DATABERG
Artificial intelligence (AI) and big data penetrates manufacturing floors in the form of robots, sensors, and machinery that make production systems faster and more efficient. Many of these technological advances in smart factories result in continuous data collection of the production systems, and big data and artificial intelligence further advance the smart manufacturing process by integrating the system and helping businesses better leverage the data collected.
Big data analytics are utilized to process the data captured and perform advanced capabilities such as analysis of non-conformance reports, enhanced security, predictive and preventative maintenance, plant load optimization, improved supply-chain management, financial risk analysis, and operational monitoring to improve overall equipment effectiveness.
Big Data also helps to integrate the previously siloed systems to give companies a clearer picture of their manufacturing processes while automating data collection and analysis throughout. The difficulties associated with data in manufacturing systems have shifted from that of being able to collect a sufficient quantity of data, to how best utilize the now vast amounts of data available in order to make better informed business decisions.
The IoT and big data integrates systems and has many analytic applications.
One application is data analysis in manufacturing, which can be utilized to improve efficiency by watching for non-conforming events in production systems. Non-Conformance report analytics are used to reduce errors and identify faulty products on the production line, before they can be released to business partners or customers. Studying these events leads to less non-conformance instances in future production.
Another strong case for big data analytics in smart manufacturing is their application in predictive maintenance processes. Systems put in place can send out repair alerts and preventative maintenance reminders to stop equipment breakdowns before they occur. Integration of sensors for condition-based monitoring can also be integrated to monitor equipment performance and health in real time, improving overall equipment effectiveness on the factory floor.
By minimizing failures and enabling better preventative maintenance practices, manufacturers see less failures, better equipment productivity and more reliability – all of which improve the overall equipment effectiveness metrics.
Big data applicability also lies in the systems involved in manufacturing such as enterprise resource planning systems, supplier relationship management systems, and product lifecycle management systems. Sales and operations planning processes can begin to be automated by using big data analytics to review historical loads, customer data, and changes to major projects which will help companies optimize their plant loading.
This strategy helps define load forecast and establish more informed manufacturing processes. IoT can also be integrated to monitor inventory, track parts as they move through supply chain, and collect data for the ERP systems. This data driven method of supply chain management helps to reduce inventory and capital requirements.
In addition to improving the manufacturing process itself, big data analytics can also be used to improve the safety of the work environment. Key Performance Indicators collected for health and safety monitor absences, injuries, and incidents that occur on the job. This data can be stored and utilized in audits and management discussions to identify gaps in the safety of a production environment, and drive some common health, safety, and environment concerns.
Industry Examples of big data use
Big Data is being utilized by manufacturing companies of all sizes and sophistications. General Electric’s New York factory is equipped with over 10,000 sensors linked to global factories that report on manufacturing processes and building condition. Employees can view the data, receive alerts of production deviations, improving quality and the overall process as it happens in real time.
McKinsey and Company assisted a pharmaceutical manufacturing company in increasing vaccine production by 50 percent by segmenting the manufacturing processes and assessing different parameters and their impact on yield. Intel used big data in the predictive analytics context and lowered the amount of quality testing, saving an expected $30 million in chip and processor manufacturing. BMW improved quality by utilizing data from sensors on prototypes to identify design weaknesses and recognize vulnerabilities before car designs hit the market, boosting their brand reputation and ensuring higher quality products.
Big Data and IoT span a variety of industries. In engineering orientated segments (such as aerospace, heavy machinery, etc.) big data analytics are used to improve factory efficiency and examine history to optimize supply chain management. In pharmaceuticals, sensors speed up R&D trials by speeding up screening processes. In the oil and gas industries, data used for preventative maintenance on the pumps that pull oil from wells can save over $50 million per day, saving money and reducing downtime. The automotive industry can improve vehicle design by using sensors to report data on speed, fuel consumption, and oil temperatures while in use. Across industries, big data analytics and the Internet of Things are disrupting the manufacturing sector and enhancing production processes and company’s ability to compete.
