Predictive maintenance for industrial refrigeration

Case study: the contribution of Meridiana Aspiratori in the Cassandra Project

Predictive maintenance in industrial refrigeration is currently one of the most critical challenges for cold storage and refrigerated transport companies. The Cassandra project, developed by the Neapolitan startup Cogito SRL with the technical contribution of Meridiana Aspiratori and the support of the Campania Region , represents a concrete response thanks to an artificial intelligence-based device capable of monitoring electric motors in real time and preventing serious failures before they occur.

Why predictive maintenance in industrial refrigeration is urgent today

Imagine a refrigerated container laden with food products en route across the Mediterranean, or a cold storage facility operating at full capacity all night. In both cases, the electric motor powering the cooling system is the beating heart of the operation. If that motor suddenly fails, the consequences aren’t limited to the loss of goods: there’s the risk of a short circuit, a fire, or a disruption in a cold chain that can’t afford even a single hour of downtime.

This scenario isn’t uncommon. It’s the daily reality faced by thousands of companies in cold storage logistics, refrigerated transport, and the food industry. And until now, the standard response was just one: scheduled maintenance, periodic inspections, and a healthy dose of luck.

Predictive maintenance radically changes this logic. It’s no longer necessary to intervene once a failure has already occurred, nor is it necessary to waste time and resources on components that are functioning perfectly. We intervene first, at the exact moment when the precursor signals of an anomaly begin to emerge, invisible to the human eye but readable by an artificial intelligence algorithm. This is precisely where the Cassandra Project.

Real-time predictive diagnostics for electric motors

Cogito SRL is a Naples-based startup specializing in AI-based solutions for real-time industrial diagnostics. Its mission is to translate complex data—such as the electrical signals of a running motor—into useful, actionable insights to prevent failures before they become emergencies.

Cassandra is the Cogito SRL‘s flagship product, an advanced device for preventing failures in industrial refrigeration systems, designed for industrial refrigerators, vans, and refrigerated containers. Rather than requiring the installation of intrusive sensors or interrupting the production cycle, it connects directly to the three-phase power lines of the electric motors already present in the system and analyzes their signals in real time.

How it works technically

The device consists of high-precision sensors for current sampling, integrated data processing modules, and interfaces for remote management. Its operation is divided into three phases:

1. Continuous sampling: Sensors non-invasively monitor three-phase power line signals, detecting subtle changes in the motor’s electrical behavior.
2. Edge AI Analytics: This is the element that sets us apart Cassandra from a traditional monitoring system. Data processing occurs directly on board the machine, without relying on external servers or cloud connectivity. Artificial intelligence runs “at the edge,” ensuring real-time diagnostics with minimal latency and full operability even in environments with limited connectivity, such as onboard ships or in remote logistics areas.
3. Predictive identification: Algorithms analyze measurement patterns and identify anomalies that signal the onset of critical faults (short circuits, fires, overheating) early enough to allow for scheduled intervention rather than an emergency..

The institutional context

The project was not born in isolation. It was funded by the Campania Region through the Campania Startup 2023 Call and co-financed by the PR CAMPANIA FESR 2021-2027 . This institutional recognition is not a minor detail; it means that the project has passed a rigorous technical evaluation and is part of a specific regional strategy for industrial modernization through digital innovation.

Contribution to industrial predictive maintenance

Meridiana Aspiratori wasn’t just a component supplier for this project. It was an active technical partner, contributing throughout every stage of development, from operational logistics to the creation of the testing infrastructure that enabled the artificial intelligence training.

Hospitality and logistical support

The first contribution was to make available to Cogito SRL our operational headquarters in Fisciano (SA) served as a base for developing project-related activities. We offered not only physical space, but direct access to a real industrial environment, complete with equipment, expertise, and industrial extraction and ventilation know-how.

Technical design of electric fans

One of the most concrete contributions to the project was the design of three electric fans specifically developed for the experimental setting. This was not a selection process; it required a dedicated technical study to define the construction characteristics, electrical specifications, and operating conditions most representative of real industrial refrigeration systems.

The result was the creation of components that could faithfully simulate the behavior of electric motors in various operating scenarios (a prerequisite for collecting data of sufficient quality to train an artificial intelligence system).

