Managing equipment health
As equipment becomes more sophisticated, monitoring equipment health becomes increasingly important. Condition monitoring systems embedded into equipment provide enhanced management of critical assets by focusing on system wellness.
Downtime is a major avoidable cost. System wellness is the guiding principle of maintaining uptime by reducing downtime. This can be accomplished through the following approaches:
Maintaining the health of equipment
Detecting developing problems in motors and other rotating equipment
Logging abnormal events
Notifying the appropriate personnel about operating conditions.
Embedded controls are changing the face of condition monitoring. Although condition monitoring has been around for more than 20 years, it had been an afterthought. Machines were modified to accommodate the sensors required to monitor health parameters, such as temperature and vibration. However, the connection between condition monitoring and embedded control did not become widespread until recently.
Now plants are taking on more responsibility for the operation and health of their equipment. Plant engineers are realizing the importance of monitoring machine health. They are specifying machines that will diagnose themselves and report when they are not healthy. Machine health information is a crystal ball into how the machine is operating. More machine maintenance information translates into better production capacity forecasting.
Machine health has a direct effect on product quality, production capacity, and maintenance scheduling. To reliably monitor the health of equipment, it is necessary to measure key indicators, analyze data provided by these key indicators, and notify the appropriate personnel when these data deviate from predicted values. Key indicator data include parameters such as oil monitoring, temperature, motion and position, flow, and pressure.
There is more to machine health than monitoring abstract data. Decisions must be made that include information from these data.
The benefit of system wellness to a user is to reduce the total life cycle operating cost through the following methods:
Quickly detect health of motors and other rotating equipment
Provide the information to plant personnel
Identify problems before equipment is damaged or destroyed.
It starts with the sensor
Sensors convert a physical parameter or stimulus into a measurable electrical signal. Functionality of sensors has increased as more intelligence is integrated into them.
Sensor functions include current flow, liquid and/or gas flow, level, position, pressure, proximity, temperature, ultrasonic, vibration, and vision.
Current transformers (CTs) are sensors used to detect electrical current flow. A CT is a donut-shaped device that fits around an electrical conductor, such as a cable or wire. The signal it produces is directly proportional to the current flowing through the conductor.
A paddle switch is a simple device placed in a pipe that can detect flow of liquid or gas. The switch is actuated when flow is present. Flow measurement can also be more sophisticated. Flow meters determine flow of liquid or gas based on the measurement of differential pressure across an orifice mounted in the pipe. A flow switch determines if flow is present; a flow meter determines how much flow is present.
Level can be sensed using a float and switch actuation or potentiometric device. Level can also be detected using ultrasonics or radar techniques. Position is sensed using potentiometric devices. A variable resistance is measured based on the position of a motion element coupled to the potentiometric device.
Pressure transducers convert positive or negative pressure in a vessel or line into an electrical signal. A diaphragm deflects based on the amount of pressure sensed. This deflection is measured by sensing strain on the diaphragm or by mechanical means.
Proximity is measured primarily using Hall-effect sensors, which work like typical metal detectors. A high-frequency signal is applied to the sensor. The characteristics of this signal change as objects come into proximity to the sensor.
Temperature sensors can be thermocouples, resistance temperature detectors (RTDs), or thermistors. A thermocouple is constructed from wires made from two dissimilar metals. The dissimilar metals produce a measurable voltage, which is proportional to the measured temperature. RTDs vary in electrical resistance based on measured temperature. A thermistor is a semiconductor device that changes in measurable electrical characteristics based on measured temperature.
Ultrasonic sensors are typically piezoelectric microphonic devices. The sensor produces a measurable voltage, which corresponds with ultrasonic vibration or sound.
Vibration sensors are accelerometers, which are devices that sense inertial reaction to measure linear or angular acceleration. In its simplest form, an accelerometer consists of a case-mounted spring and mass arrangement. Displacement of the mass from its rest position relative to the case is proportional to the total acceleration sensed along the equipment’s sensitive axes.
Vision sensors are typically miniature video cameras or charge-coupled devices (CCDs) packaged for industrial applications. These and other sensors can combine in a system to monitor the health of a piece of equipment.
Signal conditioning adds a level of intelligence to the sensor, which is sometimes necessary to translate a signal to be compatible with a system. The next level of intelligence incorporates a sensor with communication capability. Even more sophisticated devices make decisions, perform diagnostics, and sometimes are self-calibrating or self-ranging. At a higher level of intelligence, a control function is integrated with a sensor function (Fig. 1).
