The School of Hard NOCs

Re-Invisioning the Role and Goals for Monitoring and Operations Centers

Centralized facilities to monitor and control various assets, often called Network Operating Centers (NOCs) or Remote Operating Centers (ROCs) provide critical 24/7 control and management of a vast array of assets. Given increasing volatility in markets, weather, and market pressures, today’s value for a NOC is materially greater than monitoring asset health and managing identification and prioritization of control, operations, or routine maintenance — to proactive support for “connecting” decisions/actions to requisite outcomes driven by market, fiduciary, and risk postures! But with increased importance of actions, frequently NOCs struggle to deploy OT/IT solutions that can address the full range of needs. Seemingly, there is a never- ending stream of data available – from multiple disparate sources – to analyze and make sense of; the greater the population of assets or greater number of issues/actions to be contemplated, and more work required – by humans or AI systems – to generate insights and actions at increasingly higher levels of complexity and intensity. The temptation is to double down on proven tools and methods – to leverage advanced models and AI frameworks to assess state-of-condition, identify potential areas of concern, and characterize source/criticality of the issue. BUT, such systems tend to be inherently disconnected from market or business contexts, often leaving the quantification of an increasingly important factor – criticality – to the conjecture of the system expert, or via math based on some set of assumptions. If we flip the script, so to speak, and break the paradigm by examining criticality of asset features based on business and markets, we can dramatically alter workflows, focus AI horsepower on exposing higher business value! Our goal is to enable NOC as the “brain” of systems capable of spanning near real-time, tactical and strategic decision realms and assets under management and to efficiently bring context needed to optimize decisions. Let’s begin with tools and methods to truly characterize market and its nuances: seasonal, asset-mix, nature of participation, demand profiles — to clarify precisely where performance/reliability (or lack of it) is most costly or, positively, where there may be significant rewards for committing assets toward bright opportunities. Outcome-Driven: Dynamic assessment or quantification of forward positions for asset to deliver energy or other services – derived from market insights – allows one to understand impacts of performance or maintenance actions and their timing on overall ROI. Focused: Matching areas of opportunity to concern via AI provides critical focus on areas where attention is most warranted. In other words, assure that the efforts expended by experts and traditional AI/diagnostics are naturally prioritized based on our understanding of criticality, ROI, and asset value Opportunity Aware: the ability to influence the magnitude of opportunity? Better performance, efficiency, capacity, assurance of asset starts, reliability/availability, … against a dynamic ROI model that takes costs to support additional capabilities. So back to the focus of this article … the school of hard NOC’s … to take our lessons-learned conventional tools and methods but, to flip the script, to drive additional value via reshaping the nature of the analytics stack for NOCs; in essence, to be more agile, responsive to market while retaining and building on the immense and valuable knowledge and analytics systems.


