Cómo los diagnósticos inteligentes están cambiando las reglas de la monitorización de máquinas

Editor’s note: This article has been updated to reflect the latest practices in online condition monitoring and smart diagnostics.

More efficiency through automation. Greater productivity from centralizing or outsourcing technical expertise. These evolutions of global market demand have launched a new era of connected, data-driven monitoring. Where online systems were once relegated to monitoring only critical plant assets because they were expensive to purchase and even more expensive to install, advancements in sensor technology, edge computing, and AI-based diagnostics now allow plants to automate condition assessments and monitor remote installations with confidence.

Historical Barriers to Online Monitoring

Once upon a time, high sensor and cable installation costs made online monitoring systems an unattractive alternative to portable systems or limited their use to only the most critical of plant assets. Back then, traditional online systems consisted of a central processor to which all of the sensors were connected. This meant laying a great deal of cabling, often in conduit, to get the signals to the central processor. The cable installation costs would often dwarf the actual hardware and software costs required for the online system. But sensor prices have dropped considerably, and design innovations greatly lowered both system and installation costs. Today’s modular, edge-based systems use Ethernet, Wi-Fi, or industrial wireless protocols that make installation faster, smarter, and far less costly.

Wireless Sensors vs. Wireless Systems

Two models share the common aim of reducing cabling costs or removing cables entirely. Wireless sensors contain small transmitters that relay data back to a central processor, often through a wireless access point. The benefit of these sensors is that there really are no wires. They easily move from machine to machine if necessary.

There are challenges, however. For one, these sensors are configured to collect and transmit data on a schedule, not based on what the machine is doing. Though this saves battery life, it creates a significant technical limitation of the devices themselves for applications involving variable-speed or load machines.

In these applications, the benefit of an online system is that it checks the machine state before testing to ensure repeatable test conditions for trending, as well as identifying whether the operating state itself has changed. Without repeatable test conditions, vibration monitoring is not very useful.

Wireless systems assume there will be multiple sensors per machine and multiple machines to be monitored within a relatively small area. In this case, it makes sense to use off-the-shelf, inexpensive sensors (including non-vibration related sensors) that are right for the application, and cable them short distances to a processor located on or near the machines of interest. The processor then sends diagnostic results and/or data wirelessly or via Ethernet to a central server or to individual workstations. Multiple processors are installed plant-wide, close to the machines they are monitoring, then integrated on a higher level at a server or control center.

Modern systems now combine these approaches—leveraging adaptive sampling, long-life power solutions, and edge analytics to maintain precision while reducing network load.

Diagnostics vs. Data and Alarms

Now that system and installation costs are so affordable, the next question is what is the system used for? Many older online systems use simple overall alarms to determine increases in vibration, but because many operators already know that certain increases do not necessarily indicate mechanical problems, these alarms are often ignored.

If these alarms are not ignored, operators face the challenge of deciding what to do next to determine, what, if any, problem exists with the machine. This often requires more technological knowledge and time than is currently available onsite. Some older online systems simply send heaps of data to a central site for analysis, as if they were just replacing portable systems. But what does one do with multiple machine tests per day when hundreds of machines are being monitored? Who analyzes this data? How does this manual approach to data analysis improve efficiency?

Today’s most advanced systems go beyond alarms—they use hybrid diagnostic engines that combine rule-based logic with AI pattern recognition to automatically identify and classify faults. This patented concept independently monitors the machines and sends data to a central site only if there is a change of status and/or on a predetermined time basis. It can detect hundreds of fault types, across pumps, motors, blowers, compressors, and generators—with accuracy approaching that of expert analysts. Reports include specific faults, with corresponding severities, plus an overall repair recommendation.

Remote Monitoring & Control

Effectively monitoring critical pumps in remote pipeline stations is very challenging. To perform high-quality machine condition assessments, vibration and other data must be sampled at very high rates. With modern cloud infrastructure and bandwidth availability, large datasets can now be streamed securely and efficiently to centralized monitoring hubs.

A key design feature of the newest online system is its proactive nature. It continuously assesses machine condition locally, but alerts specific individuals of a machine’s health problem only when a repair recommendation is made. This distributed intelligence—processing data at the edge and sending only actionable results—means complex analysis happens closer to the asset. This information, including concise fault diagnostics and severity, is sent to monitoring centers in near real time. Raw data becomes available periodically, on demand and – most importantly – when a machine’s status changes. If a problem is detected, it is quickly confirmed so machinists can be deployed with proper parts to repair the problem before a catastrophic failure occurs.

Today’s systems are also remotely managed and updated via secure over-the-air software delivery, allowing engineers to adjust configurations, update baselines, and troubleshoot assets anywhere in the world.

Interface

Results produced by new online system technology can be incorporated into a wide array of user interfaces, depending on the site requirements. As a generic out-of-the-box solution, the new technology produces real-time web dashboards and mobile-ready visualizations containing machine status and fault information that are viewed on workstations in the plant and beyond.

A real-time data server presents live data streams through cloud dashboards or integrated CMMS/ERP systems, which can be further analyzed by efficiency, power and differential calculations. Dynamic visualization tools now replace static spreadsheets, using color-coded health indicators, trend graphs, and machine schematics that update automatically.

Modern online systems integrate through open protocols such as OPC UA and MQTT, ensuring that diagnostic data and machine health status are seamlessly shared across platforms and control systems.

Join the Revolution

New online system technology offers a unique solution to ever-increasing demands for efficiency, automation, remote monitoring and consolidation of technical expertise. Affordable sensors, edge intelligence, modular design, and cloud-based connectivity are redefining how facilities approach condition monitoring. What was once reserved for critical assets is now scalable, predictive, and accessible to every operation aiming to protect uptime and extend asset life.

Author Bio: Alan Friedman is a senior technical advisor for Azima DLI. With more than 18 years of engineering experience, Friedman has worked with hundreds of industrial facilities worldwide and developed proven best practices for sustainable condition monitoring and predictive maintenance programs. Friedman contributed to the development of Azima DLI's automated diagnostic system and has produced and taught global CAT II and CAT III equivalent vibration analysis courses.