<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=1205042&amp;fmt=gif">

Having covered the theory and key principles of Industrial DataOps, it’s time to see how organizations are putting these into action in the real world, to achieve tangible improvements in the industrial value chain. This chapter explores stories from the field, including examples from oil and gas, power and utilities, and manufacturing.

The complex machinery and supply chains necessary to deliver the desired outcomes make general industrial value chains fall into the category of problems described as “systems of systems”. In these, overall challenges are affected both by the complexity of individual processes and tasks, and by their interdependence.


Overcoming these challenges successfully calls for the combination of deep niche software (such as the control system of an individual oil or gas pump) with high-level software that orchestrates and governs enterprise-wide processes (such as an enterprise asset management system).

One example would be developing a model to predict the optimal maintenance for an individual pump, where its output should trigger the automatic creation of relevant tasks in the enterprise asset management system (EAM), considering supply chain, resource, and risk interdependencies. Industry is full of such potential improvements that can be made by linking relevant, contextualized data with real-time operations.

Taken in aggregate, such fusions of technologies, data, and automated processes are central to the concept of Industry 4.0, or the Fourth Industrial Revolution. At the core of the necessary infrastructure to handle these improvements at scale lies industrial DataOps.

Effective Industrial DataOps is imperative to ensure the development, scaling, and managing of data-driven improvements across different applications.

Opportunities for Industrial DataOps

Let us first briefly discuss where the opportunities are. Industry’s core driver is production, be that the extraction and processing of oil and gas, the generation and delivery of power, or the manufacturing of goods.

Optimization of throughput is crucial, both in terms of production volume and, when the output is not a raw material, production quality. Your assets and equipment are vital in this, so naturally the maintenance of those assets is one of the main levers to increase throughput and reduce costs. The chief objective of a maintenance program is therefore to maximize uptime and minimize cost in an asset life cycle perspective. Production optimization and maintenance are both key areas ripe for data-driven improvement with an Industrial DataOps approach.

Managing field-worker efficiency is another key area of opportunity for improvement with Industrial DataOps. Often, practitioners separate equipment maintenance and asset integrity maintenance, where there will be different experts assuring the proper prioritization, risk assessment, and planning of the two.

The objective remains the same: to minimize negative production impact at lowest lifetime cost. The ultimate manifestation of the process is that some task is performed in the field by personnel, whether that task is a change of equipment oil, or the corrosion treatment of a handrail.

As well as production optimization, maintenance, and field worker efficiency, supply chain is also on the list of areas where Industrial DataOps can be transformative. We can add capital project planning and execution (in certain industries), as well as subsurface oil and gas with its exploration, drilling and wells, and reservoir domains.

The promise of Industrial DataOps is to increase the pace of deployment of data-driven improvements, including the scaling of any individual improvement across the fleet of assets, whether that asset represents equipment or larger facilities. Now let’s take a closer look at some real-world examples of data-enabled use cases and the data operations necessary to support them.

Extending the lifetime of critical assets and reducing service time



Like many manufacturers, Aarbakke was facing performance challenges caused by underperforming critical equipment: specifically, computer numerical control (CNC) machines. As a supplier to the oil and gas industry, Aarbakke must meet strict quality requirements compounded with short timelines.

Previously, service managers depended on operators to log critical issues through emails or notes. Service managers would then physically go to each individual machine and manually review the local log to see the alarms. This legacy work process was limiting response time and resulting in lower throughput and poor quality events.

Aarbakke needed a new work practice to understand their critical equipment and provide sitewide visibility for active machine alarms, historical alarms, and near real-time sensor data. The team knew that using Industrial DataOps to increase visibility would enable them to better prioritize work, adopt more efficient work processes, and gain learnings to prevent repeat poor quality events.



Aarbakke’s chosen Industrial DataOps solution was used to integrate data from the manufacturer’s source systems, contextualizing it all into a unified data model which was made accessible to all users. The data sources included both process (time series) data and events data (equipment alarms and maintenance activities).

With a unified data model incorporating all the necessary data, it became possible to quickly develop an application to make Aarbakke’s management of critical equipment more efficient. The delivered application provided a contextualized, site-wide view of all CNC machines that grouped alarms and events by assets. Service engineers could then use persona-based filters to take targeted maintenance actions.


