How manufacturers can unlock the value of big data to boost efficiency and resilience
Importance or big Dates in Industries
Big data has been a constant talking point in the IT world and now manufacturing industries are also racing to leverage it for better decision-making. The industrial sector is realizing the potential impact of big data analytics to improve complex plant operations and R&D activities.
Whether it is monitoring the shop floor operations, managing the supply chains, or understanding customer preferences, big data analytics has a potential to tremendously impact the industrial arena. Industries of all types are either planning or actively thinking to leverage data and analytics for their internal processes. Besides the evident benefits like improved insights and streamlined plant operations, big data can also aid in optimizing critical business metrics.
Areas of Impact†
- Revenue and Profits: Before deciding upon investing in a new technology, the Technology Executives and Research & Development Heads of the industrial enterprises need to be sure if it is a good investment for a long-term revenue generation. For example, if an OEM has only 10% of its field devices connected to software systems, should it invest in adding next generation communication interface like Ethernet APL to its field devices or should it rather consider edge devices for outlier use cases? With deep insights into industry use cases, customer behavior, industrial product ecosystem and features, a rational decision can be made to derive at a nuanced decision.
- CapEx and OpEx: Revenue and profits cannot be separated from CapEx and OpEx that directly impact the plant operations and business decisions in the long run. Suppose a plant asset is showing frequent downtimes and hence reduced productivity and the plant engineer needs to know the root-cause of the failures and implement measures to identify these failures beforehand. The details can be pulled from the asset performance logs and using analytics this huge amount of data can be processed to find the frequency of downtimes, and accordingly, steps can be taken to prevent the episodes.
- Multi-dimensional Data Analysis: The parameters to measure the success of plant operations can vary for different plants. The analysis of data generated at every level of plant architecture equips the management with deeper understanding of various parameters that can be controlled to ensure efficient shop floor operations. Reasons of wastage, factors affecting efficiency, realizability, security, turnaround time, etc. These insights allow the manufacturers to save a lot of cost directly and indirectly.
To gain a competitive edge with the help of data, industries need to have an efficient data strategy in place. How can that be achieved?
Ideal Data Strategy for Industries in the Face of Digital Transformation
Even if you have big data, it does not really mean that the game is won. Gathering data is merely the first step in the roadmap towards digitalization. In IIoT-centric world every system, device, or equipment generates data. These volumes of data can easily become too much to handle. Monitoring no more than ten plant assets can generate up to 10 million records per day. However, not all this data is needed to address business objectives, hence considered as data noise.
It is possible to not have the clarity on which data points have value. Any industry looking forward to use ‘data’ as the key weapon, needs to associate it with a business value. Some of the questions that may help accomplish it are:
- What is the business objective(s)?
- What data is needed to solve this business objective?
- Is the data available or needs to be acquired?
- Can this data be reused for other business purposes?
- What is the roadmap to achieve this business objective?
Data that is directly or indirectly associated with a business objective should be analyzed and only then meaningful outcomes can be expected.
Managing Big Data with Technologies
The vision of achieving promising business value with the help of big data can be achieved only if the industrial enterprises have an information foundation that supports the rapidly growing volume of data. Some of the top trending technologies that manufacturers across the globe are preferring are Hadoop, MapReduce, NoSQL engines, Tableau, PowerBI, SAP HANA, etc.
Challenges in Implementing Big Data
Despite the data strategy and big data infrastructure in place, there are certain factors that limit the manufacturers from unleashing the full potential of big data. These factors are:
- Lack of Common Context: Manufacturers will have evolving needs to add new equipment/devices to the existing infrastructure. The more the number of devices, the harder it is to integrate and coordinate all of them. Due to heterogeneous make, logics, and languages, their data are not in sync. Imagine as per your algorithm, ‘1’ means “the temperature is normal”, but one of your sensors returning value ‘1’ which means “the temperature is high”. It increasingly becomes difficult to handle when the number of connections rise.
- Infrastructure Not Suitable to Handle Big Data: It not just about gathering big data, but also about all the supportive systems that make the processing, storing, and analyzing of this data possible. Is the network capable enough to stream such amount of data without frequent downtimes? Is the database efficient enough to store this huge amount of unstructured data? Will the visualization platform be able to represent the varied datasets with veracity? Most manufacturers invest in big data technologies without analyzing these situations.
- Data Not Impacting the Decisions: Once the data is gathered, the next step must be to make informed decisions. However, industrial OEMs usually are unsure of how this data can have an impact on the business. The multi-dimensional data can surely tell you that there is a potential problem, but even with all that data in hand the decisions still get made on “gut feeling”. Most of the times, OEMs implement big data analytics to analyze historic events to backtrack and find the patterns. However, this data is not necessarily reflective or indicative of the real-time shop floor scenarios. As a result, critical real-time decisions cannot be confidently made based on the retrospective analytics. This prevents the OEMs from whole-heartedly embracing big data initiatives.
- Legacy Systems Unable to Incorporate Big Data Capabilities: Most of the equipment or devices in any plant have a lifespan of 10 to 20 years. This timeframe is substantial if you think about it from an evolution point of view. While some systems are legacy, others do not have much connectivity options. Therefore, it becomes a real pain-point to incorporate IIoT capabilities in such systems. As a result, achieving the value of big data is only possible for a part of the ecosystem that falls in line with the modern IIoT expectation.
Industrial operations are complex. In the shop floor, where downtimes and even small mistakes can be expensive, data becomes an integral part of the plant operations. Data is the microscope to reach the digital corners of an industrial enterprise and find out what is working, what is not, and why. Decision makers in the industrial arena can use big data to control costs, optimize or alter the operations, formulate effective strategies, and predict the damages or failures before the red flags. However, big data isn’t a magic wand and needs due diligence, thinking, planning and proper execution to reap the benefits.
Views expressed above are the author’s own.
END OF ARTICLE