To introduce a hydraulic oil leak detection system in a factory, first determine which hose, fitting or power unit could leak, where the oil would travel and who would act on an alarm. Comparing sensor catalogues first will not solve a missing collection path, a sensor that reacts to the wrong liquid or an undefined shutdown boundary. This guide is for presses, moulding machines and other equipment in Thai factories. It addresses external oil escaping from hoses, joints and drain pans, plus PLC integration, procurement and acceptance. It has a different scope from hydraulic oil condition monitoring, which examines particles, water and temperature inside the circuit.
Define what counts as a leak
Here, a leak means hydraulic fluid escaping from piping, hoses, fittings, seals, valves or pumps to an external or unintended location. Contamination, water ingress and temperature change inside the fluid are different problems. A reservoir level drop alone does not prove external leakage: temperature, cylinder position, topping up, oil replacement and internal leakage can alter level. A small external leak may also take time to trigger a tank-level alarm. Neither “low level equals an external leak” nor “normal level equals no leak” is a sound rule.
Leaks take different forms: drops from a coupling, wetness on a hose, spray from a moving assembly, oil flowing across a floor or accumulation inside a guarded enclosure. Each requires a different collection or observation point. A fine high-pressure jet may be difficult to see. Parker warns against searching for it by touch because hydraulic fluid can penetrate skin; pressure can remain after the pump stops. OSHA identifies hose, fitting and line deterioration as inspection concerns. A detection system should help maintenance staff recognise an abnormal condition without approaching a pressurised hazard. See Parker’s hose safety guidance and OSHA’s hydraulic systems bulletin.
“Detect leaks early” is too vague to specify a system. A useful goal might be to notify maintenance before oil reaches product, keep oil out of a walkway or identify accumulation in the power-unit pan before the next inspection. The response depends on the hazard to people and equipment, how oil spreads, the fluid and hazards created by stopping. Sensor selection alone cannot justify one universal automatic-stop rule.
Survey equipment, fluid and flow paths before buying hardware
Start with an asset register. Trace the power unit, pump, manifold, fixed pipes, moving hoses, cylinder connections, return lines and floor drains on the actual machine. Maintenance changes can make drawings incomplete. Give monitoring zones distinct IDs, such as “below the power unit”, “moving-hose outlet” and “below the valve cabinet”. Photographs should show their direction and exclusion zones. Check whether installation, cleaning and replacement can be done safely. Follow gravity, but also look at guards, cable ducts, pipe slopes, spray and wash water.
Record the actual fluid manufacturer and product, replacement fluids, cleaning agents, cutting fluid, water and condensation. A hydrocarbon detection cable is not automatically compatible with every hydraulic fluid. Chemelex (Raychem TraceTek)’s TT5000 literature mainly describes liquid hydrocarbon fuels; it does not establish performance with the particular oil in your factory. Request the candidate’s compatibility information and a test with the real fluid. The Chemelex (Raychem TraceTek) sensing application guide illustrates the distinction between point probes and sensing cable.
Put previous leaks, hose replacements, floor cleaning, oil top-ups, operating conditions and downtime on one timeline. Separate events during pressure, after repair and while idle. If records are missing, record an inspection hypothesis instead of inventing a leak rate. Establish whether escaping oil enters a recoverable pan or runs into a drain. A sensor on the floor has no value if the liquid flows elsewhere.
Choose the detection principle for each leak location
For each candidate, distinguish what can reach the sensor from what the sensor directly detects. Fluid viscosity, temperature, dirt, washdown, mounting height, required liquid volume and maintenance all affect suitability. Specifications narrow the shortlist; approval should follow a test using the real fluid in a representative installation.
| Method | Suitable location and observation | What to verify |
|---|---|---|
| Drain pan and inspection point | Oil directed below a power unit or manifold | Slope, lowest point, drainage, cleaning, other liquids |
| Liquid point sensor | Oil reaching a small collection area | Fluid compatibility, detection volume, dirt, bubbles, diagnostics |
| Liquid sensing cable | Longer runs beneath pipes or several risk points | Fluid compatibility, routing, damage, location, restoration |
| Reservoir level | Supporting evidence of a substantial oil-volume change | Temperature, cylinder position, topping up, other causes |
| Camera or inspection route | Spray or wetness beyond a pan | Light, obstruction, dirt, safe observation, consistent records |
ifm illustrates leakage monitoring by placing a capacitive point-level sensor in an overflow basin. That example is not a promise of performance with every hydraulic oil or cleaning regime. Confirm the exact model, medium and installation, then test a sample. See ifm’s LI applications. Chemelex (Raychem TraceTek)’s guide similarly helps compare probes with cable over risk areas; its stated fuel applications still require separate qualification for hydraulic oil.
