Need for smart field devices
At present, innovative concepts in the field of process engineering and in particular process intensification are being promoted for the analysis and design of innovative equipment and production methods. This leads to a considerable improvement in sustainability and efficiency. Environmental performance will be improved, for example, through alternative energy conversion. Such concepts also improve hydrodynamics and heat transfer within processes.
Current research focuses on intensifying continuous processes. Compared with conventional batch processes, it allows for more intensive continuous production, the adoption of new compounds that are difficult to produce (see example in the “Knowledge from the digital twin: hybrid models, artificial intelligence and machine learning for process intelligence” section) improves product uniformity and significantly reduces the consumption of raw materials and energy. In addition, continuous processes often meet higher safety requirements, as they only produce hazardous chemicals when needed and do not have to be stored in large quantities. This is a clear advantage for plants located near or in urban areas, which is typical for most major chemical parks in Europe. Flexible (modular) chemical plants can produce multiple commodities with the same equipment, with short inter campaign downtime and fast time to market for new products. Typically, such plants are smaller in scale than basic chemical plants in batch production, but are still able to produce between kilograms and tons of specialty products daily.
A very high degree of automation is therefore a prerequisite for the realization of such benefit from intensified continuous plants. In continuous flow processes, continuous, automated measurements and strict product quality control are indispensable. If these are not readily available, there is a high potential danger of producing large volume of out-of-spec (OOS) products.
In pharmaceutical production, the agreed long-term vision is “continuous manufacturing” (CM), based on “real time release testing” (RTRT), i.e., risk-based and fully integrated QC in every process unit. For flexible production, modular facilities based on standardized architectures to be developed in the near future will enable plug-and-produce approaches suitable for small batches. This provides a flexible connection of smaller production plants, production transfer to fully autonomous plants, less user intervention, less non-productive time, and a continuous process knowledge over the product life cycle, future expertise, and a faster market entrance. It is also supposed that the costs for quality control within a CM concept will be substantially reduced at the same time.
The increasing demand for automation in general and PAT sensors in particular will changes the way PAT engineers design, install, and operate sensors in the field, especially in the context of modular plants and the reusability of respective PAT modules. A smart field device, which is capable to fulfill this, does know what its colleague devices in the smart labs—as defined in Part 1, our preceding contribution —did during the process development and connects to the sensors and actuators around it. Furthermore, smart field devices have process connections, built-in safety, and regulatory specifications already implemented to ease the installation and enhance the reusability of PAT. This will increase the yield of chemical information and will significantly support the respective control strategy, if combined with process information.
“Chemistry controls chemistry”
A production plant in the process industry usually changes the chemical and physical properties of the used substrate. Thereby, an increase in value is achieved by using knowledge how to transform raw materials into value products and goods. In chemical plants, for example, a substrate like propene is oxidized to acrolein or acrylic acid. Therefore, the chemical plant breaks and makes chemical bonds to achieve an added value. Still, control strategies within the process industry mainly focus on classical instrumentation measuring variables like pressure, temperature, level, or flow and not specific information (i.e., “chemical” such as physico-chemical properties and chemical reactions) that present the changes in chemical structure and composition. For example, measuring the functional groups that are changed by the plant (C=C, C=O, and COOH) during propene oxidation would obtain the concentration of main or by-products.
Future processes will need to be more flexible in using a broad range of raw materials, such as bio-based materials for circular economy solutions. So far, the substances added to a process show very small ranges of fluctuation in their specifications. In contrast, renewable and recycled input materials are subject to severe changes and fluctuations of the bandwidth of their properties—depending on their specific origin. However, since these raw materials are increasingly used, new requirements for PAT tools need to be met—e.g., due to fouling of measurement probes that need to be overcome .
The flexibility with respect to the use of raw materials is, for example, shown today in energy industry using secondary fuels or biofuels. This results in new requirements for the chemical analytics of the raw materials in discontinuous processes (e.g., incoming goods inspection) and continuous processes (e.g., material flows in pipelines and on conveyor belts).
For the chemical industries, the monitoring of chemical information is the key to “chemical” process control. A chemical factory that performs chemical reactions is controlled in a closed loop by specific information. As an example, a case study was made of a given aromatic coupling reaction step (lithiation reaction). The project challenge was to integrate a commercially available low-field NMR spectrometer from a desktop application to the complete requirements of an industrial automation environment, including accurate interpretation of measurement data .
