Analytics for Engineering Asset Management

"Asset management can be de?ned as the systematic and coordinated activities and practices through which an organization optimally and sustainably manages its assets and asset systems, their associated performance, risks and expenditures over their life cycles for the purpose of achieving its organizational strategic plan" PAS55:2008

The Business Challenge for EAM

In today’s business climate, asset intensive industries (Rail, Oil and Gas, Renewable Energy, Manufacturing, etc...) are faced with global competition, increased pressure from stakeholders to improve performance and reduce cost, strict regulations in terms of Health, Safety and Environment and an imperative demand for excellence and world class performance. Although these industries are different in the way they generate revenue, the way they conduct business and their regulatory and organisational aspects, they all share a unique trait; Asset intensive industries invest heavily in their physical assets. Whether its production critical assets, safety critical assets or other supporting physical assets, organisations need to make sure their assets are efficiently utilised, able to perform their production goals, produce the required quality standards and achieve the appropriate level for safe operations.

Asset Management is a strategic approach to the optimal allocation of resources for the management, operation, maintenance, and preservation of asset infrastructure. The way in which organizations collect, store and analyse data has evolved along with advances in technology, such as mobile computing, advanced sensors, distributed databases, and spatial technologies. These technologies have enabled data collection and integration procedures necessary to support the comprehensive analyses and evaluation processes needed for Asset Management. However, in many cases, the data collection activities have not been designed specifically to support the decision processes inherent in asset management, or the technologies and research to leverage value from data in the asset management domain are lagging behind other sectors. The use of the aforementioned technologies has led organizations to collect very large amounts of data that have not always been useful or necessary for supporting decision making processes; Moreover, they created vast databases that made separating the signal from the noise a challenging task. Although organizations have placed a large emphasis on collecting and integrating data, little effort has been placed on linking the data collection to the organization’s decision making processes.

A recent Study titled “From Overload to Impact: An Industry Scorecard on Big Data Business Challenges” surveyed 333 US and Canadian executives from the oil and gas industry as well as other industries. It found that 74% organisations are collecting more business data today than 2 years ago, with 96% increase in collected/managed business data in the last two years. It also reported that executives were most frustrated in regards to:

1. Not having the right systems to gather the data they need 38%.
2. Access to information, need to rely on IT 36%.
3. Systems are not designed to meet the needs of the industry 29%.
4. Can’t make sense of the information provided and unable to transform it into actionable insights 25%.
5. Information is no longer timely by the time they reach stakeholders.

Executives were asked to highlight the most needed changes to improve their data strategy:

1. 43% Greater ability to translate information.
2. 38% Improved tools to collect more accurate information.
3. 36% Customised applications to meet their business demands.
4. 36% Direct access to relevant information and insights for business managers.

A key finding form the study points that Big Data is key to revenue growth, 93% of the executives believe they are losing revenue at an average rate of 14% annually as a result of having to make decisions not having access to the right information in the right time.

The promise of Big Data Analytics

Advanced Big Data analytics engines can combine data across disparate systems in the organizations (information silos) to enable the collection, processing, management, analysis of the data and visualization of the results to empower decision making and automation interfaces. The essence of the solution is finding the hidden relations and patterns in the data to give better understanding of the problem that one is interested in. Predictive analytics includes a variety of statistical and machine learning techniques (e.g. Regression models, time series models, Bayesian networks, fuzzy logic, genetic algorithms, neural networks, etc...) the algorithms analyze current and historical data to model the business, provide insights and make predictions about the future. Predictive analytics frameworks can handle large volume of varied business related data transforming them into actionable knowledge which are futuristic and probabilistic in nature. 

In asset intensive industries, effective decision making across all the levels in the organization becomes much easier with better availability of informative insights about the business, these insights becomes available by choosing and following applicable predictive analytics solution methodology, adopting an open data strategy with the right choice of algorithms applied on the data coming from various sources related to the business. Some examples can be equipment temperature, pressure, vibration variations in the field, load fluctuations in the electricity network impacting the transformer failures, well monitoring data, engineering reports, inspections, maintenance reports, audits, etc...

