How Manufacturing Engineers in Chemical Manufacturing Should Evaluate Condition Monitoring Solutions

Selecting a condition monitoring system for chemical process rotating equipment is an engineering specification problem, not a procurement comparison exercise. The evaluation criteria that matter in a continuous petrochemical or specialty chemical environment are not the same as those for discrete manufacturing. Classified area certification, process boundary integrity, vibration spectrum data depth, and data architecture for FMEA integration are evaluation dimensions that a monitoring vendor should be able to answer at the specification level, not just the marketing level.

This guide provides the technical evaluation framework that a manufacturing engineer in a chemical plant should apply when assessing condition monitoring solutions for centrifugal pumps in critical process service, compressors, agitators, and heat exchanger drivers. The framework covers hardware requirements, data requirements, installation engineering requirements, and the data output requirements that support your FMEA and PHA workflows.

A monitoring system that cannot document its zone classification for your specific process area, cannot provide full vibration spectrum data for failure mode identification, and cannot export alert history in a format suitable for PHA failure rate validation is a system that will not serve your engineering requirements in a chemical process environment, regardless of how capable it is in other industrial settings.

Hardware Evaluation: Classified Area Certification

The majority of process-critical rotating equipment in a continuous chemical plant is installed in electrically classified areas. Equipment that handles or transfers flammable or combustible process fluids, or is located in spaces where such fluids can accumulate in the atmosphere, requires electrical equipment certified for the zone classification.

ATEX and IECEx certification for European and international installations:

ATEX certification categories for gas atmospheres:

• Category 1G (Zone 0): Explosive atmosphere present continuously or for long periods
• Category 2G (Zone 1): Explosive atmosphere likely to occur occasionally during normal operation
• Category 3G (Zone 2): Explosive atmosphere not likely to occur during normal operation but may occur briefly

Most process-critical rotating equipment in chemical plants is installed in Zone 1 or Zone 2. Centrifugal pumps transferring flammable process fluids, compressor trains in flammable gas service, and agitators in flammable solvent service are typically Zone 1. Verify the zone classification for each installation location against the plant's area classification drawing before specifying a sensor category.

Gas group requirements:

ATEX certification specifies the gas group for which the equipment is certified:

• Group IIA: Propane atmosphere and equivalent
• Group IIB: Ethylene atmosphere and equivalent
• Group IIC: Hydrogen and acetylene atmospheres (most demanding)

A sensor certified for Group IIA is not acceptable in a Group IIC classified area. Verify the gas group at each installation location against the plant's area classification documentation.

Temperature class requirements:

The temperature class specifies the maximum surface temperature the equipment can reach:

• T1: Maximum 450°C
• T4: Maximum 135°C
• T6: Maximum 85°C

For installations adjacent to high-temperature process equipment, verify that the sensor's temperature class exceeds the maximum expected surface temperature at the installation location. Bearing housing surface temperatures on high-temperature service pumps can be 30 to 60°C above ambient.

North American certification equivalents:

For North American installations, NEC Class I, Division 1 is equivalent to ATEX Zone 1. Division 2 is equivalent to Zone 2. CSA certification provides equivalent classification for Canadian facilities. Verify that the sensor certification document covers the Division and gas group applicable to each installation location, not just a generic "Class I, Division 1" claim.

Evaluation questions for certification verification:

• Provide the ATEX certificate number and certification body for your sensor product in this gas group and temperature class.
• Can you document the temperature class for the junction enclosure and cable assembly as well as the sensor body?
• Have you installed in Zone 1 environments at other chemical plants with the same gas group classification? Can you provide a reference site?

Hardware Evaluation: Process Fluid Compatibility and Pressure Boundary Integrity

Chemical process rotating equipment presents two fluid exposure risks for monitoring sensors: incidental exposure from seal weeping and process drips under normal operation, and potential exposure during maintenance events or abnormal process conditions.

Pressure boundary integrity:

No condition monitoring sensor should be installed in a location that requires penetrating a pressure-containing component. The correct installation approach is external mounting on non-wetted surfaces: bearing housing exterior, structural brackets mechanically coupled to the machine frame, or motor frame adjacent to the bearing. All of these locations provide vibration transmission from the rotating assembly without requiring any modification to the pressure boundary.

Confirm with the vendor:

• Does sensor installation on this equipment class require any drilling, welding, or modification to a pressure-containing component?
• Is the mounting method entirely external to the pressure boundary?
• Does the installation generate a formal MOC requirement under your installation procedure for this equipment type?

For PSM-covered equipment, any installation that could be interpreted as a modification to mechanical integrity documentation may require a formal PSM MOC review. Clarify the plant's MOC procedure applicability before committing to an installation approach.

Process fluid compatibility for incidental exposure:

Sensors mounted near pump mechanical seal faces, compressor dry gas seal vents, or agitator shaft seals may be exposed to process fluid weeping during normal operation. The sensor housing material and sealing specification should be compatible with the specific process fluid at the plant.

