Using PCs for machine condition monitoring: Part 1

Editor’s note: This is the first article in a series on PC-based machine monitoring.

By Plant Engineering Staff July 20, 2004

By Armando Valim, National Instruments, Austin, TX

(Editor’s note: This is the first article in a series on PC-based machine monitoring. The next article in the series will appear in the next issue of Plant Engineering HotWire.)

There are several reasons to monitor machinery and probably the most simple one is the old “time is money” adage. It is critical to keep machines running during production use and to stop them when maintenance is really necessary and when parts and personnel are available.

Advances in personal computers and related technologies have had a significant impact on plant maintenance operations — specifically machinery vibration monitoring. The widespread adoption of PCs, which is now a given, has driven progress in their reliability, flexibility, and performance. Their use is extended beyond the traditional office environment to everyday operations on the plant floor.

Plant engineers and managers benefit from the flexibility of software, integration of modern networking, and accuracy of analog-to-digital converters (ADCs) with 24 bits of resolution. Today’s combination of technology delivers not only higher performance but also lower cost compared to traditional turnkey systems.

This series of articles examines the components of a modern machine PC monitoring system and discusses how advancements provide higher performance at a lower cost.

Fig. 1. As the cost of machine monitoring technologies decrease, applications for PC-based monitoring increase, enabling plant engineers to provide more capacity.

Advancements in measurement

Plant engineers, maintenance supervisors, and instrument technicians look for ways to improve reliability and increase equipment uptime through condition monitoring and asset management. Understanding the transducers used for these reliability technologies and using them correctly help make an implementation successful. As we all know: “garbage in, garbage out.” For example, signal conditioning for the transducers is one of the most important — and most overlooked — components of a plant data acquisition system.

Properly applied signal conditioning brings real-world signals into a digitizer/ADC. Many sensors require special signal conditioning technology. However, no instrument has the capability to provide all types of signal conditioning to all sensors. For example, thermocouples produce very low-voltage signals, which require amplification, filtering, and linearization. Other sensors, such as strain gauges and accelerometers, require excitation power, amplification, and filtering.

Signals from other sensors may require isolation to protect the system from high voltages. No single instrument can provide the flexibility required to make all of these measurements. However, with front-end signal conditioning, you can combine the necessary technologies to bring these various types of signals into the data acquisition system.

A recent development that improves data management related to measurement and automation systems is the adoption of smart sensors and increased integration of data from the sensor into the asset management system. Making sensor information an integral part of the asset management system increases plant efficiency, productivity, and reliability.

As computers get faster, technology in general advances in rapid succession. As a result, we see the benefits of sensor manufacturers using several related technologies. While processes are improving, costs are coming down, and some sensors now offer built-in intelligence. IEEE 1451.4 has paved the way for smarter sensors that self-identify, store configuration information, and minimize time and effort for set-up and troubleshooting. By simply connecting the sensor to measurement hardware, technicians and operators eliminate the need to manually enter configuration data.

Using modern PC-based monitoring equipment, engineers, technicians, and operators can take advantage of the benefits of smart sensor technology and integrate it into asset management systems through industry standard protocols such as XML. Connection of the measurement system to the analog sensor is defined by Transducer Electronic Data Sheet (TEDS) IEEE standard 1451.4. The standard describes the interface for creating smart sensors, with built-in EEPROMS that contain information about the sensor such as calibration, range, etc. — essentially an electronic data sheet you can use within your test application.

Improvements in resolution

More and more, the plant engineers can use commercial technology in their applications. The historical price of ADCs continues to go down, while new technologies are surfacing as prime candidates for plant applications (Fig. 2). However, it is important to understand the key specifications that provide information on both the capabilities and the accuracy of the data acquisition products. Basic specifications, which are available on most data acquisition products, list the number of channels, the sampling rate, the resolution, and the input range.

Fig. 2. As a reflex of Moore’s Law, ADC prices continue to decrease — leading to more resolution and better performance of data acquisition systems in plant applications.

The lower-cost ADC trend opens a new front for digitizers with up to 24 bits of resolution that deliver nearly 120-dB dynamic range. This makes machinery protection and fault prediction more accurate than ever before. Advanced filtering techniques and built-in signal conditioning are other hardware improvements making modular, PC-based instruments the modern choice for machine-condition monitoring. Most of this new technology comes from audio developments made for the exigent audio industry. This technology is delivered in a more efficient form of ADCs for plant applications called delta-sigma converters.

In general, delta-sigma ADCs are most compatible with high-resolution (> 16-bit), low-to-medium speed (& 1MS/s) (MS/s = megasamples per sec) applications and are available in two broad “flavors” — dc and ac measurements. Delta-sigma ADCs for dc are obviously optimized for absolute dc accuracy and generally returns data very accurately but quite slowly, approximately 10 kS/s or less (kS/s = kilosamples per sec). Delta-sigma ADCs for ac are optimized for high-resolution ac spectral measurements at or slightly above the audio frequency range to about 1 MS/s. Both types incorporate powerful digital filtering because of the noise shaping that is part of delta-sigma design, which makes increasing resolution as easy as adding specific digital filtering. Delta-sigma techniques also facilitate low-cost designs because they can be built around single-bit data converters, which are inherently linear, simple, and inexpensive.

