back in those days of yore that we used to call the 1980s. One of the prototyping techniques we employed was called wire-wrap, which we thought amazing at the time, but which now seems incredibly “clunky” in hindsight.

The Internet of Things (IoT) is helping industries worldwide become more efficient with managed and scalable digital solutions. More specifically, the Industrial Internet of Things (IIoT) focuses on connecting machines and devices in key industries such as oil and gas, hydropower, as well as manufacturing. In factories, the application of connected sensors to machines is being used to collect valuable data for condition monitoring and predictive maintenance purposes.

Industry 4.0 technologies give various industries the opportunity to address safety concerns and simplify manual tasks affecting productivity. The construction industry is no exception.

An Israeli semiconductor specialist has developed an AI-based vibration-monitoring chip that processes vibration data on-board, greatly reducing the amount of data that needs to be transmitted to the cloud, thus cutting power consumption and supporting energy-harvesting designs. Polyn Technology says that by reducing the amount of data transmitted by a factor of around 1,000, its VibroSense chip will cut costs and improve ROIs.