  • 20-30% lower service cost by extending the lifetime of the machines, with a reduction in the number of breakdowns and overall downtime.
  • Faster response times to issues with organizational visibility.
  • Improved quality and increased throughput.
  • Prioritization of assets to intervene before unexpected failures.
  • Better support for maintenance workers in the field.

Supporting maintenance workers in the field



Large-scale digitalization of the manufacturing industry will require all relevant data to be accessible by users out in the field.25 However, for many manufacturing companies, data is trapped in complex, siloed systems without the ability for maintenance personnel to easily access the information they need when diagnosing equipment, performing repairs, or completing inspections. Siloed data creates inefficient work processes for maintenance personnel in their day-to-day activities, often requiring them to access multiple systems to find the information they need.


In less than two days, the right Industrial DataOps solution liberated and contextualized data from Yokogawa’s source systems, including process variables, equipment information, historical events, and instruction manuals. This short turnaround was achieved using contextualization services to automatically create relationships between process variables, equipment, and events for plant assets.

This contextualized data could then be connected to a field application designed to meet the needs of the digital worker. Accessible from tablets and mobile devices, this made real-time process data, historical data, documentation, CMMS work orders, and pictures available to maintenance personnel out in the field. Assets could now be identified simply by scanning the tag on any piece of equipment to see all relevant information.

Having this data at the tips of their fingers significantly increased the efficiency of maintenance personnel and equipped them with the information needed to perform their daily tasks.In addition, an operational digital twin was created, combining the liberated, contextualized data with a 3D model. This was built in under one hour, after capturing some 400 pictures of the Kofu plant. The contextualized real-time process and historical data was then overlaid in the 3D model, giving users a powerful visualization tool to explore the plant. These 3D models can be used by maintenance personnel to better understand how a piece of equipment fits into the overall process and see other work orders created in the same area—all from their mobile devices.


  • Maintenance personnel are able to access all of the relevant information they need when completing work out in the field. With mobile access to all the process and asset data, they can quickly diagnose errors and conduct maintenance work more efficiently, resulting in 30–55% increased worker productivity.
  • The 3D model with contextualized data saves field workers time with more efficient processes for locating equipment and planning work to be completed in the field, improving off-site planning and support.
  • The solution delivery time is less than one week, delivering value within weeksof implementation.

Enabling data-driven slug prediction to boost production



‘Slugging’ is a common production challenge in the oil and gas industry. This means a separation of the three-phase flow (gas, oil, water) in the pipeline, and the accumulation of one or more of the phases blocking the flow. The factors that influence slugging can be either transient (such as the opening or re-routing of wells) or steady-state. Therefore, simulators and real-time production data need to be combined to monitor and prevent slugging.


Aker BP’s Industrial DataOps solution enables access to thousands of live and historical time series, continuously analyzing these for pattern recognition and statistics, ensuring the necessary data quality for operational decision-making. The live production data was integrated with a data-enabled third-party application to deliver the complex simulations and models required for slug prediction.

Live operational data was fed into hybrid models with self-learning algorithms that could identify field behavior and generate predictive models to identify slugging scenarios. Aker BP was able to develop optimization models which deliver real-time, actionable insight to production engineers, to avoid slugging incidents and the associated production losses.


  • The company saw a 1% increase in production.
  • Slug handling and prediction capabilities improved significantly.
  • Production engineers gained real-time, actionable insight.
  • Operators received user-friendly decision support and algorithms that deliver early warning signs of imminent slugging.

Preventing expensive transformer failures for an energy grid operator



Transformers are some of the most expensive and critical components in a power grid. Often weighing in at more than 200 metric tons, these massive devices are situated at critical points in the grid, transferring electricity between alternating-current circuits and increasing or decreasing the voltage as necessary.

Grid operators sometimes experience transformer failure. These events can lead to power outages for consumers and production losses for power companies. In the worst-case scenario, a malfunctioning transformer can catch fire and even explode. Both repair and replacement are expensive and time-consuming.

A major grid operator, responsible for hundreds of transformers, was experiencing about one malfunction a year. The grid operator was conducting regular maintenance of the transformers, and it also invested in replacement components to ensure that power could quickly be restored in the case of an outage.

The grid operator expressed an interest in improving both of these processes. How could data help the organization identify early signs of transformer failure, and how could it optimize its spending on replacement parts?


