The Symbiocen Age (Part 3)

The Symbiocen Age (Part 3)

This is the third, last part of a three-part series on the new Symbiocene age.

Digital Pangaea.  The previous article expounded on the four tectonic plates making up the Symbiocene age, which are the four pillar upon which the actualization of the age will be realized. In this article, the realization is discussed – the digital industrial transformation.

The quiet revolution.  The internet changed humanity and has become so pervasive that it seems invisible.  Yet, it is utterly necessary to the world, on par with electricity, indoor plumbing and the rule of law.  Surprisingly, one late comer to the benefit buffet table has been the industrial realm.  That is not to say that industrial companies are disconnected from the internet; since everyone with a desk stares at a computer and by extension, the internet.  That access is utterly dissociated from the true potential of the digital world but is slowly changing.  A profoundly disruptive transformation is afoot and unfolding far and wide under the moniker industrial internet of things – or IIoT for the short. 

The IIoT can be summed up in one sentence: machines dealing with machines independently of humans.  The machine could be the simplest on-off switch or a monitoring device.  It could be a sensor or a chemical treatment.  It could be a Bluetooth transmitter or an entire control network.  Whatever its nature, its size or its function, the machine is connected to the rest of the world’ industrial network, exchanging information continuously, in fantastic volumes.  The transactions occur automatically and autonomously, without any interventions by humans.  Five billion devices are estimated to be connected in this manner right now, a number that is expected to reach thirty billion within a decade, with a $10 trillion windfall in profits and cost savings for their owners.

The autonomous paradigm.  The digital industrial is much more than untold widgets connected together over the internet.  It is a transformative paradigm shift where the machine network exists as a complex living organism.  The analogy is closer to reality than lyricism suggests.  Take the human body for instance.  It is a fantastically complex chemical factory.  It is autonomous, self-correcting, self-healing, self-aware, and environment-analyzing.  It is chemically powered, centrally controlled, and de-centralized in actions.  It is incredibly connected through a network comprising trillions of synaptic connections.  Furthermore, it is self-replicating and able to manufacture its replacement parts on its own.  The human body does all of this, and more, with only five essential inputs: oxygen, water, nutrients, DNA and sleep.  Five inputs that have given us you, the reader, Einstein, Raphael and Mozart.   

Consider now a traditional factory that produces a physical product (like a cell phone or a tire).  Traditionally, each step in the production sequence has not needed to be operated through a sophisticated digital framework.  But classicism has a cost: extensive human interventions.  In a traditional setting, manufacturing operations require lots of humans making lots of decision continually to make sure that the steps are in sync, in the right quantities, and to the correct specifications.  What might a digital industrial transformation look like?  Something like figure 1 (derived from a tire fabrication setup).  Each node is a network unto itself that is embedded into the bigger manufacturing network for the tire operation.  Each nodal network runs is independent of the others but remains in constant contact with the other nodal networks through the information transaction that transit across the overall web.  The connectivity between the nodal networks makes the entire operation digital.  But the autonomous nature of each node, interacting in real time with all other nodes, is what makes the whole thing digital industrial.

Structure of the information node.  The nodal representation of figure 1 is inherently fractal.  That is, the structure is the same whether the node is simple device like a sensor or a valve, a system such as a metering station, an installation like a back-up power system, a plant, or the network proper joining them together.  The structure is shown in figure 2.  It applies to machines, to humans, to processes and to information flows, each one falling under the general descriptor “node”.  The node can be static (bolted to a wall for example) or mobile (an inspector travelling to perform a calibration check).  This kinematic state belongs to the first level of the structure called the presence.  The structure is comprised of seven layers through which the embodiment of a node is realized and the information arising out of it flows, outward. 

The inner layer is the presence layer.  It is the physical manifestation of the purpose of the device.  The presence layer includes the physical, algorithmic and inner control functions associated with the purpose.  It also includes the node’s interactions with its immediate nodal neighbors.

The deployment layer comes next.  It surrounds the presence layer by encapsulating the physical and informational transactions between the node and the nodes that serve to control it externally.  Control is exerted through signals that arrive and leave the node, to govern the node’s normal and abnormal operations.  Control is also exerted by the physical inputs and outputs required by the node’s actions, such as power supply, cooling fluids, effluents and waste products, emissions and noise, lubrication, etc.  The ensemble of signals and physical inputs/outputs goes under the moniker “deployment flows”.

The third layer is called Autonomy.  The autonomy layer is where the digital industrial paradigm appears first, by augmenting the node with the capability to self-assessment its performance, in real time.  The autonomy layer renders the node aware of its presence and its effect on its nodal neighbors.  The capability can include such functions as monitoring the performance limits; self-regulate when operational exceedances or variations are detected; determine the adequacy of deployment flows; re-route commands to bypass operating problems (backup & redundant capacity activation; and reduce performance to maintain system integrity. 

Next comes the visibility layer.  This layer takes the monitored data produced by the lower layers and performs comprehensive assessments of performance, operability and reliability metrics. The layer also produces probabilistic predictions on the root causes of metric excursions beyond their design ranges.  The layer is heavily reliant on algorithms (A.I. or others).  The layer also packages the outgoing information streams to be sent to the next layer (what data is to be sent where, in what format, at what frequency and on what feedback basis).

