Decoding OT data secrets
In the past, ensuring the reliable transmission of operational technology (OT) data was the only requirement; now, it is necessary to also ensure the data’s quality.
Amidst the COVID-19 pandemic, system integrators in the industrial sector moved toward one of the most influential changes in recent history, industrial digital transformation (industrial DX). However, before engaging, it is important to understand what industrial DX entails, starting with its building block: operational technology (OT) data.
Consider a factory that converts toxic waste into organic fertilizer. If a slight temperature change occurs during the production process, the neutralization of the active agent would be affected and damage production capacity. In the past, inspections and adjustments had to be made by onsite staff. However, this method has long been proven to be ineffective in responding to sudden and immediate changes. Accurately predicting temperature changes within six hours before an actual fluctuation is crucial to making the necessary adjustments to maintain the optimal neutralization process. Several kinds of data are required to achieve accurate predictions including equipment operation data, controller data, field temperature data and weather forecast data. These types of operational data are the fundamental building blocks of industrial DX.
To create effective business intelligence, OT data not only needs to be collected but also analyzed to formulate a relevant strategy. Industrial DX can be viewed as a process to find and interpret the value of the existing OT data. This is easier said than done. As system integrators have discovered from connecting OT data, when the value of the OT data increases (i.e., expecting more out of the data), the difficulty to connect it also rises. As a result, OT data connectivity technologies’ original responsibilities — securely collecting, processing and labeling the data prior to transmitting it in a timely fashion — are now divided into more precise steps.
To respond to the influx of incoming data, great strides have been made regarding the method and speed of transmissions. OT data connectivity is now a hybrid specialty integrating domain knowledge and the latest technological capabilities. These advances in “OT data connectivity” are carrying the key elements of success in industrial DX.
Monitoring the present versus optimizing the future
In the past, the purpose of collecting OT data was to monitor and control the existing operational system. OT data is used to ensure machines are operating stably by keeping track of the devices’ current state on the factory floor. It also can be used to control the flow rate of an oil pipeline to comply with oil production targets. In other words, it only focuses on maintaining the “right now.”
However, industrial DX takes it one step further into the future. Obtaining OT data is no longer for the sole purpose of monitoring and controlling the present; integrating data to analyze the future is the main aim. By finding key factors that affect the operating efficiency, optimizing and creating new business opportunities, OT data has allowed many early industrial adapters to create brand-new business models. Consider a leading power system integrator as an example. The company applied the data it collected from the historical usage of methanol in hydrogen energy batteries to estimate the future energy usage of every customer. A new personalized charging plan was developed for each customer. The original charge-per-usage plan was upgraded to a machine-as-a-service (MaaS) monthly charging plan, creating a win-win transaction model for both parties.
Transitioning from a digit to an actual value
The line between IT and OT is blurred when OT data is used for further analysis. In the past, ensuring the reliable transmission of data was the only requirement. Now, it is necessary to also ensure the quality of the data. This has become one of the biggest obstacles for industrial DX. The lifecycle of industrial equipment is often very long, which means it accumulates a lot of OT data that is incomplete or in unrecognizable formats. It is up to IT professionals to perform additional data cleaning and conversion before it can be used.
The best-case scenario is data scrubbing takes a little longer, which costs money and precious time, but the data is useable. The worst-case scenario is the data cannot be understood and becomes useless altogether. For example, if the output data shows “5” without any labeling, it is impossible to decipher what this number represents. Without further investigation, one may never know the number indicates the machine speed. This miscommunication often is caused by a different format not recognized by the IT system. Since this phenomenon is common, one solution is to preprocess such data by converting it into the required format through a built-in program in the OT data connectivity device. Thus, data preprocessing gives the data context and makes it recognizable. The process of turning OT data into usable OT data — allowing it to have “analytic usability” — is an important step at the onset of the OT data revolution.
Complex data sources and types
The traditional control system already relies on a multitude of OT data to maintain daily operations. Simple data, such as the position of the water tank gate, daily oil production, etc., shows basic information about operating status. Complicated data, such as production recipes or processes, also gets generated. However, industrial DX requires more.
Consider the renewable energy industry. To quickly remove the shadows or stains on solar panels, more data is required. In addition to monitoring power generation, it needs environmental information such as temperature and humidity. This data coupled with the live feeds from surveilling drones and analysis by an AI platform can find the more precise position of the contaminated solar panel. Armed with this information, real-time and precise maintenance can be scheduled. Hence, large volumes of OT data from diversified sources reduce traditional capital expenditures and improve production efficiency.
Changing from linear control to circular feedback in real time
Traditional automation systems place great emphasis on real-time control. OT data is often used as an indicator of a specific time slot on the linear control process. The data’s purpose ends when the specific process is over. However, industrial DX emphasizes a different type of real time by focusing on the “OT data collect/analysis/feedback” loop.
With the big data processing technology, faster networks and maturing industrial computing capabilities, IT can analyze uninterrupted OT data and provide immediate feedback to the operational equipment after analyzing the data. This loop of receiving data, analysis and feedback allows enterprises to perform real-time adjustments. Consider small to medium-sized manufacturers served by KPMG. To reduce wasting manpower hours and material resources caused by defective products, more OT data, such as vibration, temperature, speed, current, etc., is collected, uploaded and analyzed by the AI platform. Through analytics, we have learned when the tool-current frequency of a certain machine is too high, this means the tool is worn out. The tool can then be replaced in advance to ensure high-quality output.
There will only be more, not less
In the era of Industry 4.0, large-scale automation systems, (such as a distributed control system (DCS) in an oil refinery, are capable of processing large volumes of OT data per second. However, this data is only used while the equipment is running. Once the operation is over, interpretation of the data also ends. OT data is only used to interpret the present.
However, industrial DX takes it one step further. Using a larger amount of data, simulations and analytics can be performed to improve real-time operation efficiency and control operational risks. For example, to avoid crowded carriages during the pandemic, Taiwan’s railway company installed pressure sensors on its trains to measure carriage loads. Before a train enters the station, the sensors will send information along with the feed from the onboard CCTV of each carriage to the control center. This way, the control center will have an accurate view of how congested each carriage is and provide this information to the passengers waiting on the platform or notify management to help evacuate crowds.
Data security equates enterprise and national security
Although cybersecurity is not always the main concern when talking about OT data, it is a big priority for industrial DX. Much of the OT data comes from critical infrastructures (such as equipment monitoring in water plants and power plants) or important operational information in key manufacturing facilities (such as oil refineries and semiconductor factories). Such information if maliciously altered could cause immeasurable and monumental losses.
In February 2021, hackers entered the supervisory control and data acquisition (SCADA) system of the Oldsmar public water treatment plant in Florida by taking advantage of outdated versions of Microsoft Windows operating systems and poor network security. The hackers planned to raise the sodium hydroxide content in water to a range that could harm humans. Fortunately, the onsite operators discovered the abnormality and prevented the threat from being carried out. As the threat of cyberattacks increases, cybersecurity should be prioritized as more industries may fall victim to such attacks, which could yield significant consequences.
Industrial DX is shining a spotlight on previously mysterious and unnoticed OT data. This transformation has catalyzed the integration of information technology/operational technology (IT/OT) in knowledge, operation, security and personnel mentality.
Original content can be found at Control Engineering.