Today, in process industry, spectroscopic methods are increasingly applied to measure specific chemical information online. For example, NIR spectroscopy is often applied for online analytics in the liquid phase and is an excellent industry-proven tool for a wide range of applications. The need of future flexibility demands more chemical sensors revealing comprehensive chemical information, like Raman spectroscopy, mass spectrometry, or the above-mentioned NMR spectroscopy. It will be a joint task of the PAT users, research, and equipment and software manufactures bringing these “too complex” analyzers to the field through encapsulation of the complexity by modular approaches.
Current and future requirements to smart process sensors and actuators
Sensors are the sensing organs of industrial automation. There are presently profound developments in information and communication technology that provide great opportunities for optimized process control and added value with dedicated cross-linked communicative sensors. These types of “intelligent” sensors can provide services within a particular network and use information from there. Consequently, smart process sensors enable new innovative business models for users, device manufacturers, and service providers.
Recently the technology roadmap “Process Sensors 4.0” was published. It describes the necessary requirements as well as the communication capabilities of such process sensors—from simple temperature sensors to state-of-the-art technologies still being developed. Important smart features are shown in Fig. 2. Many of these requests have meanwhile been successfully and reliably realized in first case studies being evidence of excellent cooperation between users in the process industry and their device and software manufacturers in committees and joint projects.
The cost of connectivity is dropping dramatically, providing the great chance to connect people, assets, and information across the industrial enterprise. While only providing add-on information for the initial stage, future comprehensive cloud services may not require a high disposability or real-time capabilities. Though, worldwide known tech companies already demonstrated the available cloud computing power with respect to real-time data analysis—but for (process) industry, real-time transmission of complex and a large number of signals/data is a major challenge. If these are given in future, even process control tasks will be possible using cloud services, e.g., when complex computing algorithms are needed, which require more computing power than can be possibly provided by edge computing devices.
Future requirements to communication/connectivity
In automation technology, there is currently a large number of process control systems (PCS), such as automation and control systems, operating and monitoring systems, and manufacturing execution systems for controlling a process, while process information management systems (PIMS) are used for data acquisition and evaluation. The transitions are smooth. Meaningful data acquisition takes place in PIMS together with stored process steps (“recipe”), which are called up as required and specify all raw materials, materials, and production equipment used (plant, reactors, plant equipment). In this way, deviations can be stored in the system accordingly. The database contains further fields for all relevant data from production, quality control of substances and materials, and the product.
Access to the system is ideally secured by, e.g., controlling user roles or traffic monitoring and can be logged in an audit trail if required. In critical cases, it is even today already possible (but not yet implemented as industry-standard) to only allow raw materials to enter production if their identity and specification are unique—for example by using fast fingerprint methods such as Raman spectroscopy. Of course, data safety needs to be solved as well, but is out of scope of the present contribution. In areas subject to regulatory supervision, opinions on the need to store in-process data or to document OOS results are often less sharp than those for release data . From a technical point of view, a complete intermediate storage of all measured data is not always necessary if, for example, the frequency of data collection is very high. Depending on the dynamics of the process step, representative data should be stored.
How can complex process analytical devices be integrated? In the field of industrial communication, an unmanageable variety of bus systems are used to transmit complex information: Industrial Ethernet, Profinet, Modbus, AS-Interface, IO-Link, Industrial Wireless Communication—just to name a few. For some procedures, which originate from the laboratory environment, no professional bus systems are (yet) available and one is dependent on the connection of a (local) evaluation computer. This is unacceptable from the control technology point of view if no status control of the evaluation computer is installed in order to achieve the necessary robustness and to establish a safe operating point in the event of system failure.
In order to achieve this, steps were taken towards simplification and standardization, because today automation components in a plant are not at all standardized. Thus, a uniform protocol and a uniform fieldbus are required for trouble-free communication between all automation components. Meanwhile, the standard OPC Unified Architecture (OPC-UA)  is considered to be set and can be regarded as a small triumph of industry 4.0. Non-ethernet field buses are still dominant today against the background of a grown landscape in existing plants and the often very special requirements for power supply and explosion protection.
In the figurative sense, OPC-UA is comparable to the PDF standard or HTML standard, which define the properties of graphical objects, e.g., print products. It is also independent of manufacturers or system suppliers, programming language, operating system, or communication standard (e.g., fieldbus) and standardizes the underlying data format, e.g., for online measured values. The German Federal Office for Information Security (BSI) confirmed in 2016 that OPC-UA can be used to implement IT-safe industrial 4.0 communication .
Because the maintenance and operational functions are of great benefit, some innovative companies in the process industry are currently covering their plants with additional network access, mostly wireless technologies. Companies have also started to completely digitize their asset and plant plans. The question whether the high amount of information needs to be integrated in the classical pyramid automation structure or there might be another interface between the classical PCS domain and the monitoring and optimization domain is covered by the NAMUR open architecture concept (NOA) [10, 11].