Even a sensitive device cannot detect liquid that never reaches it. Verify a low point in the pan, whether drops run along a guard edge and whether moving hoses can strike or abrade the device. Wrapping a sensing cable around a moving hose can complicate bending, replacement and safety. A fixed collection point may be easier to maintain, but spray beyond the pan needs guarding and another observation method.
Test both false alarms and missed leaks where water, coolant or cleaning liquid enters. A device that ignores water may still fail to detect the specified oil. Water may also carry the oil past a dry sensor. Include actual fluids, mixtures, cleaning and ageing in FAT and SAT instead of extrapolating from a marketing label.
Monitor hydraulic hoses from outside the hazard zone
Hoses and fittings experience movement, heat, vibration, abrasion, bending and replacement. Do not ask a worker to feel a pressurised hose, even through gloves. Parker describes near-invisible pinhole jets, injection injury and residual pressure after shutdown. An OSHA accident record describes a worker whose finger was injected with hydraulic fluid while checking a leak with a cloth.
The best location is therefore not simply the closest one. Balance capture of escaping oil with a safe maintenance distance. Follow the equipment OEM and local isolation and pressure-release procedures. Test ports should allow a function test without damaging a pressurised hose or reaching into the hazard area. At SAT, separately establish whether liquid released at that port follows the same path as a plausible leak.
Hose shields and spray guards can change where oil falls. Retake photographs and verify collection paths after modification. Record hose identity, fitting, abrasion, bulging, twisting, supports, pan condition, leak marks and replacement date. A detector does not repair a degraded hose. It links an event to inspection and replacement decisions; it does not replace routine physical inspection.
Connect power-unit leak detection to the PLC
Specify the meaning of each input before wiring: healthy, liquid detected, device fault, cable break, power loss, communication loss and maintenance bypass. If a device cannot diagnose some failures, state that limitation in the RFP and test what a broken wire looks like. For contact outputs, document the normal and alarm states and the consequences of loss of power in the I/O schedule. For analogue, IO-Link or network signals, define raw value, quality, timestamp, tag and stale-value behaviour.
Define the PLC state transitions rather than one rule saying “input ON, stop”. Include signal persistence, cleaning, alarm latching, investigation, manual reset and restart authorisation. Do not copy generic thresholds. A longer delay may suppress nuisance alarms while also delaying the response. Test both extremes. Keep the event time, asset and zone ID, input and alarm states, operator, findings, repair and reset approval in history.
A notification tag and a safety-related stop are different responsibilities. Pilz explains that the required reliability of a safety function follows risk assessment and is evaluated under standards such as ISO 13849 or IEC 62061. This article specifies no safety performance level. Assess hazardous motion, oil reaching an ignition source, residual pressure and any hazard created by stopping with the OEM and safety team. Then decide what to stop, how pressure is managed and what permits restart. See Pilz on functional safety.
Establish the alarm and shutdown boundary
It helps to discuss three kinds of response: inspection request, operating decision and safety-related stop. These are project discussion categories, not a universal standard. Whether a small amount in a pan warrants a full-line stop depends on leak position, spread, isolation, machine movement and effects on product and floors. A suspected high-pressure jet invokes the plant’s safety procedure and OEM guidance. A “normal” display must not override an operator’s observation of oil.
Build an alarm matrix with zone, machine state, detection condition, first responder, safe approach conditions, equipment affected, pressure check and restart approver. Add separate cases for missed detection, nuisance alarms, communication failure, power loss and bypass. Identify the recipient outside normal working hours and the drawing and asset ID they need. Wiping oil and clearing an alarm is not completion. Record source identification, repair, cleanup, pan restoration, trial run and release.
A shared power unit makes the boundary particularly important. Stopping one pump can change workpiece holding on another machine. Review actuators, check valves, accumulators, gravity loads and trapped pressure with the OEM. An ordinary leak sensor should not be assumed to substitute for a required safety circuit. Verify alarm functionality and machine safety functions under their respective ownership.