Ground truth is critical! Characterization of asset-based outcomes (and risks) associated with state-of-condition, actual weather conditions, alternative operational regimes markets … continues to drive our understanding of asset condition, risks, performance and reliability remedies. Leveraging the plethora of sources of data available from combination of sensor, SCADA, environmental (air/water) management systems, and similar Industrial Internet of Things (IIoT) sources. There is also substantial data available from digital sources that addresses the environment the asset is participating (market interfaces, weather and weather forecasting systems, other exchanges (fuel, emission credits, …). Value generation through structuring/organizing the data and adding additional, valuable context. o Conventional monitoring and diagnostics platforms and digital twins are an ideal source of additional intelligence about the state of condition or performance of an asset or systems. o Asset health monitoring (vibration, oil monitoring, air effluent monitoring systems) also feature specialized analytics to translate raw data to key factors of interest: efficiency, life expended/remaining life, capacity, emission rates, etc. o Fully leveraging the rich legacy and evolving toolbox from original equipment manufacturers (OEMs) and others that focus on providing additional intelligence: monitoring/diagnostics, performance packages, and add-ons to Computer Maintenance Management Systems (CMMS), Enterprise Asset Management (EAM), Computer Aided Facility Management (CAFM), Integrated Workplace Management Systems (IWMS), and Financial Information Systems (FIS). New features and capabilities we add are focused on developing and analyzing a myriad of forward views driven by differing sets of assumptions and commissioned via robust simulation space to examine potential decisions or actions or policies; we leverage scenario exploration and optimization frameworks to identify the most profitable or set of actions – operations, utilization of assets, nature of investment, timing – to consider: Different objectives that will focus on different time dimensions and different resources yet all service the greater mission; this includes short-term or tactical actions as well as operational or capital-prioritized replace/repair decisions. How priorities or objectives may change in importance, in composition, or both Must be calibrated or re-calibrated to changes in external factors influencing the nature and timing of asset value that can be realized. As a result, analytics are able to consider and explore changes in strategies, objectives, and requirements, stakeholder needs and expectations across a range of possible outcomes. The emergence of scalable Generative AI technologies, specifically Generative Pre-training Transformers (GPT) and Reinforcement Learning (RL), provides unprecedented ability to deploy embedded targeted solutions to characterize and optimize future actions and timing/sequence of such actions. GPT offers exciting if not unprecedented ability to apply AI-enabled advanced analytics to identify patterns in data that: Mathematically describe relationships between different aspects of operations, maintenance, capital expenditures, and behaviors against performance outcomes Sponsor powerful models that can leverage these insights and predict how specific aspects of asset performance and reliability will be transformed under different operations and maintenance and utilization scenarios. Relate asset value and its contribution to value-based objectives at location over time using dynamic market-based models that can take weather, market, technology, and services profiles. Can be leveraged to derive expected outcomes from a very broad set of factors, based on ground truth. The value of GPT is additive and incremental. GPT-driven analytics has successfully identified time-critical features of the market to optimize market participation and bid strategies; for NOCs, we extend this GPT fabric to incrementally address asset condition/health, and performance/utilization features of the underlying assets. Reinforcement Learning (RL) brings optimization to the forefront. RL applies credit functions that can be applied to track outcomes overtime as a function of actions taken under different operating conditions across a range of contextual parameters. This allows the system to learn optimal responses to a wide variety of situations and to exhaustively explore the effect differences actions taken can influence performance, behavior, and outcomes. This opens exciting and new pathways to engineer and generate a more agile, learning NOC!
· · · · · · · · · · · ·

Key Topics and Perspectives

March 3, 2024: Bridging the Gap: Aligning Asset Management Systems with Analytics for Optimal Business Outcomes -- How to unleash the power of analytics and AI within the context of dynamics and outcome-driven asset management systems. February 1, 2024: The School of Hard NOCs -- Bringing business outcomes focus to asset portfolios monitoring and operations.

Blog Articles

Bridging the Gap: Aligning Asset Management Systems with Analytics for Optimal Business Outcomes

