Drives & Controls October 2022

26 n BIG DATA October 2022 www.drivesncontrols.com How to liberate stranded data T he automation systems deployed by manufacturing organisations can generate vast amounts of process and plant data, which can be used to improve operational performance and profitability. However, achieving these aims depends on being able to access, manage and analyse this wealth of data effectively. Unfortunately, it is all too common for this valuable data to remain stranded and inaccessible. This can be as a result of legacy technologies being incapable of handling the data, or because of traditional design philosophies that focus on basic functionality at the expense of data connectivity. Organisations are now looking to build big data analytics capabilities into new systems and add it to existing operations. Data sourced at the edge can now be pre-processed and contextualised before being transmitted to the cloud for deeper analysis, which can lead to new insights and improved business outcomes. How data becomes stranded Until recently, most manufacturing data was obtained from PLCs, HMIs, Scada and historian systems in the operations technology (OT) domain. Because the aim of these systems has been to maximise operational efficiency and uptime by providing control and visibility, their ability to access and analyse data beyond the realm of immediate production goals has been somewhat neglected. Because their OT infrastructures having been designed and scaled according to operational needs, organisations have typically chosen proprietary communications protocols that meet performance requirements but do not support cross-vendor connectivity and interoperability. Collection of control and sensor data has been minimised to increase system reliability and simplicity. Localised on-premises architectures have been implemented to minimise cybersecurity risks, and vendor lockouts have been introduced to protect intellectual property. The resulting systems may perform admirably in achieving operational goals, but by failing to collect valuable sources of data such as environmental conditions, condition monitoring information and utility consumption, organisations are missing out on some significant benefits. Let’s examine some of the various forms in which stranded data exists. Isolated data Data becomes isolated when plant assets have no network access to any type of OT or IT system. For example, a standalone temperature transmitter with a 4- 20mA connection or even Modbus capability would need to connect with an edge device such as a PLC, edge controller or gateway to allow this datastream to be accessed. The data is often not critical to machine control and therefore unavailable via traditional legacy PLC/Scada data sources. Accessing the data through the nearest machine PLC risks voiding OEM warranties because of the need to change programming logic. Ignored data Many intelligent edge devices provide basic data – a smart power monitor, for example, can provide information such as volts, amps, kilowatts, kilowatt-hours and more. However, extended datasets, such as total harmonic distortion (THD), may not be transmitted due to a lack of application requirements, low bandwidth, or limited system data storage capacity. The data exists, but is simply not accessed. Under-sampled data When a smart device is supplying data to supervisory systems via a communications bus, the sampling rate may be too low, the latency too great, or the dataset too large for the results to be usable. The data may be summarised before being published, resulting in a loss of fidelity. The data produced by automation systems can unlock new insights – but only if it can be accessed effectively. Silvia Gonzalez, global product management director for software at Emerson, explains how edge-to-cloud IIoT systems are simplifying access to this valuable resource.