Test bench construction

The collaboration led to the construction of a dedicated test bench, an experimental facility specifically designed to conduct systematic and controlled measurement campaigns. The test rig allowed us to reproduce different operating conditions and acquire large volumes of data on the motors’ electrical signals under each condition.

Training artificial intelligence

A predictive AI system cannot be developed in the abstract. Its algorithms must be trained on real data, acquired from real electric motors, under real operating conditions. Only in this way does the model learn to recognize the difference between the signal of a healthy motor and that of one approaching failure.

The measurements performed on the test bench built with the technical support of provided exactly this: a dataset of real data, collected with high-precision sensors on electric fans designed specifically for the project. This data was the “fuel” with which the algorithms Cassandra have been trained to recognize anomalous patterns in food signals.

In other words, without the physical infrastructure we helped build, AI Cassandra it couldn’t have learned. The data produced by Meridiana Aspiratori ‘s systems are literally incorporated into the device’s predictive capabilities.

Real-time diagnostics

The tangible outcome of the entire collaboration is the creation of operational real-time diagnostic devices. These tools continuously and non-invasively analyze the three-phase power signals of electric motors, using Edge AI to identify anomalies that precede critical failures. Their strength lies in the combination of three normally mutually exclusive features.

• Precision: High-quality sensors and advanced algorithms.
• Non-invasive: no interruption of the production cycle for installation or monitoring.
• Autonomy: on-board processing without network dependence. dalla rete.

Project results and impact

The system is designed to anticipate the most critical failures in electric motors of refrigeration systems:

• Fires: Identifying abnormal electrical patterns that precede overheating of components.
• Short circuits: Detecting variations in three-phase signals indicative of insulation deterioration.
• Unplanned machine downtime: allowing scheduled maintenance before a breakdown occurs.

The operating result is twofold: on the one hand, it reduces extraordinary maintenance costs, which by nature are always more expensive than preventive ones, and on the other, it improves operational safety by eliminating the risks associated with electrical faults in industrial environments.

A solution already available

Cassandra it’s not a laboratory prototype; it’s a device developed, tested, and available to companies operating in cold storage logistics, refrigerated transport, and the food industry. Direct application sectors in Southern Italy include port and maritime logistics, large-scale retail trade, and the food industry.

The significance of this collaboration for Southern Italy

The partnership between Meridiana Aspiratori and Cogito SRL it represents something that goes beyond a single project; it is a model of industrial collaboration that demonstrates how the most advanced technological innovation can arise from the combination of complementary skills.

On one side, an small to medium enterprise with 25 years of experience in the industrial ventilation and extraction sector, with in-depth technical expertise in the mechanics of electric motors and their real-world applications. On the other, a talented startup specializing in artificial intelligence, capable of transforming physical data into sophisticated predictive systems.

The synergy between these two entities has created an AI device trained on real-world data and validated in the field. A unique innovation that neither company could have achieved alone.

The value for Meridiana Aspiratori

The financing of the Campania Region it has provided access to the resources necessary for technological development, which requires time, infrastructure, and data—elements that a startup can hardly finance entirely with its own resources in the initial stages.

For the industrial ecosystem of Southern Italy, this type of tool represents a concrete opportunity to retain technological expertise in the region and modernize traditional sectors through digitalization, reducing the innovation gap compared to the more advanced areas of the country.

Participating in the project Cassandra has meaning for Meridiana Aspiratori much more than a technical collaboration. It meant positioning itself within a cutting-edge innovation ecosystem, developing direct expertise in applying artificial intelligence to its components, and demonstrating that a company specializing in the traditional industrial sector can be an active participant, rather than a mere spectator, in the digital transformation of manufacturing.

The three electric fans designed, the test bench built, and the data collected are concrete contributions that have made Cassandra what it is today.

The project demonstrates that predictive maintenance is no longer a technology of the future reserved for large multinationals, but is an available solution, developed in Campania by a Neapolitan company and an industrial partner from Southern Italy, with the support of regional and European institutions. For companies in the cold storage logistics, refrigerated transport, and food industry, the question is no longer whether to adopt predictive maintenance, but rather when do it.