Increasing demand for natural resources including oil, agricultural products, minerals, gas, metals, has led to an increase in the volumes, complexity, and velocity of data that is a challenge to handle and process. Likewise, lots of data from the manufacturing industry is simply ignored. The underutilization of this extremely valuable information prevents improved quality of products, energy efficiency, reliability, and better profit margins.
Big Data application in key industries
In rapidly evolving industries, big data enables businesses to solve today’s manufacturing challenges and to gain a competitive edge. With big data and analytics, companies have got a chance to make better real-time decisions about asset usage and operations scheduling. Nowadays, data can be generated by hundreds of thousands of machines and parts, from valves to monitors equipped with sensors and wireless capabilities.
In such a way, this data helps chemical and oil/gas manufacturers optimize production levels, reduce waste, improve accuracy, and manage energy consumption. With data flowing from multiple sources, data-driven industrial companies can now gain a better understanding of supply chain operations, resulting in streamlined processes and better distribution channels (e.g., data on exchange rates let the costs be reduced and suppliers optimized). As a result, distribution within large enterprises and to hundreds of destinations can run more efficiently.
Equally, in oil and gas industry, large amounts of data that are constantly generated from oil and natural gas upstream, midstream and downstream processes can be quickly processed and analyzed to reveal new insights to prevent equipment malfunctioning and improve operational efficiency. For instance, by integrating IoT into offshore equipment, employees track and monitor lifespan and other elements that can affect production, such as wave heights, temperature, and humidity.
In like manner, once automation is introduced to remote-control systems, the personnel is removed from areas of risk, and time-critical responses are constantly generated. In other words, captured data is stored in an active database, linked to maintenance scope, and accessed through 3D representations of the platform, making it accessible to all authorized parties.
Video surveillance technology is another big manifestation of the industrial data presence. It’s crucial for manufacturing sectors to protect critical infrastructure that cover large areas. To encompass these spaces, numerous video surveillance cameras need to be deployed. Video analytics can help analyze the video streams of those cameras to provide real-time alerting, as well as operational insights for maintenance purposes.
Why is big data crucial for a company success?
Managing things centrally makes sense. All companies are scrambling to become part of Amazon. We want to make our product available on AWS. Redundant, virtualized systems in the public cloud. This is a Capex versus Op Ex versus core competency decision. Offload your IT architecture to AWS and focus on your business. Trust someone else to maintain the environment and use as a service.
Moore’s law in computing power continues to enable the handling of larger datasets, containers, and micro-services. Integration challenges are being solved. IT needs to focus on the right problem, set-up, configure, maintain, and manage disparate data sources to innovate like Tesla and Uber.
Continue to help companies take different formats – normalize, tag, and put in a repository to use and analyze. Analysis of data by computers by enriching the data by enabling lookups.
The immediate opportunity is to find effective, efficient ways to correlate use cases, and leverage common big data stores to address multiple needs. The concepts behind big data are fabulous for collecting a wide range of data, and now the challenge is to break away from legacy siloed thinking to recognize and leverage data relationship to solve the remaining hard problems that are out there today, such as complex multi-domain proactive incident prevention and multi-faceted cybersecurity threats.
Insights everywhere for everyone – not just the elite. Take big data, business intelligence, and analytics to 100% of the population. Everyone can use analytics just like everyone can read and write.
Democratization of data. Able to get useful information from data via the clould and solutions as a service (SaaS). Catalog data in NoSQL and Cloud, in addition to Hadoop so analysts can get access to all big data.
We will be collecting more data, getting actionable insights, with repeatable (automated) processes, to see near-term value from data. In the next generation, we’ll have real-time streaming and decision making. More of a real-time view.
Tremendous. Data is king. Evolution of tools with machine learning and NLP. Make it easy for people to use our infrastructure and focus on data mining. Cloud, Flash, soft edifying makes infrastructure invisible with APIs.
