Decision Intelligence is the most advanced result of Artificial Intelligence techniques applied to risk mitigation processes and the achievement of objectives
Maintenance now plays a central role in the management of complex machines and systems. In modern plants, it is in fact increasingly important to guarantee pre-established levels of productivity and availability also in consideration of the fact that these are often parameters included in the contractual terms.
In addition to these aspects, it is also essential to be able to control the overall management costs of the plants since, in an enlarged economic scenario, competition is fierce and globalization leads to a direct confrontation with Emerging Countries that have labour at lower costs. compared to Western countries. The issues related to effectiveness, measurable in terms of availability and productivity, and efficiency, which instead can be assessed on the basis of specific management costs, are increasingly interconnected and their control opens the scenario to multiple trade-offs that, especially in the case of complex systems in which the operating variables are many, it is not possible to fully govern with traditional methods.
Industry 4.0 and Artificial Intelligence are making a truly significant contribution to the evolution of maintenance, which has gone from a purely reactive operation to a preventive, predictive and finally prescriptive operation. This evolution is fundamental for identifying and controlling the risks associated with the operational life cycle of complex systems by organizations, both in terms of magnitude and frequency of unwanted events. And it is precisely in this context that the new machine learning techniques of Industry 4.0, based on Artificial Intelligence, can represent a valuable tool for automating the analysis of large amounts of data from plants and machinery and for prevention. associated risks.
Naturally, Artificial Intelligence must be powered by learning models including deep learning, a model based on the stratification of artificial neural networks.
Outlined the technological environment in which we operate, what Decision Intelligence is and what it represents today in such an environment? Decision Intelligence is the most advanced result of Artificial Intelligence techniques applied to risk mitigation processes and the achievement of objectives. The advanced management of complex systems through innovative Decision Intelligence techniques can revolutionize the way an industrial organization coordinates its strategies and operations. It is therefore a discipline that encompasses a vast range of advanced techniques aimed at modelling, executing, monitoring and refining models and decision-making processes. In structured organizations – for the most part, multinationals, more rarely local – this discipline orients the corporate decision-making process towards defined objectives both from a strategic and operational point of view thanks to descriptive analysis, diagnostics and predictive data coming from the field. The outputs of the analysis related to the Decision Intelligence technique are then screened in a context of iterative process management, which consists in carrying out simulations of alternative scenarios in relation to an objective to be maximized or minimized. These alternative scenarios can be analyzed through an advanced Artificial Intelligence calculation of a descriptive or predictive type of future trends.
Maintenance, understood as an area of competence to which the management and responsibility of the assets are entrusted, can, in turn, be equipped with these new technologies, becoming, in its form of prescriptive maintenance, a useful and necessary tool for the effective management of complex systems. in various fields of application which include, among others, the industrial, maritime, civil and energy sectors.
When we talk about ” prescriptive maintenance “, however, we do not refer to a maintenance strategy defined by the regulations, but to a specific methodology, to a particular approach, like proactive maintenance, which, starting from the purposes of two industrial maintenance strategies such as Reliability Centered Maintenance and Prognostic Maintenance, uses the new tools of Industry 4.0 to generate value. So here we are at the true purpose of Decision Intelligence as an evolution of Artificial Intelligence: to generate value for a complex industrial reality.
This new approach to maintenance, which represents the most modern evolution of predictive maintenance, offers the opportunity not only to predict the onset of problems and failures but also to prescribe solutions and provide decision support, identifying actions to be undertaken and selecting the optimal risk management policy in terms of safety, availability, profitability and sustainability.
Prescriptive maintenance, therefore, makes it possible to simulate alternative scenarios on the basis of certain actions and to evaluate their impact both from a financial and safety point of view, along different time horizons and under certain conditions. In this way, it is possible to calibrate decisions such as the decommissioning of a specific structure, the replacement of a system to reach the end of life, the remodelling of the amortization plan and the revision of some items of the income statement. Thanks to the possibility of making inferences, that is to draw deductive conclusions, supported by Artificial Intelligence in a decision-making network, Decision Intelligence also allows for the evaluation of the possible impacts of simulations including the”Life cycle assessment” and the “life cycle cost”.
What has been described up to now underlines the extreme importance of data in the creation of knowledge bases and information assets for organizations that intend to apply the discipline of Decision Intelligence virtuously to their complex systems. The availability of data and informative feedback are therefore the enabling conditions for prescriptive maintenance. The complex systems that prescriptive maintenance examines are all those industrial assets whose components and related interconnections potentially determine interference that can translate into risks for organizations, people and the environment.
The possibility of having systems capable of collecting large amounts of data in real-time and of acquiring information feedback from these systems can be favoured by the use of IoT platforms that also support Decision Intelligence algorithms. In addition, according to a model of circularity of information that is consolidating more and more in structured organizations, IoT platforms can also fulfil the requirement of “data storage and visualization”, representing the results of simulated scenarios through appropriate graphic interfaces with the dual purpose to facilitate the understanding of the final decision maker and to enable the creation of a historical archive. In cases with a high automation content, the selection of the best decision can also be delegated to machine-to-machine systems.
Decision intelligence thus becomes a data-driven tool, in which decisions, variables and impacts are monitored and predicted in relation to the achievement or deviation from the objectives set a priori. Having digital and innovative tools that exploit this discipline therefore becomes a fundamental step in governing the stochastic processes on which the objectives of many organizations depend. However, the effective management of such processes must first pass through the creation of a methodology (see Figure 1) which includes collaboration between the various company stakeholders, the mapping of information resources and enabling technologies.
In conclusion, an advanced approach to prescriptive maintenance, if it fully incorporates the innovation deriving from Decision Intelligence processes, can represent a real paradigm shift at an industrial level and bring concrete added value to traditional maintenance strategies.