Not every application requires an intelligent sensor. It is the discerning application of sensors, hardware, and software that enables the embedding of useful intelligence into equipment in the plant.
Condition monitoring and plant asset management
A plant asset management (PAM) system provides timely information to maintenance and operations personnel to aid them in safely increasing total production output at a reduced cost-per-unit of output. Benefits occur as the plant makes optimum operating and maintenance decisions.
A PAM system turns equipment measurement data into actionable information, then issues advisories to maintenance and operation personnel regarding equipment health (Fig. 2). PAM is a term originated by the ARC Advisory Group to address the following questions:
What equipment may fail if it does not receive maintenance intervention?
What intervention should be taken and how soon?
What parts should be ordered and how soon?
What is the optimal blend of condition-based maintenance, calendar or usage-based preventive maintenance, and run-to-failure/breakdown maintenance for a given piece of equipment?
Should any adjustments be made to a process now to prolong the life of equipment critical to the process?
How much can a process output be increased without incurring an unacceptably high risk of unexpected process slowtime, downtime, quality problems, or safety shutdowns?
What is the risk of successfully producing X amount of product next week given a projected process utilization rate of Y?
Figure 3 shows the structure of a typical PAM system and how the operational blocks interrelate. The following paragraphs summarize the relationship among these blocks.
Storing and categorizing equipment information
A typical PAM system uses a register storage method to provide the rest of the blocks with information about the identification, location, and criticality of the equipment. The registers also store measurement location and transducer information.
Harvesting the data
The data-harvesting block gathers data from equipment sensors. The harvester module periodically extracts data from process data historians as well as from protection, high-speed transient, periodic surveillance, and control device monitoring systems.
The data harvester synthesizes data from various monitoring technology systems, including shaft displacement, casing vibration, ultrasonic, electrical circuit, thermographic imaging, oil particulate, and oil chemical analysis. It correlates this condition-based monitoring data with current process data so the PAM system can properly associate the dependent variables, such as vibration, with the independent variables, such as speed and load.
Calculating key indicators
A key-indicator calculation block makes calculations on data from raw measurements and dynamic spectra. It also calculates macro indicators derived from multiple measurements, such as from differential pressure transducers. Calculations of rotating shaft and bearing vibration, sound, and electrical frequencies allow for a sophisticated fingerprint analysis of dynamic frequency data. These computed indicators are vital in discovering early abnormalities in equipment.
Archiving the data
A data archive block provides long-term data storage of plant equipment measurements with options for data error flagging, compression, and expiration. Archive blocks manage data expiration and allow physical deletion from the online database. However, many plants prefer to keep data in the archive block for up to five years to look at long-term trends in equipment condition monitoring and performance data.
State-of-the-art archive block schemes use industry-standard relational databases, such as ORACLE and Microsoft’s SQL Server. They allow external access for distributed database management and other database administration functions.
Condition monitoring is the heart of the PAM concept. This block enables the creation and maintenance of an equipment health baseline and searches for abnormalities whenever new data or indicators enter the PAM system. It allows plant engineers to establish normal and abnormal conditions for all measurements and computed indicators in the database. The measurements can range from temperature and oil particle counts to complex data such as vibration spectra or infrared images. The objective is to determine what is normal for the equipment and identify various abnormal alarm states.
Advanced PAM systems include inputs from process control data historians and sophisticated, embedded, state-aware condition monitoring technology — automatically setting multiple baselines for equipment based on variable operating loads, speeds, and other process conditions.
Analyzing equipment health
If exceptions are found, an equipment health analyzer is required. This module facilitates and permanently archives an analyst’s evaluation of the current health of the equipment in question. It integrates relevant data into information displays, which allow multidisciplinary data (lubrication, vibration, thermographic, ultrasonic, process data, etc.) to be visually compared in multi-parameter plots and graphs. Equipment health analysis is also aided through the use of automated diagnostic tools and rule sets.
After performing a diagnosis, a prognostic assessment is also needed to determine the future health of the equipment in question and its projected time to failure as well as failure mode. If the equipment failure mode will cause an impact on operations, an additional prognostic assessment is required.
Alerting maintenance when there is a problem
By incorporating the ability to perform these analyses into equipment control, these conditions can be quickly detected and appropriate people notified so action can be taken.
