Enabling Real-Time Intelligence Using OPC

Capturing industrial equipment telemetry data in a standardized manner has been a challenge ever since machinery started to be instrumented. During the 1990s various efforts to standardize interfaces emerged. One of the most popular and enduring was the Open Platform Communications (OPC) standard.

OPC is maintained by an industry consortium called the OPC Foundation. The membership includes equipment manufacturers, partners, and software vendors. The foundation manages several standards, certifications, and products. But in the streaming analytics world the most commonly seen specifications are OPC DA, UA, and HDA.

OPC Data Access (DA) - The original specification from 1996, originally based on Microsoft DCOM technology, and still closely aligned with the Windows ecosystem. This API is designed to provide real-time access to live data from various data acquisition devices.

OPC Historical Data Access (HDA) - A specification for the exchange of historical, archived data rather than live data. Like DA it is based on Microsoft DCOM technology.

OPC Unified Architecture (UA) - A cross-platform version of the Data Access API, based on open standards and enabled for modern cloud-based architectures. Like DA, this API provides real-time, event-oriented access to live data.

OPC Unified Architecture Historical Data Access (UA HDA) - Like DA HDA, a specification for the exchange of historical data, but following the cross-platform UA architecture.

Integrating industrial data acquisition equipment such as PLCs with OPC servers enables client applications to access real-time data in a standardized manner. This opens up a variety of use cases for equipment owners to monitor and manage the performance of their infrastructure. OPC technology is widely used across various industries, including manufacturing, oil and gas, and renewable energy.

One useful open-source tool for visualizing OPC UA server hierarchies is OPC-UA Client UI. Using the UA protocol this client presents the user with the tree hierarchy of nodes with their respective values.

Spotfire® Streaming is an event processing and streaming analytics platform which is highly scalable, flexible,  and efficient at processing large volumes of data with low latency. The product includes an adapter set that integrates with OPC UA servers. This allows developers to browse node hierarchies, read and write individual node values, and create live subscriptions which produce sensor data streams.

OPC UA adapters do not work directly with OPC DA servers, and Spotfire® engineers are often asked for guidance by customers about integrating with their DA servers. Commercial software such as OPC Expert provides OPC protocol bridging, and allows a UA adapter to connect to a DA server in a fault-tolerant manner. OPC Expert even provides a limited free trial which you can try out with your own DA server.

OPC UA HDA is supported by Spotfire® Streaming allowing retrieval of historical data from UA servers. Historical data is important because it provides the basis for root cause analysis upon which rules and alerts can be developed. This data is also important when building out data science models for detecting anomalies and implementing condition-based and predictive maintenance strategies.

Spotfire®  is a full-featured analytics and dashboarding solution which closely integrates with Spotfire® Streaming. When combined, the two products integrated with an OPC UA server provides a powerful tool for analysts to monitor real-time and historical equipment performance. 

Together these products provide a canvas to build out a sophisticated monitoring and analysis platform. But to get started more quickly, Spotfire® offers fast-start, open source components called Accelerators illustrating real industrial equipment monitoring use cases. They include a bundled, open-source OPC UA server based on Eclipse Milo, sample data, rules, data science models, and real-time dashboards. 

The Data Historian Accelerator provides an easy to use interface for setting up node subscriptions. Single data nodes can be subscribed from the server tree hierarchy, and then monitored with univariate rules and calculated statistics. Additionally, there is the capability to group nodes together into virtual rows called Feature Sets. These rows can be monitored with multivariate rules, and data science models can be applied to detect anomalies and predict behavior.

The Intelligent Equipment Accelerator also shows how the OPC UA adapters can be used with more concrete use cases. There are sample multivariate rules for monitoring the performance of oil well pumps, and predicting outages before they occur based on live telemetry. There are sample data science models for monitoring the state of a power plant. A clustering model implemented using PMML classifies the emissions state, and an autoencoder model implemented using H2O detects anomalies in high frequency data streams. In both cases data is fed from a simulator through the included OPC UA server, and into the running Accelerator for monitoring and display on various real-time dashboards.

These Accelerators provide a great way to try out some of the capabilities of OPC UA servers and live streaming data. They can be downloaded right now from Spotfire® Community free of charge. Combined with evaluation licenses for Spotfire® Streaming and Spotfire®, we encourage people to try them out on their own hardware or virtualized environment. In as little as an hour you will have a fully functional OPC UA demonstration with real rules, models, and live dashboards.