The operability layer is next.  This is the first layer beyond the node proper.  The layer receives the information streams from individual nodes’ visibility layers.  The streams are captured in a live database.  The information is aggregated by groups of related nodes and subjected to additional analyses for presentation, visualization (including virtual reality (VR) and augmented reality (AR)), and archiving.  This layer effectively introduces the big data into the digital industrial’s framework.  Note that there may be more than this layer may be constituted of two more layers, depending on the analytical strategy selected.  For example, each external node in figure 1 represents a supply chain partner (independent or not), set up with its own nodal structure and supplying the results of its metadata analyses to the main operator.

The high-level data mining and crunching occurs next, at the actualization layer.  The layer operates on data and meta-data sets alike via data mining, statistical analyses, deep learning neural networks and other A.I.-type algorithms.  Global operating, economic and maintainability metrics are derived.  Undercurrents, trends and hidden correlations are uncovered.  Simulations of what-if scenarios and their outcomes are performed.  Intervention strategies are formulated, weighed and estimated.

The final layer is called, appropriately, management.  The knowledge sets and intelligence information produced by the actualization layer are evaluated by managers and executives.  Decisions are made as to what must be changed, fixed, replaced, altered, abandoned or re-designed (possibly with the help of additional A.I. tools).  Action plans are quantified, implementation timelines are chosen, and the continuous improvement process is closed.

The point.  The initial picture that emanates from figure 2 is one of information dominance.  The initial impression is certainly one that puts all the emphasis on the generation and manipulation of computer-based data outside the reach of human interactions.  While the impression is correct in light of the preponderance of data to the node structure, the underlying motivation is not about data per se, but about valunomy of decision-making.  Companies have hitherto been captive to a reactive process in matters of operational effectiveness, owing entirely to the lack of actionable feedback from their daily operations.  There simply is too much information being produced at any given time to be captured in real time and subjected to analysis also in real time.  So, companies have had to triage these mountains of information to whittle them down to manageable bits captured in some ersatz real-time fashion.  Thus were born production reports, material inventories, productivity data and time sheets.  However, real-time analysis of those record sets remains for the most part impossible. 

The point of the digital industrial is to maximize real-time optimization.

Optimization in this context takes on a specific meaning: anticipatory management of potential issues that threaten profitability.  Hence, the digital industrial is, above all, a commercial paradigm that rests on the principle that a firm’s profitability can be maximized at all times through pro-active management.  The entire argument rests on the key word: pro-active.  All data and records are by nature passive.  They just sit there until some action results pursuant to them.  If an unwanted condition is the progenitor of the data, any corrective action will necessarily be after the fact and reactive.  Any negative economic impact has already been incurred.  The best that can be done is to limit the value of this extraneous cost.  In another scenario, the incoming data may be indicative of a potential negative trend (say, a tool is showing signs of excessive wear).  Taking the appropriate measures before the trend becomes a fact illustrates a pro-active action.  The ability to do so is usually dependent on a short window of opportunity, among other production considerations.  It is entirely possible to know what to do ahead of time but be unable to do (the replacement tool is not available for example).  Hence, pro-active management is often more difficult to exercise than its reactive cousin. 

Greater benefits still accrue to the digital industrial organization.  The authority delegation that is built into the digital framework lays down the shortest bridge between the decision-making process and the conditions most impacted by the decision.  The machine intervention can be instantaneous and autonomous (i.e., without human inputs), including the creation of replacement parts via additive manufacturing systems (discussed earlier in the chapter under the A.I. header).   The net effect to the bottom line of the entire operation is a tangible cost reduction coupled to enhanced commercial effectiveness. 

  • Asset reliability is increased across all nodal scales.
  • Production throughput is maintained or optimized to the conditions on the ground.
  • The digital industrial network provides a pervasive wireless connectivity to everything and everyone, everywhere at all times.   Anticipatory management by machines or by humans becomes the de facto modus operandi.
  • The connectivity permits remote monitoring in real-time at all nodal scales, which in turns improves security, compliance (regulatory, environmental, operational).  Monitoring is performed through node presence data, smart video and sound, drone surveillance, and satellite tracking.  A.I. applications are applied to perform image recognition, diagnostic predictions, and situation analyses in real time, autonomously.
  • The connectivity supplies the power and the engine to enact constant predictions on equipment failures, leakages, downtime, bottlenecks and other degraded conditions.
  • Constant predictions feed the data required by asset integrity systems that manage, autonomously, maintenance programs in an anticipatory fashion, through real-time analytics.
  • Mobile nodes (humans, vehicles, material handling, etc.) are equally monitored in real-time, throughout the wireless network, for fleet efficiency, deployment optimization, problem detection, and geospatial tagging.
  • Personnel safety, availability and productivity can be measured through wireless real-time tracking, video analytics and autonomous incident assessments.
  • Global metrics can be compiled from aggregated nodal data and calculated intelligence information.  Hidden trends, silent underlying currents, and optimization correlations can be inferred through plenipotentiary A.I. inference systems.

The combined effects of these benefits yield an order of magnitude improvement in overall operational effectiveness, cost valunomy, personnel optimization, and regulatory compliance.  In other words, improved commercial performance and capital investment returns.