Cutting Fluid Concentration Monitoring: Measurement to Safe Top-Up

When planning to introduce cutting fluid concentration monitoring, the first question is not which sensor to buy. Can the plant name the water-miscible fluid in each CNC, its supplier-approved working range, its product-specific refractometer factor, a representative sampling point, and the person authorized to correct a deviation? A dashboard cannot resolve those questions. This guide is for machining and grinding plants in Thailand that want to move from disciplined manual checks to continuous monitoring and, where justified, controlled automatic top-up. The subject is the working fluid in the cutting-fluid sump, not hydraulic-oil condition or generic machine uptime.

Establish four conditions before managing coolant concentration

A water-miscible metalworking fluid starts as concentrate mixed with water. Working fluid travels from the machine sump to the cutting point and returns with chips. Some water evaporates; some working fluid leaves on chips and parts. An identical fall in sump level can therefore require different makeup mixes. A level switch detects volume, not concentration. Master Fluid Solutions identifies current and target concentration, current volume, and full volume as distinct inputs for calculating makeup fluid.

For each machine, record four sets of facts: (1) fluid product, supplier data-sheet revision, approved working range and measurement method; (2) product-specific refractometer factor and the mixing water; (3) sampling location, operating state, and treatment of chips, foam and tramp oil; and (4) who approves a correction and how the action is logged. A percentage quoted in another manufacturer’s case study is not a setting for your fluid. The approved range depends on the specific product, workpiece material and operation.

Define the displayed number as carefully as the sensor. Keep the raw Brix reading separate from calculated working concentration. A screen that says only “concentration 5” leaves the operator guessing whether it means 5 Brix or 5%. Store timestamp, machine ID, fluid product ID, sample point, operator or instrument ID, raw reading, quality flag, applied factor, calculated concentration, and fluid temperature in one record. This is the minimum context for later investigation.

Refractometer readings need the fluid’s Brix factor

A handheld refractometer measures how light bends through a sample. For water-miscible metalworking fluid, the reading often requires a product-specific multiplier. The calculation is working concentration (%) = refractometer reading (Brix) × product refractometer factor. For illustration, 4.5 Brix with a factor of 2.0 yields 9.0%. Whether 9.0% is acceptable must then be judged against the approved data for that particular fluid. Neither the example factor nor the result is a general coolant target.

Master Fluid Solutions’ Technical Bulletin 110 explains that factors differ among products and do not rank fluid quality. Blaser Swisslube likewise explains the product-specific factor and zeroing a handheld refractometer with water. In the plant master data, keep the factor, its source document and revision, registration date, and approver together. Review the factor whenever fluid product or supplier instructions change. Unrestricted editing of a factor on an operator screen can turn every converted value into a misleading figure.

Choose a consistent zero-setting procedure with the supplier: whether to use the actual mixing water or the specified high-quality water. Document it. Dirty prisms, temperature differences, unstable emulsion and tramp oil can make a reading unclear. If the refractometer boundary is diffuse, do not force a precise number: collect a clean sample and inspect the fluid condition. Continuous refractive-index measurements still need the correct fluid factor and periodic comparison with an independent handheld reading.

Can level, pH or conductivity stand in for concentration?

No. Level describes volume; pH describes acid-base condition; conductivity responds to electrically conducting constituents. They are valuable complementary signals, but none alone replaces product-approved concentration measurement. A low level does not reveal how much was evaporation versus carryout. A pH trend cannot simply be converted into concentrate percentage. Quaker Houghton’s QH FLUIDCONTROL 100 data sheet lists refractive-index concentration, pH, conductivity and temperature as distinct measurements.

Tramp oil is oil from other systems that has entered the cutting fluid. It can blur a refractometer’s boundary, but the volume of tramp oil is not the working-fluid concentration. Monitoring oil within a hydraulic power unit is a different application; see our hydraulic-oil condition monitoring guide. Even if both applications appear on one plant dashboard, separate their fluid IDs, units, limits and response owners.