The grid operator worked with an Industrial DataOps provider to liberate information about transformers from its source systems, including temperature, load, dissolved gas analysis, technical specifications, and inspection logs, and ingest it into a data fusion platform.

With access to all the data relevant to transformers in a single location, the development team was able to calculate a health index for every transformer in the power grid. That health index was then visualized in a dashboard, giving the grid operator’s engineers the ability to monitor the entire fleet of transformers at a glance and see which components should be prioritized for maintenance.


  • $2 million savings per year driven by failure reduction of 20-50%.
  • The health index helps the grid operator make data-driven decisions about how to plan its transformer maintenance activities.
  • Each transformer failure costs the grid operator at least $5 million. The grid operator has set a goal of reducing the chance of failures by 20-50% over the next five years, which in the short term will save the company about $2 million a year.

Enabling data-driven maintenance and performance-based service delivery



Leading drilling technology and service provider MHWirth landed a new performance-based maintenance contract with a strategic customer, for surveilling and maintaining their fleet of drilling equipment on rigs across the world.28 To meet their contractual commitments, and keep the account profitable, MHWirth needed to rethink how they used data in their maintenance decision-making. The need for taking a condition monitoring and predictive maintenance approach was a given, but there the challenges began. How could they set up their IT architecture to efficiently survey and detect anomalies from a worldwide fleet of assets? It was also crucial to ensure the necessary data governance and quality to make actionable decisions, which could potentially make or break the customer relationship.


MHWirth implemented an Industrial DataOps solution to capture and organize live equipment data. The provider set up live extractors of the necessary data, creating asset templates and the supporting data models as foundational pieces of the underlying data operations.

With all the relevant data in the system, MHWirth could use the visualization and analytics tools of their choice to create digital representations of the current state of their drilling equipment. The tools enabled MHWirth to perform predictive analysis on the data to better plan the right maintenance program.The stream of real-time data fed directly into the model, and the results were made available for other visualization tools, applications and machine learning models.

MHWirth dashboards now use both historical and real-time data to inform experts about the true condition of equipment. This is the prerequisite for ensuring maximum uptime of drilling equipment in operation, balancing appropriate maintenance with a profitable maintenance contract.


  • By bringing all industrial data together in one place and linking it automatically to an asset hierarchy, MHWirth gained a better understanding and fuller control of their industrial reality.
  • Experts analyzing the data in MHWirth’s system can rapidly identify which equipment needs service, prioritize their actions, and advise on optimal maintenance.
  • The company has now the opportunity to develop truly insightful maintenance programs, keeping maintenance costs in check, extending the lifespan of equipment, improving equipment reliability, and minimizing unplanned maintenance and downtime.
  • MHWirth now has continuously innovate on the services they provide to drilling companies, bringing condition-based maintenance services into the mix.

Scaling operational impact from visualization software



Visualization software, like Power BI, Tableau and TIBCO Spotfire, is commonplace for all companies attempting to become more data-driven.29 However, visualization software is nothing without data. Unless the data can be found, insights cannot be generated; and unless the data can be trusted, the insights are not actionable. These problems are well known in large organizations with many different data domains, data stores, and data owners.

Getting IT to provision the data can be cumbersome in itself, but just as often, the user does not even have the full picture of what data exists. When the data is found, visualized in Power BI and used to generate insight, constant assurance of data quality is necessary to create sufficient trust to utilize the data in operational decisions. All these challenges need to be overcome to scale the operational impact of visualization software.


Using the right Industrial DataOps solution, business users are easily browsing all available data and provisioning it to their visualization software of choice. Thanks to data management and data quality tooling, users can easily set up data lineage and quality monitoring. This ensures the visualization is actionable at any given time, and ripe for operational decision-making.

As the data is provisioned through the solution, users are building visualizations independent of the operational and IT systems where the data resides. Integration with AD providers enables easy sharing or restricting of data access to visualizations. Enterprises are able to become more data-driven without the need for complex infrastructure projects on the IT-architecture side and long, expensive training programs on the user side.


  • Organizations gain access to visualization of complex, large-scale data sets.
  • Large-scale data visualization becomes available to all necessary users in the organization, in a simple browser view.
  • Users gain instant access to use cases, to fuel innovation.