As many other industries are experiencing, the world of manufacturing is also undergoing massive transformation with respect to technology and the way the business gets conducted. Knowledge graphs or as we may like to call them “Industrial Knowledge Graphs” are now increasingly seen as an integral part of a manufacturing organisation’s digitalization strategy towards smart engineering.
“Industrial Knowledge Graph at Siemens
Leveraging semantic technologies and the metaphactory platform, the Industrial Knowledge Graph has become an integral element in Siemens' strategy towards intelligent engineering and manufacturing. It powers various business use cases, including gas turbine maintenance, building automation, and internal R&D management, and generates business value in real-life applications by significantly increasing transparency, reducing redundancy, improving internal processes, and reducing the time required to find and analyse information and reach smart business decisions. “
Data structures in Manufacturing
In manufacturing entities, there are multiple data silos corresponding to the line of business or the functional areas. Though this data in multiple and distributed systems is a very important asset for the manufacturing entity, its value is sometimes limited to line of business of functional area it corresponds to. If by any means, such strategic data sources can be combined to create a larger connected structure, its value for the company would grow leaps and bounds. Distributed data in silos is bad and needs to be integrated to create a knowledge hub for the company so that it can move to next level to transformation and growth.
Looking at manufacturing, it seems that a data structure based on graph is one realistic solution to create a unified data structure that can manage data of a manufacturing entity in a connected way and creating value out of that never seen before. A graph data structure which connects everything and simulates the real world.
Google and Facebook are two shining examples of how organisations can use knowledge graph systems on scale and Large global web companies (Google, Facebook and even LinkedIn) are the best examples of knowledge graph systems. If manufacturing wants to go the google way of using knowledge graphs at scale, then they need to integrate all data stores and convert it into a graph format where everything is connected. This is an enormous exercise but since the results are outstanding it’s a must do now for manufacturing.
Industrial Knowledge graphs use cases
Let’s quickly have a look at some of the different areas where knowledge graphs can really make a difference for a manufacturing entity.
Risk assessment and mitigation
Over the last two decades, manufacturing industry in general tended to focus heavily on advanced management and production activities to increase organizational flexibility, enhance product quality and to increase innovative capability. Although manufacturing industry incorporates project management techniques within its operations (for product development and for investment), risk management is still considered an anomaly. Needless to say, that Industry projects need careful planning and constant monitoring of all potential risks. The risks may come from the financial side, stakeholder issues, investors or partners, currency fluctuations, government regulations, societal change and many other parameters. Since, the threats faced by manufacturing industries is so varied a knowledge graph will be able to help manage and mitigate risks leaps and bounds. This knowledge graph will collate all the information considering almost all the contributing factors and company specific variables. If we can augment this knowledge graph further with external data from third-parties to cover additional areas, then a manufacturing entity will have a holistic understanding of the risks that can potentially impact their operations, as well as knowledge of how to effectively mitigate those risks to protect their assets and their brand, and to keep operations running smoothly.
Manufacturing Process optimisation
Talk to any process engineer and he will say transparency is one of the most important prerequisites for the optimization of manufacturing operations. Transparency not only increases visibility, but it also forms the basis on which all key decisions are taken. In a simple language it translates to, the more information available about manufacturing processes, the more accurate decisions can be made and the easier it becomes to optimize the manufacturing process. Who else can help achieve this kind of transparency if not knowledge graph? Each machine or a transducer generates large amount of data which if integrated to the graph model describing each physical component and device in one knowledge graph allows a very user friendly and complete process monitoring. It will also help in understanding which processes to optimize, increasing yield and reducing defects in production.
Supply Chain and Logistics
The supply chain of the future, while more complex, should be less expensive and more efficient. How money is spent across the four key variables of the supply chain—energy for transportation, labour, inventory carry, and rent—is undergoing notable change. Knowledge graphs powered by graph databases are a natural solution to manage the ‘nodes and edges’ of a complex supply network. With the use of knowledge graphs in supply chain, manufacturing entities can create a single platform to gain precise information of deliveries and quality within the chain. They can also help in making audit of suppliers easy and more transparent which will have a cascading effect on quality improvement across the supply chain.
Maintenance, repair and operations (MRO)
Every manufacturing entity owns both large and small machines which are integral part of the entity. In all sectors, effective MRO (Maintenance, repair and operations) involves performing routine actions which keep devices, equipment, machinery, building infrastructure and supporting utilities in working order (known as scheduled maintenance) and prevent trouble from arising (preventive maintenance).All these are part of maintenance programs that run all along their years of operations. All events related to servicing of devices, equipment, machinery, building infrastructure and supporting utilities are well documented. If a manufacturing entity has a knowledge graph which portrays information about devices, equipment or machine models, their location, usage, last and next date of servicing etc it will allow for better maintenance cycles in terms of prediction and error detection, which yields higher availability and lower costs.
Health & Safety
Manufacturing entities face some of the strictest restrictions around health and safety of their workers and for the people who live near to their location. It is not uncommon to see failures which require their workers to perform repairs and maintenance in either dangerous or hazardous conditions. Also, in many cases, manufacturing entities have paid billions of dollars for certain accidents at their sites which caused issues to people living nearby. A knowledge graph can be a very useful tool for manufacturing entities. A Knowledge Graph which has a unified data structure comprising of failure data, operational hazards data, as well as safety and compliance rules related to previous accidents and causes, at a single place can be a remarkable tool to drive process, maintenance, and safety improvements throughout operations and facilities.
Integrated Knowledge graph for Plant Monitoring
Once we have models for machines and processes, the next step to have the entire plant integrated into a knowledge graph. This can be done both at production stages and even earlier when the factory is under construction. The knowledge graph thus can be the great tool for getting quick insights into dependencies and will allow a comprehensive supervision and monitoring.
For example, the questions, “Which machine in assembly line is experiencing problems?” and “Which batch of raw material will arrive late?” can be answered much more accurately by looking at the knowledge graph as it showcases how the individual components in the entity are connected and impact one another.
Graph of graphs: Enterprise Wide knowledge graph
The graph use cases mentioned above bring fantastic results to the line of business they belong to. However, knowledge graphs specific to line of business are still in silos. To experience and use the power of graphs all silos need to be eliminated and there is a need to build one graph for the enterprise which will be graph of graphs and will truly work as enterprise wide knowledge graph.
By integrating an enterprise’s multiple knowledge graphs into a single Knowledge Graph platform, an enterprise can manage itself well and can support a much wider and deeper range of products and services.
What I have described above is, the tip of the iceberg. Knowledge graphs in manufacturing can really help an organisation to become a knowledge driven digital entity.
For each of the use cases described above, we see challenges when an organisation moves towards knowledge graphs driven organisation. To address those challenges successfully, get technical solutions and get insights into the generated business value, its extremely important to work with an expert in this area.