Typical gateway managers include e-mail and paging interfaces to notify plant personnel of urgent, impending equipment failures. Message templates can be configured offline. Then, when equipment conditions produce alarms, these previously designed messages can be used to alert the appropriate maintenance technician.
The bigger picture
Downtime is the single largest avoidable cost element in many industries. Unplanned downtime costs include the cost of lost production and, in process industries, can include the cost of material that is rendered useless when a process is shut down.
Equipment reliability is critical to reducing downtime. It not only reduces downtime but can maintain the quality of the product or material being produced.
PLANT ENGINEERING magazine extends its appreciation to Cutler-Hammer, eMation, Rockwell Automation/Entek, and Siemens Energy & Automation for their assistance in the preparation of this article.
Condition monitoring schemes embedded into equipment help plant engineers maintain equipment wellness.
There is more to maintaining machine health than monitoring abstract data.
Plant asset management (PAM) systems provide information that aids in reducing downtime.
Documenting financial gains
Implement a standardized process to analyze and document cost savings that are a direct result of a condition-based maintenance program. Institute a financial analysis process that is based on deferred maintenance cost. Through this ongoing process, continually prove the value of the program by relating it to bottom line business results. Evaluate where problem equipment is and how to strategically plan for the future. Develop formulas that are used to calculate the cost savings across material, labor, and fuel.
Upper management support
With the support of senior management, continually reinforce the strategic importance of a condition-based maintenance program. By relating maintenance performance to individual financial performance, management sends a clear message to everyone in the organization that maintenance cost savings are an important competitive component.
Full integration of condition-based practices into an overall maintenance philosophy
Incorporate condition-based maintenance practices into the existing maintenance organization and philosophy. This accelerates the acceptance and understanding of condition-based techniques into an organization.
Integrated test technologies
Embrace a wide range of test technologies (oil analysis, vibration analysis, motor monitoring, etc.) throughout the facility. Use a variety of condition-based technologies and apply them to equipment as appropriate. A primary objective should be to integrate condition-based maintenance information from multiple sites to help in making quality maintenance decisions. This technology is the foundation that allows a tie-in to all hardware, software, and information to give a clear view of the status of all critical equipment. Approximately every three years, conduct a formal assessment of the condition-based maintenance program. The assessment identifies specific strengths, weaknesses, opportunities, and threats. The plan should guide tactical activities and ensure the plant stays focused on the business objectives.
Where condition monitoring technology is being used:
Limit switches (sense actuation speed, overtravel, eturn to original position)
Motor control centers (sense drive info)
Bearings (rotating equipment)
Condition-based maintenance provides strategic advantage
Plants need to minimize costs and at the same time maximize their capacity. Since maintenance is one of the largest controllable costs, plants should look for ways to improve operational efficiency and reduce operations and maintenance expense. One way is to initiate a condition-based maintenance program.
There is value in integrating information from multiple plants and from a variety of condition-based technologies such as vibration and oil analysis. Move from a single, stand-alone condition-based maintenance system, to a local area network (LAN), and ultimately, to a wide area network (WAN) to integrate all technologies and sites into one coordinated system.
Keys to a successful condition-based monitoring program
There are four elements that are considered key to the success of a condition-based monitoring program.
Sensing pump and motor control health
Because of their numbers, pumps are one of the largest users of energy in industry. In many process industries such as refineries, paper plants, water treatment facilities, and chemical plants pumps are the primary pieces of equipment. Pumps can suffer from mechanical abnormalities such as broken impellers, debris lodged in the impeller, worn bearings, misaligned couplings, or loose foundations. Pumps also can cavitate, which is an abnormal operating condition that can be destructive.
Monitoring power delivered to the pump can indicate multiple problems such as low flow, excessive flow, and dry running. Monitoring current alone is not an adequate method since changes in voltage and power factor can significantly impact current.
Algorithms that analyze the power spectrum can indicate problems such as cavitation, unbalance, excessive bearing wear, and misalignment.
Embedded intelligence in motor controls
Embedded intelligence in motor control devices is being used to reduce mechanical and electrical stress on the motor, load, and motor control devices. Increasingly sophisticated protection algorithms are being implemented to avoid nuisance tripping and provide more effective protection. These directly provide enhanced system wellness by maintaining the health of the control equipment.
The ability to connect motor control devices to factory networks is being used to provide critical motor operating parameters to the enterprise level. This is a critical enabler for the rapid identification, notification and response to operational issues such as overcurrent and fault conditions.