Plant applications use a class of delta-sigma converter that has excellent ac performance and works well in applications requiring faithful reproduction of the spectral content of a signal. High resolution and dynamic range, combined with excellent linearity and digital filtering to remove aliases, are important qualities for any sort of spectral analysis. Common configurations have four or eight channels with more than 110-dB dynamic range for making high-accuracy frequency domain measurements. The input channels of these devices simultaneously digitize input signals over a bandwidth from dc to 45 kHz. Engineers use these devices based on delta-sigma products in such diverse applications as acoustic measurement, rotational vibration analysis, and structural vibration analysis.

Increased intelligence

Keep machinery running and available for production requires you to predict when and where faults will occur, allowing you to make repairs during scheduled shutdowns. In addition, capturing vibration signatures and related machine parameters provides insight into machine performance and optimal operational speeds and settings. This insight can lead to lower operational costs and increased consistency in product quality.

To monitor machines and control processes, there are several hardware routes — PCs, programmable logic controllers (PLCs), portable data acquisition devices, wireless transducers, proprietary microcontrollers, and others. However, the key factor is software — not hardware. Software enables generic PLCs and PCs to perform specific tasks. Software gives the hardware the intelligence and ability to respond to different inputs. Software links hardware such as SCADA, fieldbuses, and PLCs to manufacturing execution systems (MES), enterprise software, and CMMS. In fact, engineers look for functionality and flexibility that are only possible on systems where the programming software is full featured and easy to use.

Some programming languages also work in real-time environments to provide reliability and hardware timing to critical applications. A graphical programming language can be used to program PLCs and to deploy in rugged environments. The developing language can also target field programmable gate arrays (FPGAs), which reduces development times and increases control speeds. FPGA technology deploys custom applications to reconfigurable I/O (RIO) hardware, as well as control timing, synchronization, and priority of operations on the hardware level. FPGA hardware leverages digital and analog inputs and outputs, hardware timing, and multithreaded operations for deterministic applications.

For example, significant improvements can be obtained in the design phases of plant machines. Engineers can design machines to facilitate calibration and monitoring. In the gas turbine market, successful monitoring depends on turbine design due to the complexity of the application. While calibrating, software assumes its key role of acquiring process and failure data to prioritize and streamline maintenance activities. Software links all the dimensions of the system and, when programmed with intelligence, affects cost dramatically. An easy-to-use package brings intelligence to the system — from simple logic or neural networks to fuzzy logic. It also facilitates measurement of a variety of variables and communicates with several devices and buses. The result is more reliable information that can be integrated faster into not just the data acquisition system, but also the entire MRO/asset management system.

Unifying technology

The PC brings machine monitoring to plant applications. PLCs, fieldbus, and SCADA benefit from PC improvements. The computer not only provides analysis, networking, and slots for data acquisition monitoring hardware; it also provides greater flexibility in integrating a wide array of information and scalability in building toward a plant-wide proactive maintenance program.

Many companies are using PCs to improve technical data management and increase reliability. For example, with the dissemination of networking technology, different plant sites can work together and achieve new levels of efficiency. Engineers, with the use of an embedded web server, can quickly see what is happening at a remote data logger or can compare machines at different locations. They also can use virtual remote panels to view and/or control an operation through a standard web browser in the middle of the night from home. They also can use data logging and supervisory control software to quickly create a sophisticated client/server architecture based on their needs. Ethernet-enabled plants can now connect process control, data acquisition, and vibration monitoring/protection systems. Engineers and technicians can now control and monitor plant processes and assets from a single location.

As PCs improve, engineers are using hardware more in PCs and taking advantage of the high-volume analog components of the telecommunications and industries. The wide use of networking and the internet within manufacturing plants provides an environment where data sharing is easy and commonplace. With new tools that are compatible with internet or intranet, it is now easy to monitor machinery from a remote location in real time using a browser. Several manufacturers are already setting up internet-based machine diagnostic services for supplied machinery. The results of these technologies are lower-cost, open-architecture hardware platforms that facilitate the widespread monitoring of machinery. Combining the power of today’s software and computers with the data acquisition and sensing capabilities yields the ability to quickly analyze the data and integrate the results immediately into plant floor operations.

Armando Valim is a group manager responsible for sound and vibration products at National Instruments, Austin, TX. He holds a bachelors in engineering as well as graduate degrees in engineering, marketing, and finance. For more information about using PC-based machine monitoring, Armando can be reached at 512-683-5861 or .