Put detection conditions, not just a parts list, in the RFP
“Sensors and PLC connection” allows a supplier to quote hardware without a comparable outcome. Attach site photographs, fluid details, zone map, normal coolant or wash-water entry, operating states, network, shutdown policy and maintenance permissions. Mark unverified drawing details as unverified and include site survey in the scope.
| RFP item | Required supplier response or evidence |
|---|---|
| Coverage | Assets, zones, target joints, hoses and pans; exclusions and reasons |
| Media | Compatibility data and sample tests for actual and replacement oils and cleaning agents |
| Detection | Point or cable principle, liquid flow path and installation drawing |
| Electrical | Power, contacts or protocol, failure and broken-wire outputs |
| Response | Test amount, injection point, equipment state and measurement method |
| Control | Owner of notification, operational stop and safety-related stop |
| Maintenance | Cleaning, replacement, retest, diagnostics and spares |
| Records | Asset and zone IDs, timestamps, states, work and reset history |
| Acceptance | FAT/SAT procedures, pass criteria, open issues and handover |
Do not promise a fixed volume or alarm time before the fluid and installation are defined. Derive targets from risk and the flow path, then describe the exact test. If real oil cannot safely be released at the machine, agree on separate media compatibility, flow-path and I/O tests. Record what that substitution cannot reproduce.
Compare complete cost: pans, oil recovery, wiring, cabinet space, PLC and HMI changes, communication, cleaning, training and spares. A few point sensors may require expensive pan fabrication; a long cable may require a shutdown to replace. Compare covered zones, maintainability and visibility of failure after clarifying exclusions in each quote.
Define responsibility among the buyer, machine OEM, sensor supplier and controls contractor. Who checks spare I/O capacity, backs up the PLC, approves the stop logic and schedules production access? Specify installation and wiring drawings, I/O lists, change records, test sheets, cleaning instructions and spares as deliverables. Record uncovered areas and their compensating inspections. Make any later change to hose routing or pan position trigger a review of the detection path. Do not copy the first installation blindly to identical machines: floor slope, covers, washdown, hose replacements and cable distance may differ. Standardise the survey and acceptance record as well as the part number.
FAT checks logic and fault paths
Factory acceptance testing checks the sensor, panel, PLC, HMI, notification and history before installation. Use the real sensor where possible and decide explicitly which tests use simulated inputs. Apply an approved test medium to a point sensor, then follow the contact, zone indication, alarm, log and reset. Ask the cable manufacturer for a safe test method. Handle test oil under plant requirements for ventilation, collection and disposal.
Test healthy, detected, broken wire, loss of power, communication failure, momentary input, sustained alarm, simultaneous zones, maintenance bypass and PLC restart. Acknowledge must not silently release a condition whose cause remains. Check timestamp order, duplicate records and whether zone names match terminals and drawings. Use labels that operators working in the plant can identify.
FAT cannot prove that oil will reach the sensor on the actual machine. Record the medium, conditions, pass and fail, exceptions, tests not run and SAT follow-up. Map every requirement number to a test number so site issues cannot disappear between stages.
SAT confirms real flow paths and maintenance actions
Site acceptance testing uses the installed pipes, pan, supply, network, PLC program, HMI and notification route. Agree safe shutdown, isolation, pressure release and fluid handling with the OEM and safety team. Do not loosen a pressurised connection to manufacture a leak. If an approved test port is available, introduce a controlled amount at representative positions and observe its path to the sensor. If a substitute liquid is used, record how viscosity and surface tension may change flow and response.
Put liquid arrival, alarm time, PLC input, notification and operator action on one sheet. If nothing is detected, follow the escaped liquid before changing sensitivity. Guard edges, fixings, floor slope and post-cleaning condition may require a revised path or sensor location. Where a realistic path cannot be reproduced, record the limitation and an inspection or guard as a compensating measure.
Test a shutdown interlock’s resulting safe state separately. A single pressure gauge reading does not establish absence of trapped pressure everywhere; use the OEM’s isolation and release procedure. Verify workpiece holding, moving axes, accumulators and shared supplies in accordance with the risk assessment. Qualified, authorised personnel must approve tests affecting machine safety functions. Leak-detector acceptance alone does not validate the functional safety of the whole machine.
Have both maintenance and operations personnel demonstrate alarm recognition, safe approach, first report, cleanup, sensor restoration and restart request. Check after-hours contacts, spare identification and current drawings. A system that cannot be maintained by the next shift is not ready to hand over.