For risk-averse markets, how to incorporate detailed data to do more experimentation and test more hypotheses. Enable experiments to go into rapid prototyping. Don’t trigger false positives. Don’t open security issues. Take advantage of big data technology. Plan for voice of the customer, sensor data, biometric signals with changes. Don’t hard code the data feed, stay flexible. Think ahead about how our kids are the next generation and how the user interface will change.
Anyone involved realizes we’re in a boom time. Remain adaptable and flexible. Don’t be dogmatic about a particular solution. Evolution of production analytics. Predictability, scalability and how to get value out of network technologies. Embedding of knowledge to make real-time decisions and have real-time knowledge.
Convergence of software, data, and machine learning. AR/VR via mobile. Blockchain is getting less buzz but being able to share data to get an outside perspective without moving the data is a tremendous opportunity. Attributes and scores are based on data that’s secure and encrypted.
More real-time convergence of disparate data sets for real-time analytics. Machine learning. Being able to ask questions of big data and get answers back. IoT will provide real-time traffic, threats, and traffic. Make computer systems smarter. Build enterprise data hub architecture. Average users employ Hadoop to do one thing realizing the multi-tenant state of the data. Open source will continue to drive big data. Need to move to the cloud but maintain flexibility with hybrid solutions.
The distant future is the application of machine learning, AI, and automation. There’s still a way to go for this not to be science fiction. We get there by having tools make the analysis and collection of data more similar. Machine learning is currently siloed away from the analysis tools. This will need to change for machine learning to be integrated into real-time decision making.
Life sciences – struggle with large data sets. Able to predict microbiotics’ impact on humans. Learning systems across industries are building cognitive systems where all training and machine learning is based on user behavior in real-time.
Resurgence of small businesses integrating with online businesses. It’s easy to open a store online, look at the analytics and then replicate in brick and mortar. Bonobos is doing this as are craft businesses in coffee, chocolates, and bedding. Retail will be able to use the same technology developed for casinos (e.g., Space Meter) to track walk byes and drive byes. There will be more analytics involved in monitoring the performance and potential of retail stores even down to the use of shelf space.
Lower barrier to entry for majority of organizations and users, so that Big Data analytics can truly be integrated into day-to-day operations.
Automation of the generic use cases will change the perception of big data projects and how they are carried out in the next couple of years. Additionally, moving away from custom implementation to more generic implementations, which in turn can be tuned to the customers’ needs.
Better education at the grassroots level. Computer engineering degrees need to teach the concepts and provide future computer engineers with experience in big data processing. It takes a significant mind shift to move away from traditional data processing concepts, such as relational database, and to move toward big data processing concepts. Cloud providers need to keep refining their tools to make them easier and easier to use. To be fair, they are doing this. Machine learning and advanced analysis as a follow-on phase to Big Data processing is already a revolution. I believe that the longer-term impact of this technology will be politically and economically profound and is currently grossly underestimated by most people.
Summing up
Forbes predicts that digital content will have multiplied by 50 times by the year 2020, further necessitating strides in big data analysis to manage the uptake in content.
Simply collecting the data doesn’t allow manufacturers to compete, it is the transformation of data into tangible ways that will improve efficiency that will gain manufacturers an industry advantage. Sophisticated data analytics will continue to help manufacturing companies make more informed business decisions, minimize production risks, and improve their manufacturing processes by increasing efficiency, reducing waste, and improving product quality. Big data analytics is crucial in allowing businesses to compete in the ever-evolving manufacturing industry.
It goes without saying that gradual transition towards big data application may not be easy for many industries, for many of them still lack manpower and capabilities for hiring the required personnel that can handle big data. Personal and cyber security also need attention since this remains a considerable barrier in realizing the value of big data analytics. All in all good expertise and strategic mindset while using big data tools, will surely ensure success and reduce the margin of error.