Request the sensor manufacturer's material compatibility data for your specific process fluid. Generic chemical resistance charts cover common solvents but may not address specific process chemistry combinations. For chlorinated solvents, hydrocarbon process fluids, and corrosive chemical streams, verify material compatibility against your specific fluid's corrosion data, not a generic chart.

IP rating for chemical process environments:

Standard IP67 (dust-tight, temporary immersion protection) is the minimum acceptable for chemical process environments. For installations with regular washdown exposure or in areas where process fluid accumulation is possible, IP68 (continuous immersion rated) provides additional protection margin.

Hardware Evaluation: Temperature Rating for High-Temperature Service

Condition monitoring sensors in chemical process environments face two temperature exposure modes: elevated ambient temperature in high-temperature process areas and elevated surface temperature at the sensor mounting location.

Ambient temperature rating:

High-temperature process areas, particularly around reactor systems, fired heaters, and high-pressure steam systems, may have sustained ambient temperatures that exceed the typical industrial sensor rating. Verify the sensor's maximum ambient operating temperature against the expected temperature in the installation location, with margin for peak process conditions.

Mounting surface temperature:

Bearing housing surface temperature on high-temperature service equipment is a function of bearing operating temperature, thermal conduction through the housing material, and the thermal insulation properties of any jacketing. For pumps in steam service, hot oil service, or high-temperature reactor feed service, the bearing housing surface temperature at operating load may be 50 to 80°C above ambient.

The sensor's specified maximum mounting surface temperature must exceed the expected bearing housing surface temperature under worst-case operating conditions, not just typical conditions. Request the vendor's specification for maximum allowable mounting surface temperature, and verify it against the bearing housing temperature measured during normal operation at your plant.

Thermal isolation options:

For installations where the bearing housing surface temperature approaches the sensor's limit, thermal isolation hardware can provide sufficient margin. Verify that any thermal isolation approach does not degrade the vibration signal transmission quality required for frequency-domain analysis. Soft or compliant isolation materials absorb vibration energy in the frequency ranges relevant to bearing fault analysis and may compromise the monitoring system's ability to detect the failure modes you are relying on it to identify.

Data Evaluation: Vibration Spectrum Depth for Failure Mode Identification

The distinction between overall vibration amplitude and vibration spectrum analysis is the most critical data evaluation criterion for manufacturing engineers using monitoring data in process reliability and FMEA work.

What overall vibration amplitude tells you:

Overall vibration amplitude is the total energy content of vibration across all frequencies, expressed as a single value (typically acceleration in g, velocity in mm/s, or displacement in microns). It is useful for detecting that a condition has changed from baseline. It is not useful for identifying which component is causing the change or what the failure mode is.

What vibration spectrum analysis provides:

Vibration spectrum analysis decomposes the total vibration into its frequency components using Fast Fourier Transform (FFT) processing. Each failure mode in rotating equipment produces vibration energy at frequencies that are mathematically predictable from the machine's rotational speed and component geometry:

• Bearing inner race defect frequency (BPFI): Predictable from shaft speed and bearing geometry. A bearing fault on the inner race produces vibration energy at a specific frequency relative to shaft speed.
• Bearing outer race defect frequency (BPFO): Similarly predictable. Distinguishable from inner race faults by frequency ratio.
• Vane pass frequency: For centrifugal pumps, vibration at the vane pass frequency (shaft speed multiplied by number of impeller vanes) indicates hydraulic excitation that can signal cavitation or off-design operation.
• Rotor unbalance: Appears at 1X shaft speed (once per revolution). Distinguishable from bearing faults by frequency.
• Shaft misalignment: Produces 2X shaft speed vibration amplitude, distinguishable from unbalance.
• Cavitation precursors: Broadband high-frequency noise elevation, identifiable in the high-frequency portion of the spectrum.

For each of these failure modes, the engineering response is different. A bearing inner race fault calls for planned bearing replacement within the detected degradation window. A vane pass frequency anomaly calls for hydraulic system investigation. A cavitation signature calls for process parameter review. Overall amplitude data cannot distinguish between these response requirements.

Minimum spectrum requirements for chemical process monitoring:

• FFT spectrum coverage from 1 Hz to at least 10 kHz (for high-frequency bearing fault detection on compressors and high-speed pumps, coverage to 20 kHz is preferable)
• Frequency resolution sufficient to distinguish adjacent bearing defect frequencies on the specific equipment's shaft speed and bearing geometry
• Spectrum storage on a continuous or near-continuous basis, not just at alert events, to allow baseline comparison and trend analysis

Data Evaluation: Continuous Collection for Transient Event Capture

Periodic data collection, even at frequent intervals, creates windows during which transient events can occur and partially recover without being captured. In chemical process environments, several failure modes develop primarily during transients.

Startup and shutdown transients: Bearing and seal loads during startup and shutdown differ from steady-state operating loads. Failure modes that are initiated or accelerated during startup transients may not be detectable from steady-state operating data.

