Analytical solutions VS Data Analytics – What are the differences?
As the industry evolved, new buzz words are being used to describe new techniques and methods to support the decision-making in the asset performance management area. Two key (and very similar) terms can lead to confusion – Data analytics and Analytical solutions. What are the main differences?
Intelligent Analytical solutions
Analytical solutions have been available in the market for decades now. Methods such as RAM analysis and Risk-Based Inspection are being utilized by the oil and gas industry for over 30 years now. The main analytical methods used in the Asset Performance Management solution are described below:
Risk Based Inspection
In 1994, the American Petroleum Institute (API) conducted a research and development project to establish a Risk-Based Inspection (RBI) methodology. The innovative RBI methodology completely changed the landscape of inspection plans from the traditional time-based to a condition-based approach. The most recent versions of the recommended practice are the “API RP 580: Risk-Based Inspection” (API, 2016) and “API RP 581: Risk-Based Inspection Methodology” (API, 2016). These methods can be used to identify and manage the risks associated with the integrity of pressure systems such as equipment and piping. Professionals with a good understanding of RBI, when applying these techniques in the industry, have achieved reduction in the frequency of inspections while ensuring the risk does not increase whilst reducing operating cost.
Reliability Centered Maintenance
Gives the ability to thoroughly analyze the effect of failures and optimize the maintenance strategy. The methodology is supported by standards such as “ISO 14224:2016: Collection and exchange of reliability and maintenance data for equipment” (ISO, 2016) and maintenance packages can be defined to go through a screening process before entering the detailed Failure Mode Effect Analysis (FMEA)or Failure Mode Effect and Criticality Analysis (FMECA), giving companies the ability to focus on what matters.
Performance ForecastingMake informed decisions by forecasting the performance of your assets based-on reliability, maintainability and operability factors. By predicting the production availability, revenue, operational expenditure, unit and maintenance utilization, you can act before risks increase to unmanageable level. This process can be supported by the “ISO 20815: Production assurance and reliability management” (ISO, 2008).
Safety Integrity Level
The substantial increase in machine automation which is part of the next industrial revolution, Industry 4.0, and the extensive connection through the Industrial Internet of Things (IIoT), big brings new challenges to the safety of industries. Safety, for one, will now depend on the correct function of electronic components or systems. SIL is a methodology which supports the analysis and design of safety systems. The approach is described in several standards such as IEC 61508 and IEC 61511.
Some of these methods can be implemented using qualitative and quantitative considerations. This provides a low-entry barrier to companies that want to start their Asset Performance Management program. As companies evolve in their process of collecting and treating data, they tend to move to a quantitative approach. The industry has acquired a lot of knowledge over the years. This is translated in the extensive list of standards which are used to support companies in their journey to excellence.
The main problem with the current implementation of these solutions is the decoupling amongst them. Some of these methods share very similar input data but still there is little interaction between their inputs and results. Existing departments which are responsible for these methods typically operate in silos. This creates major inefficiencies such as entering the exact same data twice in the software product. Furthermore, one department cannot see the benefits of implementing each one of these analyses and the company neglects the full picture.
Data Analytics
Data analytics refers to the process of collecting, cleaning and post-processing data with the goal of generating business insights. Data is extracted and categorized to identify and analyze data patterns, providing insights to the business.
There are several examples of the application of data analytics. One key example, which will support an Asset Performance Management solution, is the ability to calculate the reliability trend of components. This involves being able to fit historical data of equipment failure to an existing probabilistic curve, that will describe the failure pattern for prediction methods.
Methods involving regression analysis of historical failure data have been extensively applied in the industry. However, as it stands nowadays, current methods tend to be time-consuming mostly because of inconsistencies in the recording. This is one of the main areas where Data Analytics can help; the process of data cleaning will enable these methods to be more efficiently deployed.
With this information, the causes of these failures can be further analyzed and the company can take actions to mitigate the potential increase in the failure rate.
Combination of Analytical Solutions with Data Analytics
There is a clear distinction between Analytics and Analytical solutions:
- Analytics (or data analytics) is a new buzzword which relates to transforming data into insights;
- Analytical solution is a mathematical approach to describe a problem based on specific assumptions;
The analytics approach requires data to yield information. In some cases, it requires vast amounts of data to achieve information of statistical significance. This is not necessary when taking an analytical approach to the problem. The required input data to an analytical approach are readily available; it includes process flow diagrams, pipeline and instrumentation diagrams and process description documents. Information can be generated and acted upon when combined to a set of assumptions,.
In a nutshell, analytical models aim to represent the physics of the problem by utilizing an analytical equation. Data analytics for the most part focus on using statistical approaches to explore possible correlation between inputs and outputs. Thus, analytics require vast amounts of data and analytical solutions do not. At the early stage of operational-phase, it is not possible to run analytics because of the lack of data. One argument could be to make use of real-time data. However, this approach typically involves training a model which, again, requires initial data. The development of an industrial algorithms, which could be the starting point of analyzing data as it becomes available, is reality but worldwide data is still required to train the models.
So, analytical solutions can support decisions related to asset performance throughout the entire asset lifecycle. Data analytics is a powerful method that will provide extremely useful insights as the asset becomes mature and more data becomes available. So, they will have different impact in different stages of the project lifecycle. Combined, the two approaches offer a unique solution to increase reliability and availability.
Author: Victor Borges