Electrical Monitoring

How to Choose an Electrical Monitoring System

A practical evaluation framework for selecting electrical monitoring technology for industrial motors and power distribution equipment, covering technique selection, deployment approach, and integration requirements.

Key Principle

The right electrical monitoring system depends on what you need to detect, how critical the assets are, and whether your team will interpret data or needs automated alarming. A system that is too simple misses faults. A system that is too complex generates data no one uses.

Step 1: Define What You Need to Detect

Different electrical faults produce different signal signatures. No single monitoring technique detects all fault types equally well. Start the evaluation by listing the faults you are actually trying to catch, then work backwards to the technique that detects each one.

The table below maps common fault priorities to the monitoring approach best suited to detect them, along with the minimum implementation required to get a useful result.

Fault Priority Best Monitoring Technique Minimum Implementation
Rotor bar faults Motor Current Signature Analysis (MCSA) Periodic portable measurement; sample under stable load conditions
Stator winding faults Electrical Signature Analysis (ESA) or offline IR/Polarization Index testing Periodic or continuous; ESA for online detection, IR/PI for annual offline baseline
Supply quality problems Power quality monitoring Panel-level revenue-grade meter with voltage and harmonic logging
Insulation degradation IR testing (offline) combined with continuous MCSA Annual offline IR/PI test plus continuous online MCSA for trending between outages
MCC and panel health IR thermography combined with ultrasound Periodic survey; at minimum annual during scheduled shutdowns

If your primary concern is rotor bar integrity on high-horsepower motors, MCSA is the correct starting point. If you are seeing unexplained motor failures with no clear mechanical cause, stator fault detection via ESA becomes the priority. Trying to cover both with a single technique often means compromising on both.

Step 2: Assess Asset Criticality

Not every motor warrants the same level of monitoring. Applying continuous monitoring uniformly across a facility is expensive and rarely necessary. The practical starting point is a criticality assessment that segments assets into three tiers.

Critical assets: continuous monitoring justified

An asset is critical when failure causes production loss that exceeds replacement cost within hours, when no installed redundancy exists, or when failure creates a direct safety risk. Compressors on single-train lines, large cooling tower fans in data center or pharmaceutical environments, and main drive motors on high-throughput production lines typically fall here.

For these assets, the cost of a single unplanned failure often exceeds the entire first-year cost of a continuous monitoring installation. The business case is straightforward.

Semi-critical assets: periodic monitoring sufficient

Semi-critical assets are disruptive to lose but manageable. The failure gives some lead time before full production impact, or a spare can be brought online within an acceptable window. Pump motors with installed spares, ventilation fans on non-clean-room processes, and conveyor drives on multi-path lines often fit this tier.

For semi-critical assets, a structured periodic inspection route using portable MCSA or ESA equipment, combined with annual offline testing, is usually the most cost-effective approach.

Non-critical assets: standard PM sufficient

Assets with installed redundancy, easy swap-out from local stock, or minimal production impact do not need predictive monitoring investment. Standard preventive maintenance intervals are adequate.

Document the criticality classification for each asset before selecting monitoring equipment. It directly determines whether you are buying continuous sensors, portable instruments, or neither.

Step 3: Choose Between Periodic and Continuous Monitoring

Once you know which assets warrant monitoring, the next decision is deployment model: periodic route-based inspection, permanently installed continuous sensors, or a hybrid of both.

Periodic MCSA

Periodic measurement using portable MCSA equipment has a lower upfront cost per asset. The tradeoff is that coverage depends on measurement interval. If a fault develops and progresses between inspection visits, it will not appear in your data until the next scheduled measurement. Intermittent faults that appear only under specific load conditions are particularly likely to be missed.

Periodic approaches also require either trained in-house personnel or a contracted service program. The quality of the result depends heavily on the technician's ability to set up the measurement correctly and interpret the output.

Continuous online monitoring

Permanently installed sensors eliminate the sampling gap. The system captures every operating condition, including faults that appear only at specific loads or temperatures. Automated alarming means the team is notified when a fault signature crosses a threshold, rather than waiting for the next scheduled inspection.

The cost per monitoring point is higher upfront. For assets where a single failure costs more than the monitoring installation, continuous monitoring is typically cost-justified on the first event it catches.

Hybrid deployment

Most facilities land on a hybrid model: continuous monitoring on the top 20 to 30 percent of assets by criticality, and periodic route-based inspection on the remainder. This concentrates capital spend where the risk is highest while maintaining visibility across the broader asset base.

Start by identifying your top 10 to 15 most critical motors. Deploy continuous monitoring there first. Run a periodic inspection program on semi-critical assets, and use the failure data from that program to refine the criticality classification over time.

