ISO 13374: Condition monitoring and diagnostics of machines – Data processing, communication and presentation

Summary

ISO 13374 is a highly influential standard in the world of industrial IoT and condition monitoring software. It addresses the challenge of interoperability between different monitoring systems, sensors, and software platforms. Instead of defining measurement techniques, it specifies a standardized, open architecture for how condition monitoring data should be processed, stored, and exchanged. It is often referred to as the Machinery Information Management Open Systems Alliance (MIMOSA) architecture, upon which it is based. The goal is to create a “plug-and-play” environment for condition monitoring technologies.

Table of Contents (Conceptual Structure)

The standard is broken down into several parts and defines a layered information architecture. The core of the standard is a functional block diagram with six key layers that represent the data flow in any condition monitoring system:

1. 1. DA: Data Acquisition Block:

   This is the foundational layer, acting as the bridge between the physical machine and the digital monitoring system. The primary function of the DA block is to interface directly with sensors—such as accelerometers, proximity probes, temperature sensors, or pressure transducers—and to acquire the raw, unprocessed analog or digital signals they produce. This block is responsible for all low-level hardware interactions, including providing power to the sensors (e.g., IEPE power for accelerometers), performing signal conditioning like amplification and filtering to remove unwanted noise, and executing the analog-to-digital conversion (ADC). The output of the DA block is a digitized stream of raw data, typically a time waveform, which is then passed to the next layer in the architecture for processing.

2. 2. DP: Data Processing Block:

   This block is the computational engine of the monitoring system. It receives the raw, digitized data stream (e.g., the time waveform) from the Data Acquisition (DA) block and transforms it into more meaningful data types that are suitable for analysis. The core function of the DP block is to perform standardized signal processing calculations. This most notably includes executing the Fast Fourier Transform (FFT) to convert the time-domain signal into a frequency-domain spectrum. Other key processing tasks defined within this block include calculating broadband metrics like overall RMS values, performing digital integration to convert acceleration signals to velocity or displacement, and executing more advanced, specialized processes like demodulation or envelope analysis for detecting the tell-tale, high-frequency impact signals associated with rolling element bearing faults.

3. 3. DM: Data Manipulation Block (State Detection):

   This block marks the critical transition from data processing to automated analysis. It takes the processed data from the DP block (such as RMS values, specific frequency amplitudes, or spectral bands) and applies logical rules to determine the machine’s operational state. This is where the initial “detection” of a problem occurs. The primary function of the DM block is to perform threshold checking. It compares the measured values against pre-defined alarm setpoints, such as the zone boundaries defined in ISO 10816 or user-defined percentage changes from a baseline. Based on these comparisons, the DM block assigns a discrete “state” to the data, such as “Normal,” “Acceptable,” “Alert,” or “Danger.” This output is no longer just data; it is actionable information that can be passed to the next layer for diagnosis or used to trigger immediate notifications.

4. 4. HA: Health Assessment Block:

   This block functions as the “brain” of the diagnostic system, answering the question, “What is the problem?” It receives the state information (e.g., an “Alert” status) from the Data Manipulation (DM) block and applies a layer of analytical intelligence to determine the specific root cause of the anomaly. This is where diagnostic logic, which can range from simple rule-based systems to complex artificial intelligence algorithms, is executed. For example, if the DM block flags an alert for high vibration at a frequency that is exactly twice the shaft’s running speed (2X), the rule-based logic in the HA block would correlate this pattern to a specific fault and output a diagnosis of “Probable Shaft Misalignment.” Similarly, if the alert is on a non-synchronous, high-frequency peak with characteristic sidebands, the HA block would diagnose a specific “Bearing Defect.” The output of this block is a specific health assessment for the machine component.

5. 5. PA: Prognostic Assessment Block:

   This block represents the pinnacle of predictive maintenance, aiming to answer the crucial question, “How much longer can it run safely?” It takes the specific fault diagnosis from the Health Assessment (HA) block and combines it with historical trend data to forecast the future progression of the fault. This is the most complex layer, often employing sophisticated algorithms, machine learning models, or physics-of-failure models. The goal is to extrapolate the current rate of degradation into the future to estimate the Remaining Useful Life (RUL) of the component. For example, if the HA block identifies a bearing defect, the PA block would analyze the rate at which the defect frequencies have been increasing over the past several months to predict when they will reach a critical failure level. The output is not just a diagnosis, but a concrete timeframe for action.

6. 6. AP: Advisory Presentation Block:

   This is the final and most critical layer from the user’s perspective, as it translates all the underlying data and analysis into actionable intelligence. The AP block is responsible for communicating the findings of the lower layers to human operators, reliability engineers, and maintenance planners. Its primary function is to present the right information to the right person in the right format. This can take many forms, including intuitive dashboards with color-coded health indicators, automatically generated email or text message alerts, detailed diagnostic reports with spectral and waveform plots, and, most importantly, specific and clear maintenance recommendations. An effective AP block doesn’t just state that a bearing has a fault; it provides a comprehensive advisory, such as: “Inner race defect detected on motor outboard bearing. Remaining useful life estimated at 45 days. Recommendation: Schedule bearing replacement at the next planned shutdown.”

Key Concepts

• Interoperability: This is the primary goal of ISO 13374. By defining a common framework and data model, it allows a company to use sensors from Vendor A, a data acquisition system from Vendor B, and an analysis software from Vendor C, and have them all work together.
• Open Architecture: The standard promotes the use of open, non-proprietary protocols and data formats, preventing vendor lock-in and fostering innovation in the condition monitoring industry.
• MIMOSA: The standard is heavily based on the work of the MIMOSA organization. Understanding MIMOSA’s C-COM (Common Conceptual Object Model) is key to understanding the detailed implementation of ISO 13374.
• From Data to Decisions: The six-block model provides a logical pathway from raw sensor measurements (Data Acquisition) to actionable maintenance advice (Advisory Presentation), forming the digital backbone of a modern predictive maintenance program.

Official ISO Standard

For the complete official standard, visit: ISO 13374 on ISO Store

← Back to Main Index