Increased diversity of asset ownership, remaining life, and business value creates increasingly significant opportunities to better manage asset risks/opportunities – a simple concept backed by a myriad of examples across industry … yet vexing in terms of actual practice! The landscape is littered with remnants of methods, tools, solutions, and IT systems that failed to deliver needed business value or evolve to fit new needs or circumstances. So, what is the cause? I would argue that the crux of the problem lies in the combination of complexity of both operational and business processes and the inadequacy of coordinating actions, investments, and strategies against evolving business and system dynamics. And what is the remedy? In simple terms, we must be able to match the power and diversity of analytics/AI against the diversity of business and technical decisions; and we need to do that based on business outcomes. It is critical to realize that outcomes are dependent on both what one does – what actions are taken or to be prioritized – and the dynamics and evolutionary changes within the business, market, and technological environment in which companies and assets are participating. I believe the actual solution is to combine the application of transparent, agile outcome-focused processes AND the adoption of a radically more open, modern, and flexible methodology for the application of analytics/AI to create a truly cohesive framework, laser-focused on delivering business outcomes! And, as a departure, from traditional bottom-up approaches, I believe the key here is to begin this enhanced analytics and AI journey by focusing on the big picture and leveraging the power of AI to explore how combinations of actions in concert with changes in the business environment impact business outcomes. This cannot be done in business silos, but from the top-down at enterprise scale. And the key to success is? Consistent, timely, and informed business decision making methodologies driven by foundations in both asset management and analytics/AI aligned by three key requirements. 1. True Coherence: processes, tools, analytics, decision making frameworks but be aligned, suitably flexible, and goal/outcome-oriented. Translation: analytics/AI need to be able to equally goal/outcome-oriented. 2. Ability to Minimize Regret: we seek to “optimize” decisions by analyzing both the good and the bad of past decisions we did or did not make; importantly, in terms of achieving optimal outcomes, we seek to identify combinations of actions that will minimize regret, to avoid costly circumstances, and unlock maximum value. 3. Embracing Optionality at Scale: Increased dimensionality (aka things to consider) in concert with dynamic asset management conditions translates to greater optionality and value in exploring this increasing large decision space. Foundation 1: Outcome-oriented, Asset Management System Jack Dempsey from AMPS ( astutely defines the relationship between asset management systems and outcome-driven behaviors. He emphasizes the continuous quantification of expected outcomes through leading indicators while viewing inputs as a comprehensive mix of objectives, strategies, plans, standards, processes, and resources. To accomplish this goal, management systems leverage data and analytics to determine the best possible outcomes given circumstances, i.e., to minimize the price of regret, allowing one to expand their decision-making horizon from asset or system to an analysis of system behavior and global outcomes. In this manner, one can identify potential regret and quantify its impact before it occurs, thus enabling proactive decision-making to avoid or minimize the impact of regret. In recent articles, Jack goes a step further, highlighting the need to explore the full range of actions and options. Importantly, he stresses embracing the inherent optionality arising from the diverse markets and evolving needs that challenge today's asset managers. The relationship between optionality and asset management system (AMS) flexibility lies in driving choices and creating value. A flexible AMS allows for a broader range of choices in managing assets, placing the onus on supporting AI/analytics systems to cater to increasingly diverse needs. The flexibility enables organizations to adapt to changing circumstances, make informed decisions, and generate more value from their assets. Foundation 2: Advanced Analytics/AI As the complexity and timeliness requirements of decisions have increased, the limitations of traditional decision-making approaches have become more apparent. Decisions have traditionally been made in silos, with different groups independently tackling different dimensions of a massively interdependent landscape, producing recommendations from a limited understanding of the bigger picture and, hence, full of regret. This siloed approach was a practical reality, given traditional methodologies and tools, lacking the capacity to synthesize a cohesive perspective for decision-makers. Dr. Curt Lefebvre of nDimensional, an expert in the field of industrial AI, notes that the advent of next-generation AI technologies, particularly generative pretrained transformers (GPT) and reinforcement learning (RL), is revolutionizing industrial decision-making. These advanced technologies are enabling a seismic shift from siloed, bottom-up decision-making processes to enterprise scale, top-down, dynamic approaches. AI can now learn from both real-world data and human expertise to gain a systems-level understanding of entire domains at massive scale. With a comprehensive view of system behavior and dynamics, RL agents ask: “of all actions that could be taken, which sequence will result in the best future outcomes.” AI explores a huge number of simulations across massively dimensional spaces to come up with optimal recommendations. It learns and adapts to changing situations, continuously calculating the most effective strategies to maximize future returns. Curt also noted that AI-led approach is, critically, based on the very same principles of coherence, minimization of regret, and exploration! A Revamped Charter By coupling generative AI with the organization’s tailored digitalized management systems, it is possible to augment human-in-the-loop decision making, enabling organizations to game through an expansive array of potential futures to navigate their best course of action for success. Required coherence and flexibility between management systems and analytics allows us to focus on examining how combinations of actions can drive business value under diverse or changing conditions to provide critical insights to strategy, plans, or business investment priorities.