Modern Predictive analytics engines collects data from various sources, cleanse and transform the data to enable applying advanced algorithms and statistical analysis techniques to extract informative correlations from historical and live data feeds and use these correlations to generate model(s) about the data to represent patterns and predict the future trends and behaviors. Big data analytics can be used with different types of variables whether it’s descriptive (why it happened?), Explanatory (what is the current status?), Predictive (what is likely to happen?) or Prescriptive (What to do next?).

Correlations can be derived for critical equipment(s) for identifying suspected causes of failures before the actual failure occurs, make predictions about the commercial viability of exploration prospects by integrating data from seismic surveys with historical production data from fields with similar characteristics, enhance drilling and production by making available information about similar fields and wells. It is important to note, however, that the accuracy and usability of results will depend greatly on the level of data granularity and the quality of input data available for analysis as well as the applicability of the technologies used. Predictions are usually provided with a probability distribution and various outcomes and actions are usually derived along with the results to improve the prediction accuracy.

How Big Data can transform EAM?

Predictive analytics techniques can effectively measure and forecast different asset parameters by analysing historical asset data from multiple silo’ed set of processes, applications and data, creating models to identify the root causes for key events in the data. Using these models, events can be predicted for future time periods with the help of advanced forecasting techniques, the key factors impacting the efficiency of operations of the identified assets and then actions can be recommended to address issues before they arise. Analytics can optimise asset management through optimized asset oriented business process and proactive intervention in operations. Key asset management issues that big data analytics can address:

• Improved asset utilisation (how efficiently assets are utilised to generate revenue). Asset utilization can be measured and the information can then be visualised and presented to stake holders to inform them about asset utilisation. Further, by applying analytics on live data feeds the information can inform the operational staff with the best course of action in advance to address the issues before they cause interruptions in production. Identifying these correlations and can lead to reducing Non-Productive Time in operations.
• Decreases asset failures (resulting from aging infrastructure, poor asset management and operations and improper maintenance of assets). Asset failures can be predicted and prevented from occurring before they arise by identifying probable causes of failures and initiate the right type of maintenance actions well in advance with the right kind of skilled people. This is definitely an improvement over the traditional time and frequency based preventive maintenance programs.
• Increase asset availability and reliability. Increased availability and reliability is key objective of asset management professionals as these assets are costly to operate and maintain. Increasing the availability can be achieved by avoiding unnecessary routine maintenance and extending time between maintenance cycles and perform maintenance based on monitoring equipment conditions using field sensors.
• Reducing asset down time (due to a forced outage or unplanned maintenance). Analytics can schedule and priorities maintenance based on mean time before failure for safety critical elements.
• Better informed asset risk. Analytics can helps to predict the probability of major hazards and low probability high impact events, asset risk can be visualised in real time by using real time data from field sensors, action to mitigate or reduce the risk can be predicted and communicated to the right people to help addressing the challenge of critical equipment performance, life cycle, integrity and safety.

Conclusion:

Organisations need to invest in the research and development of an Integrated, Intelligent, analytics framework for asset management in asset intensive industries to leverage the data for the planning, controlling and execution of the operations and maintenance (O&M) of industrial assets in an effective, safe, environmentally aware and economically efficient way.

With the use of intelligent notifications apps, visualizations that are easy to interpret by users and potentially natural language generations techniques "once the technology has matured enough" end users can reap the benefits of analytics. The goal is to provide users "an engineer in the field, a decision maker in the office or an executive at the board level" with actionable insights to quickly understand what is happening or what is about to happen, the root cause behind the problem or the key driver for a potential improvements "analytics is not all bad news", the trends that lead to the events and most importantly suggested actions to prevent or mitigate a failure and actions to improve and gain value.

In order for Analytics projects to really provide the true potential from data requires a holistic collaborative approach with SMEs, Data Scientist and AI researchers in order to apply the most appropriate analytics techniques to answer the important questions faced by engineers and decision makers. Analytics for engineering asset management combines engineering concepts with computational intelligence principles with sound business practices to support decision making at the strategic, tactical and operational levels.

This is the third article in a series of articles focused on the potential of analytics to transform enterprise asset management. My next article will focus on the maintenance management function within EAM. For the first and second article please refer to Analytics Driven Enterprise Asset Management and Analytics, a new dawn for the Oil and Gas industry.