Cavitation events in centrifugal pumps: Pump cavitation events in chemical service may occur transiently during process flow changes, temperature swings affecting fluid vapor pressure, or upstream pressure variations. A periodic inspection that captures steady-state operation will miss a cavitation event that occurred during a flow transient.

Thermal cycling on seal faces: Mechanical seal faces in high-temperature chemical service experience thermal cycling as process temperature varies during production. Face distortion from thermal cycling is a precursor to seal failure that develops intermittently during temperature swings rather than continuously during steady-state operation.

Compressor surge precursors: Compressor surge in continuous chemical service can develop transiently during process flow changes before a full surge event occurs. The vibrational signature of approaching surge conditions may appear briefly and recover. Continuous collection captures this signature. Periodic collection may miss it.

For chemical process applications where the failure modes of greatest consequence develop transiently, a monitoring system that collects data only at scheduled intervals or only at alarm threshold crossings does not provide the full operating history that engineering analysis requires.

Data Evaluation: Export Architecture for PHA and FMEA Integration

The engineering value of a monitoring system in a chemical process environment depends partly on whether the data is accessible for analysis outside the monitoring platform. PHA updates, FMEA revisions, and equipment specification reviews require data in formats that engineering analysis tools can process.

Minimum data export requirements:

• Alert history by equipment tag: timestamp, alert type, alert magnitude, maintenance disposition (confirmed finding, false positive, or no finding), and failure mode category for confirmed findings
• Trend data export: ability to export time-series vibration and temperature data for specific equipment and time periods in a standard format (CSV or similar)
• Event documentation: for confirmed findings, a record that includes the alert date, maintenance inspection date, finding description, and estimated lead time before failure would have occurred

API access for engineering workflow integration:

For manufacturing engineers who use data analysis tools (Python, MATLAB, or engineering analysis software), API access to historical trend data enables direct integration of monitoring data into reliability analysis workflows. Verify whether the monitoring platform provides documented API access, the authentication method, and any rate or data volume limitations.

PHA-specific export capability:

For PHA and HAZOP update workflows, the monitoring data should be exportable in a format that allows tabular summary by equipment tag and failure mode. A spreadsheet export of alert history with failure mode classification, lead time, and maintenance disposition is the minimum. The ability to query by equipment class and service type allows the manufacturing engineer to build the plant-specific failure mode frequency table that supports PHA failure rate assumptions.

Specifying Monitoring Capability in Equipment FMEA

The FMEA detection column specifies the safeguard expected to detect each listed failure mode before it produces the listed consequence. When condition monitoring is the specified safeguard, the detection specification should be precise enough to allow verification.

Insufficient detection specification:

• "Vibration monitoring": does not specify what type of monitoring, what parameters are monitored, or what alert configuration is applied

Adequate detection specification:

• "Continuous vibration spectrum monitoring installed on inboard and outboard bearing housings. Alert configured on bearing defect frequency amplitude exceeding 2X baseline in the bearing geometry-specific frequency band. Shaft speed: [RPM]. Bearing model: [bearing number]. BPFI: [frequency]. BPFO: [frequency]. Collection interval: continuous."

This level of specificity allows three things: verification that the installed monitoring system is configured as specified, validation from monitoring history that the detection method has actually detected this failure mode in prior operational cycles, and formal FMEA revision when operational history requires updating the assumed detection reliability.

Including monitoring specification at this level of precision in equipment FMEA is the mechanism that makes the detection column defensible in a PHA review. It converts a qualitative detection assumption into a verifiable engineering specification.

Process throughput and OEE visibility for chemical operations: Evaluate whether the platform surfaces machine performance data at the level needed to detect throughput degradation, not just failure events, but the zone where degraded rotating equipment is running but underperforming: reduced pump flow, compressor efficiency loss, agitator mixing deficiency. Correlating machine health signals (vibration trend, power draw, temperature) with process throughput and yield data gives the Manufacturing Engineer the objective record to identify hidden process losses, the chemical process equivalent of the hidden factory where throughput is bleeding through degraded equipment before any alarm fires. Tractian's OEE solution extends condition monitoring data into production efficiency analysis.

Machine health to product quality and yield correlation: Evaluate whether the platform allows correlation of equipment health signals with batch quality and process yield data. In chemical manufacturing, the zone between healthy rotating equipment and failed equipment is where off-spec product is produced. An agitator with bearing wear producing inconsistent mixing, a pump with impeller wear producing pressure variation, these conditions affect batch yield and quality before they produce a maintenance event. The Manufacturing Engineer who can correlate a vibration trend with a batch deviation has the starting point for a root cause analysis that prevents recurrence.

Six Sigma and RCA data quality: Evaluate whether the platform produces timestamped, exportable machine health records suitable for Six Sigma projects. The maintenance-versus-process engineering blame cycle in chemical manufacturing, "the equipment was degraded" versus "the process parameters were wrong", is an information problem. Continuous machine health data covering the period of a process deviation gives the Manufacturing Engineer the objective evidence needed to separate mechanical root causes from process root causes and assign corrective action to the right function.