An industrial electrical installation almost never fails all at once. It degrades over months, sometimes years, and only the final minute of that process looks like what the plant floor calls “a failure.” A bolt that lost its torque, a lug that worked loose after thousands of thermal cycles, a contact pitted by repeated arcing: each one raises its resistance little by little, and that resistance turns into heat. Industrial electrical maintenance is, at bottom, the work of finding those signals while they are still signals and not a burned-out panel at three in the morning.
At De La Mar we perform specialized installations and electrical maintenance — preventive, corrective, and under a recurring program — in plants, distribution centers, industrial buildings, and the associated civil works. We work to NOM-001-SEDE for what gets installed and to NOM-029-STPS for how we intervene on what is already energized. This page explains what we do and, above all, on what criteria.
The real cost of a failure
When an electrical failure gets evaluated, the most common mistake is to measure it by what the repair costs. Replacing a breaker, redoing a connection, or swapping a damaged run of cable is usually cheap in absolute terms. What is not cheap is the time the line sat idle, the product spoiled in process, the shift that was paid for and produced nothing, the shipment that went out late, and what that means to the customer waiting for it. In nearly any industrial operation, an hour of unplanned downtime costs an order of magnitude more than the component that caused it.
That asymmetry is what carries the entire argument for preventive work. A loose connection starts heating up long before it fails: resistance climbs, dissipation climbs with the square of the current, the point degrades, the nearby insulation ages faster, and the process accelerates. Between the moment that point becomes detectable and the moment it becomes a shutdown, weeks or months go by. That window is the whole business of preventive maintenance. We are not avoiding a small repair; we are moving that repair out of the worst possible moment — plant running, no spare on hand, no crew available — and into a scheduled Sunday outage, with the material on site and the crew ready.
Thermography and diagnosis
A thermal camera reads the infrared radiation surfaces emit and translates it into a temperature image. In an electrical installation it works because almost everything that is about to fail heats up before it fails, and that heat shows up in the image long before anyone would feel it by hand or the insulation begins to break down.
Here is the point most often misunderstood, and the one everything else depends on: thermography only works with the installation energized and under load. A de-energized panel sits at ambient temperature, and in it a defective connection and a sound one look exactly the same. The heat is produced by current passing through the resistance of the defect; with no current there is no heat, and with a load far below normal the heat is too low to stand out against the background. That is why a thermographic inspection is scheduled while the plant is working, with the lines near their normal operating regime, and not during the year-end shutdown “because we can get access.” A thermogram taken with no load is a handsome document with no information in it.
Nor does reading one come down to “it’s hot” or “it isn’t.” What gets interpreted are differentials: the temperature of a point against the matching phases of the same circuit, against equivalent components under equivalent load, and against ambient. A thermal imbalance across the three phases of one breaker tells you more than any absolute value. That is why the load at the moment of capture is always recorded: a hot spot at 40 % load is far more serious than the same reading at 100 %, because at full load it will be considerably worse.
The camera tells you where to look; the diagnosis is made by whoever understands the installation. The inspection is completed with insulation resistance measurement, torqueing to specified values, review and coordination of protective devices, load balancing across phases, and verification of the grounding system. A thermographic finding with no root cause identified is a photograph, not an assessment.
Preventive, corrective, and a recurring program
Corrective work always arrives at the worst possible moment, and it isn’t bad luck: it’s statistics. Components fail when they are most stressed, and they are most stressed when the plant is running flat out, on the hottest day of the year, in the middle of peak season. The failure does not pick its moment at random; it picks precisely the moment when it costs the most. A program that only reacts has accepted by design that every intervention will happen under the most expensive conditions available.
A recurring program sees something no isolated visit can see: the trend. A point that rose several degrees between two consecutive inspections at the same load is information that only exists if there is a history to compare against. A single measurement gives you a value; a series gives you a slope, and the slope is what lets you decide whether something gets attention this month or can wait for the next outage. The value of sustained thermography is not in the first inspection — it is in the third.
Frequency is not declared in advance, it is determined. It is set by the criticality of the circuit — what stops if that stops — by the age and condition of the installation, by the load regime and its cycles, and by the environment the panel lives in. And the environment is decisive: dust forms layers that block heat dissipation and, if it is conductive or hygroscopic, creates leakage paths between phases; humidity and condensation oxidize contacts and degrade insulation; vibration from motors, compressors, and forklift traffic loosens hardware over time. Load cycles make the copper expand and contract, and every cycle relaxes the connection’s clamping force a little. A connection that loosens heats up, and a connection that heats up loosens further. It is a feedback loop, and it ends in an arc. A panel in a clean, climate-controlled room and a panel in a grinding area under cyclic load have nothing in common, even if the one-line diagram is identical.
Lockout and tagout
NOM-029-STPS governs safety in electrical maintenance work and requires the control of hazardous energy: lockout/tagout, LOTO. It is not paperwork. It is the procedure by which a panel is worked on without killing anyone.
Doing it properly means identifying every energy source feeding the equipment, not just the obvious one: back-feeds from another panel, a capacitor bank still charged, an uninterruptible supply, stored mechanical, pneumatic, or hydraulic energy. Then the disconnecting means is opened, the physical lock and the tag naming whoever placed it go on, residual energy is discharged, and absence of voltage is verified by measurement — with an instrument proven against a known source both before and after the test. The lock is removed only by the person who installed it. Everything else — trained personnel, protective equipment rated for the arc flash incident energy, insulated tools, work permits, and a delimited work area — is what makes that procedure real rather than a signature in a logbook.
There is an obvious tension between this and the section above: thermography demands an energized installation. It is resolved with procedure, not by ignoring it. Energized work is limited to what genuinely cannot be done any other way, executed with approach boundaries, PPE, and protocol, and where the installation allows it we use infrared inspection windows so the thermogram can be taken with the enclosure closed. Diagnose hot and repair cold is not a contradiction; it is the correct order.
Specialized installations
Beyond maintenance, we build the installations that maintenance later looks after: service entrances and substations, main switchboards and motor control centers, feeders and raceways, industrial lighting, grounding and lightning protection systems, power circuits for new machinery, and capacity expansions for plants that grew while the installation stayed the same size.
There is nearly always civil work involved — foundations, pull boxes, trenches, supports, steel structure — and we do it ourselves, which is what removes the coordination problem between two contractors blaming each other for the schedule. We operate out of Pachuca, Guadalajara, and León, with work across all of Mexico.
What we hand over at the end of an intervention is not an “everything looks fine” report. It is a documented assessment: thermograms with their associated load, measurements, findings ranked by criticality, and a clear line between what has to be addressed now and what can wait for the next scheduled outage. With that, a maintenance supervisor can defend a budget and a plant manager can decide when the line stops, instead of finding out once it has stopped on its own.
