Non-Invasive Digital Data Extraction From IIoT Devices

By Hubert Yoshida posted 09-17-2020 17:37

Information technology/operational technology (IT/OT) convergence is a rising priority for all manufacturers. Success in the digital age will hinge upon data utilization — the most forward-thinking manufacturers work on becoming data driven in their decision making.

Manufacturers view IT/OT convergence across three horizons to create a strategic road map for transformation:

  • Align, communicate, and enable cooperation for innovation, compliance, and security.
  • Maximize the value of operational data through merged data streams, AI, and OPM (Operational performance management - the alignment of all business units within an organization to ensure that they are working together to achieve core business goals.
  • Accelerate data-driven decision making to empower resilience.

The challenge for manufacturers is the merging of OT data with IT systems where IT tools like Artificial Intelligence (AI) has the potential to transform manufacturing by improving shop floor processes such as production, maintenance and quality. Industrial datasets are notoriously difficult to extract in a real-time, streaming fashion thus, negating potential AI benefits. The main example is some specialized industrial controllers that are operated by custom software which complicates the process of connecting them to an Information Technology (IT) based data acquisition network.

A PLC (programmable logic controller) is an industrial digital computer which has been ruggedized and adapted for the control of manufacturing process such as assembly lines or robotic devices or any activity that requires high reliability, ease of programming and process fault diagnosis. PLCs can range from small modular devices with tens of inputs or outputs to large rack mounted devices with thousands of I/O which are often networked with other PLC or SCADA systems.

SCADA is a central system used to monitor and run plant processes. It is typically software installed on a computer, and one of its major functions is to act as an interface with industrial machines (or Human-Machine Interface, or HMI). In other words, it allows users to track information coming in from equipment, enter commands, make changes to their programming, etc. SCADA systems are often used in conjunction with PLCs and other devices (in fact, some would say that a PLC would be part of a SCADA system). Data from PLCs and Remote Terminal Units (RTUs) are relayed to the system, and commands are entered into the HMI to make adjustments to the processes they control.

Used together, SCADA software and PLCs form an automatic system for prescribing maintenance tasks, forming the core of a predictive maintenance system. It works something like this:
  • Data from sensors on individual assets is transmitted to the PLC
  • The PLC translates that data into a format that can be used by the software
  • Users access the data through the HMI on the software
  • If the data crosses certain thresholds, a maintenance work order is created
For instance, if a turbine is showing too much vibration, sensors transmit that data through the system, and the readouts on the user end would trigger a work order. In this application, SCADA software controls the entire system, while PLCs act as relay points and controllers for specific assets.

While this data is useful for predicting future events based on historical data, the additional application of AI provides prescriptive analytics which goes beyond simply predicting options in the predictive model and actually suggests a range of prescribed actions and the potential outcomes of each action. Since a prescriptive model is able to predict the possible consequences based on a different choice of action, it can also recommend the best course of action for any pre-specified outcome.

The problem with connecting PLC and SCADA data to IT systems for AI analytics is the invasive nature of that connection. Specialized industrial controllers are operated by custom software which complicates the process of connecting them to an Information Technology (IT) based data acquisition network A large part of any complex SCADA system design is involved in matching the protocol and communication parameters between connecting devices. There are about 200 such real time user layer and application protocols. These include both proprietary and non- proprietary protocols. Security concerns may also limit direct physical access to these controllers for data acquisition. Connecting closed industrial networks to IT data acquisition networks could expose the industrial network to cyber-attacks such as the Stuxnet malware attack on the Iran’s Nuclear Facilities.

To connect the Operational Technology (OT) data stored in these controllers to an AI application in a secure, reliable and available way, Hitachi has proposed a novel Industrial IoT (IIoT) solution. In this solution, Hitachi demonstrates how video cameras can be installed in a factory shop floor to continuously obtain images of the controller HMIs. Hitachi proposes image pre-processing to segment the HMI into regions of streaming data and regions of fixed meta-data. We then evaluate the performance of multiple Optical Character Recognition (OCR) technologies such as Tesseract and Google vision to recognize the streaming data and test it for typical factory HMIs and realistic lighting conditions. Finally, we use the meta-data to match the OCR output with the temporal, domain-dependent context of the data to improve the accuracy of the output. Our IIoT solution enables reliable and efficient data extraction of OT data which will improve the performance of subsequent IT and AI applications.

For more information on this approach please see the Hitachi Industrial AI blog post by Ravneet Kaur: Digital data extraction from industrial monitors using video analytics