Step 4: Evaluate Sensor Integration Requirements

Permanently installed electrical monitoring sensors live inside or adjacent to the motor control center. The integration questions are practical: where does the hardware mount, what does it need to operate, and how does data move from the sensor to the monitoring platform.

Physical installation

Current transformers must be placed on the motor feed conductors inside the MCC. This typically requires panel access and some modification to existing wiring arrangements. Understand whether the installation requires a panel outage, how long that outage needs to be, and whether the work requires a licensed electrician or can be performed by maintenance staff.

Some systems include voltage sensing in addition to current sensing. Voltage sensing adds installation complexity but enables true ESA capability, which is required for separating supply-side problems from motor-side faults.

Power and communications

The monitoring electronics require a power source. Confirm whether the system draws from the monitored panel, requires a separate 120V feed, or uses another supply. Each option has different installation implications.

Data must reach the monitoring platform somehow. Common communication paths include 4-20 mA analog output to existing PLCs, Modbus RTU or TCP over existing plant Ethernet, dedicated wireless radio, or cellular. Cellular-connected systems require no plant network integration, which simplifies installation and avoids IT firewall and security review.

Questions to ask every vendor

Get answers to each of these in writing before committing to a vendor. Installation cost and complexity are frequently underestimated in initial evaluations.

Step 5: Evaluate the Analytics

Two systems can both perform MCSA and produce completely different levels of usable output. The hardware is similar. The analytics layer is where the practical difference lies.

Frequency calculation: automatic or manual?

MCSA works by identifying fault-characteristic frequencies in the current spectrum. The relevant frequencies depend on motor nameplate parameters: pole count, slip frequency, number of rotor bars, bearing geometry. A system that calculates these frequencies automatically from nameplate data entered at setup will identify the correct spectral lines without requiring the user to derive them. A system that requires manual frequency entry puts the burden on the operator to know what to look for and where.

For teams without a dedicated vibration or electrical analyst, automatic frequency calculation is close to a requirement.

Trending versus threshold alarming

A system that trends fault frequency amplitude over time and alarms when amplitude increases significantly from a baseline is more sensitive to developing faults than one that only alarms when an absolute level is exceeded. Many faults begin at low amplitude and progress gradually. Trend-based alarming catches them earlier.

Ask the vendor specifically: does the system alarm on absolute amplitude, on rate of change from baseline, or both?

Supply-side versus motor-side separation

Supply voltage imbalance and motor winding faults can produce overlapping signatures in the current spectrum. Separating them accurately requires voltage measurement in addition to current measurement, which is the distinguishing capability of ESA versus basic MCSA. If you are trying to determine whether a problem is upstream power quality or internal motor fault, a current-only system cannot give you a definitive answer.

Diagnosis versus raw data

Some systems display a processed fault severity score or plain-language diagnosis. Others display the raw spectrum and expect the user to interpret it. For teams without a trained analyst reviewing outputs regularly, a system that provides human-readable diagnosis is significantly more likely to generate action when a fault is detected.

Step 6: Consider Integration with Your CMMS or EAM

An alarm that generates a work order automatically is more likely to be acted on than one that sends an email. An email creates a manual step: someone reads it, decides whether it warrants action, logs into a separate system, and creates the work order. Each step is an opportunity for the alert to be deferred or lost.

Evaluate whether the monitoring platform connects to your existing maintenance management system via API or a standard protocol. Most enterprise CMMS and EAM systems support REST API or OPC-UA integration. Some monitoring vendors offer native connectors to common platforms.

Ask specifically: what does the system do when an alarm triggers, beyond sending a notification? The answer reveals how much manual process the monitoring system requires to actually generate maintenance action.

If integration is not available, document the manual workflow that bridges the gap. The effort required to maintain that workflow consistently over months and years is a real cost that belongs in the evaluation.

Common Evaluation Mistakes

Most purchasing errors in electrical monitoring follow a small set of patterns. Recognizing them before the evaluation starts avoids the most common outcomes: overspending on the wrong technology or underinvesting in what matters most.

See Continuous Electrical Monitoring in Practice

Tractian's condition monitoring platform combines continuous electrical and mechanical monitoring with automated fault diagnosis, integrated work order generation, and support from Tractian reliability engineers.

See Electrical Monitoring
ElectricalMonitoring.com is an educational resource published by Tractian. Content is intended for informational purposes only.
Specific fault detection outcomes depend on installation quality, asset condition, and operating environment. Consult a qualified reliability engineer before making purchasing decisions.