After SAT on the first machine, run a trial before rolling out to the fleet. Observe normal cleaning, changeovers, hose replacement, recovery after power loss and maintenance coverage outside ordinary hours. Set the trial duration from actual use and work cycles rather than a universal number of days. An absence of alarms during a trial does not prove that a leak would be detected. Perform approved function tests at an appropriate interval. If a real leak occurs, compare the full chain from the source and liquid path to sensor, alarm, notification, repair and release. Mark test alarms clearly so they cannot be mistaken for real leaks in the event history.
For each subsequent machine, complete a difference checklist before copying the first design. Compare fluid brand, pipe and hose routes, floor and pan slope, wash chemicals, spare panel inputs, PLC revision, workpiece holding on shutdown and after-hours owner. A material difference should trigger the relevant repeat media, flow-path, I/O or stop test. Revisit the checklist after a guard or pan modification so the change does not create a newly uncovered area. The asset register should state coverage, exclusions, acceptance outcome and next function test, not merely that a sensor was added.
Assign alarm ownership and change rights at handover. For example, maintenance may clean and replace the detector, controls may govern PLC and HMI revisions, and the safety owner may review changes to stopping functions; adapt these roles to the plant. When a sensor model or setting changes, retain the old setting, reason, retest result and approver. Repeated false alarms must not lead to an indefinite, undocumented bypass. Record why a bypass began, its expiry, compensating inspection and verified removal. These are operating conditions that make detection lead to maintenance action, not claims about a particular product.
Measure completed response as well as alarms
After startup, classify real leaks, cleaning liquid, sensor contamination, broken wires, communication failures and tests. For a missed leak, trace the actual flow path and distinguish placement from sensitivity. For frequent false alarms, investigate other fluids and cleaning before relaxing a threshold. A silent alarm screen never removes the need to inspect potentially hazardous hoses.
Useful operating measures include time to acknowledge and inspect, source identification, repair completion, recurrence, device repair and bypasses left active. Set targets from the machine’s risk and staffing; this article supplies no universal time or savings percentage. A vendor’s downtime reduction claim from another plant cannot be copied into your business case without comparable baseline and operating conditions.
Our hydraulic oil condition-monitoring article covers particles, water and temperature inside the oil. Our factory cooling-water leak-detection guide offers a comparison for flow-path and alarm design, but water and oil require different media qualification and cleanup. Link leak events to asset, zone, work order and release record so future maintenance decisions have context.
Frequently asked questions
Can a reservoir level sensor replace a hydraulic leak sensor?
It supports oil-volume monitoring but seldom establishes the position or occurrence of a small external leak by itself. Temperature, cylinder position, top-up and internal leakage affect the reading. Identify where oil from high-risk joints can collect and decide whether direct detection is needed. Use distinct alarm names when combining level and leak signals.
Should a sensing cable be wrapped around a hydraulic hose?
There is no general rule to do so. Consider bending, replacement, abrasion and proximity to a pressurised hazard. First look for a fixed, safe collection point; evaluate guards and observation when spray may escape. Avoid touch checks or a leak test made by loosening a pressurised hose.
Does a PLC signal justify automatically stopping the power unit?
A signal can be wired to a PLC, but whether stopping is safe requires separate assessment. Include shared supply, holding loads, trapped pressure, moving axes and ignition sources. Separate a standard PLC notification from safety-related control. The OEM and safety owner should approve the stop, scope, reset and validation.
What belongs in a leak-detection RFP, FAT and SAT?
The RFP needs the fluid, zones, actual flow paths, competing liquids, outputs and diagnostics, plus response ownership. FAT tests the normal and fault path through sensor, PLC, HMI and log. SAT tests site collection, installation, operator action, cleaning and reset. Agree numeric pass criteria with the fluid and test conditions attached.
Can a fuel sensing cable be used unchanged for hydraulic oil?
Do not infer compatibility from “hydrocarbon” alone. Base stock, additives, viscosity, temperature, dirt and cleaning fluids differ. Verify the product’s media documentation, test samples with the oils in use and evaluate cleaning and recovery.
Conclusion
The practical order is to survey hazards and oil paths, qualify detection with the actual fluid, define signals and actions, write testable RFP conditions, and verify in FAT and SAT. Inspect fittings, hoses and pans on the machine and set alarm and shutdown boundaries through equipment-specific risk assessment. A detector does not remove the dangers of high-pressure oil. Run it alongside safe inspection, isolation, pressure release and hose replacement.
For a retrofit in a Thai factory, you can contact TOMAS TECH while you are still mapping assets, fluid, leak-prone joints and PLC connections. We can help turn those observations into monitoring zones and acceptance tests.
