It is a cold winter morning in the plant. On the request of the top management, the plant head has called all his divisional heads for a meeting with a consultant of a renowned firm in the city. The consultant meticulously dressed in his business formals gently opens his laptop and starts the presentation. The moment he talks about cyber-physical systems and horizontal and vertical connectivity, there is a long silence followed by hushed murmurs. Wait a minute, aren’t we doing it already? I mean what’s so new about Industry 4.0 or Digital Enterprise? Have we not being collecting data for control and automation? SCADA has been around for decades! Don’t we have mathematical models for our processes and plants? Don’t we have our ERP systems that trace all our business transactions? Then why invest on another technology? What’s the big deal about going digital? These are questions that baffle the manufacturing staff.
And then the consultant shows slides after slides on how data analytics and statistical modelling can reduce costs and lead times. The quality head shakes his head disappointedly. He has been working with this kind of data for almost a decade now since he implemented Six Sigma and Design of Experiments. It seems as if the age-old technologies have been repackaged using some new terminology; old wine in the new bottle! So why should we take it seriously anyway?
Breaking the mutual exclusion between analysis and control
There have been two major disciplines in the manufacturing environment – Analysis and Control. The analysis discipline is led by the QA teams that analyze data to find gaps in the processes and run continuous improvement programs for manufacturing excellence. The Control discipline would be focused on real-time monitoring of the process and controlling it around the given set-point. So far, the two disciplines work in different contexts and time frames. The control discipline works in real-time with localized focus on one workstation or one process or one line. The analysis discipline works across the lines but not in real-time.
Typical Manufacturing IT and MES systems enable recording of information from production floor in central dashboards. This information is diagnostic in nature with little real-time prognosis application. This means that these systems can only ‘record’ an error rather than using this information to compensate it in the later stages or reject the part early on to save further processing costs. For example, until now, the material flow was disconnected from the information flow. The information that we have been capturing and analyzing was essentially from the past. Whatever insights we have gained from one part cannot be applied to another one because it has its own distinct variations in its characteristics and machine and process parameters. Hence, any information that we have captured and analyzed for one part loses its context for another part. This means the knowledge that we are generating is generic rather than contextual.
In a responsive enterprise, the information and material needs to be so tightly coupled that the information flows along with the material. Hence, the insight gained from a part is instantaneous and contextual because it can be used on the same part the very next moment. If we continuously acquire every parameter of the material, assets and the process at every point in time, we can not only predict the output, we can also control it. For a smarter manufacturing process, the systems have to be prescriptive rather than diagnostic.
Digitization allows us to extract the implicit knowledge from physical assets like material, machines and plants and human assets and express them in the digital domain with other business processes. This digital transformation interconnects all aspects of the business – material, manufacturing and business process into one integrated digital representation enabling us to represent and validate our enterprise systems and processes directly in the information domain. This would require data-driven manufacturing intelligence systems installed right on top of control systems that collect data from across the line and run intelligence algorithms to prescribe suitable actions. Some of the use-cases for such intelligent systems are
- Error compensation of defective parts
Suppose an engine block has a negative deviation in its bore diameter. Can it be compensated somewhere down the line during the machining? Could piston be made smaller to fit into this block without compromising on the performance? If yes, the system will implement these compensations at respective station when the particular component reaches it.
- Transfer, reuse and recycling
Suppose a component got wrongly assembled or machined for the given variant. Could it be used in another variant AS-IS or after re-machining? Could it be used in another sub-system? If yes, it will be sent to the respective line or stored in WIP for re-use. If it has developed a defect that can neither be compensated nor reused, it will be rejected early to save additional costs.
- Production responsiveness simulation
What would happen if the raw material costs surge by 20%? What would happen if demand drops by 15%? How should the manufacturing process be adjusted to compensate this change? Manufacturing intelligence system will simulate and emulate these business scenarios to improve organizational responsiveness for sudden changes in the business environment.
- Technology/Product innovation
What happens when aggregator comes between you and your customers and erodes your margin? What happens if technology changes or product becomes irrelevant to your current customer base? What happens if the supplier of one of the key component decides to withdraw it? The system will emulate the impact of changing components, technology, features, design or the number of variants on the costs, quality and lead times.
Re-building the competitive advantage
Traditionally, there have been three primary areas for building competitive advantages – Operational Excellence, Product leadership and customer intimacy. A company was expected to excel in one of these three and perform acceptably in others. Hence, the majority of corporate projects were focused on strengthening the competencies around one or maximum two competitive advantage. However, recently the customers have become much more demanding and require companies to perform equally on all three counts. This means companies need to work on their strengths and weaknesses simultaneously and even then it could end up facing a crisis situation due to volatility and competition in the global markets. And when an organization undergoes a crisis, not many leaders are prepared and geared up to lead in the face of an imminent threat. And this is why the internal leadership crisis hurts organizations more often rather than the external factors.
Building the ability to handle such difficult scenarios is not easy. It requires continuous training and practice to develop once survival instincts. A martial artist trains his mind and body by emulating different threat scenarios and practicing his response till he can fight instinctively. This happens because these skills get ingrained in his reflexes and muscle memory. In a similar way, even an organization needs to continuously envisage various threat scenarios and emulate them to build instinctive responsiveness to handle crisis. These organizational instincts can only be developed through a firm commitment and tenacity to build the necessary capabilities and infrastructure and deploy the right resources. If an organization rigorously works its survival instincts, it will not only sail through downturns, it might even use it to build a sustained competitive advantage.