Cutting Fluid Concentration Monitoring: Measurement to Safe Top-Up - figure 1

Sampling and pretreatment determine whether a monitoring system is trustworthy

Before choosing an online instrument, mark its proposed sampling point on a piping diagram. Tramp oil can accumulate near the sump surface; chips and sludge near the bottom. A return line may contain foam or entrained air. A point immediately after the makeup inlet may show a locally concentrated or diluted stream before mixing. A dead leg may retain old fluid while the machine is stopped. Compare at least two feasible points under real operating conditions rather than assuming the neatest pipe is representative.

During a site trial, take a manual sump sample and an online-loop sample at the same time and test both by one agreed procedure. Compare machining, immediately after top-up, shift start, and recovery after a long stop. Record why the results differ. A few matching readings do not prove representativeness over normal variation in chips, foam, level and cleaning. First ask whether two samples represent the same mixture before blaming sensor accuracy.

An online sensor may require a strainer or filter, but finer filtration is not automatically better. Clogging can reduce flow until the instrument is measuring stale liquid. Pretreatment that removes oil can change the composition relative to an untreated hand sample. Specify filter type, flow or differential-pressure indication, cleaning interval, drain route, sample-line residual volume and backflow prevention. A measurement received while sample flow is absent should carry an invalid-quality flag, even if the sensor continues to output a number.

Include maintenance and environmental safety in the sampling design: accessible isolation valves, a leak tray, clear supply and return labels, a cleaning route, and collected waste fluid. Check with the machine builder whether tapping the coolant circuit affects warranty or existing safety functions. Connecting fluid piping can affect the machine more than simply reading its uptime signal.

Move through manual checks, continuous monitoring and automatic top-up

The first stage is a repeatable manual check. Document fluid product, factor, supplier-approved range, refractometer zero-setting, sampling point, frequency and corrective owner for each machine. Blaser describes checking concentration whenever a tank is replenished and recording it at least weekly. The actual plant frequency should follow the fluid supplier’s instructions and the speed of change in its machines. Keep an independent manual check after installing online instruments for calibration and fault investigation.

The second stage is continuous monitoring. Combine refractive-index concentration with level and temperature if the decision requires them. Show the last valid measurement time, machine and fluid IDs, sample-flow status and top-up events beside the calculated concentration. Do not keep displaying a stale last value as if it were current when the network fails or the sensor fouls. Direct alarms to the people who can act. An email to quality alone does not correct the sump if the machine operator cannot see it.

The third stage is controlled top-up: water and concentrate mixing, pumps, valves, flow proof and level limits. Blaser describes equipment that monitors each machine’s concentration, temperature and level, tops up individual tanks, and records its actions. Quaker Houghton’s published product sheet describes separate outputs for water, concentrate and additives, with a time-limited manual override. These are implementation examples, not a requirement to automate every line from day one. A semi-automatic process requiring operator approval may fit the first deployment better.

Measuring a deviation and correcting it safely are separate capabilities. A replenishment mix that is too strong can create over-concentration even when the sensor works. Establish evaporation and carryout patterns from plant records, then agree the makeup rules with the fluid supplier. Low and high concentration require different actions. Do not pour neat concentrate straight into the sump against the supplier’s mixing procedure; do not add water to an already full sump without considering overflow and waste handling.

Cutting Fluid Concentration Monitoring: Measurement to Safe Top-Up - figure 2

Interlocks and manual operation for automatic coolant replenishment

Write the fault behavior before the normal sequence. What happens when level is low but concentration measurement is invalid? What if the pump runs without verified flow? What if the level continues rising after the valve-close command? Can local staff stop replenishment during a network outage? The answers belong in the control specification, not in an operator’s improvised response.

Set an upper limit per addition, a cumulative limit over a defined period, a sump high-level stop, dry-run detection, valve fault response, maximum age of a valid concentration value, and controlled manual-mode access. A fault should stop dosing safely, with a local manual procedure. Existing machine interlocks and emergency-stop functions must not depend on a cloud decision. Include leaks, slippery floors, operator contact and ingress into electrical cabinets in the assessment. Keep the current safety data sheet and machine instructions available at the point of work.

Manual override should not mean indefinite forced operation. Log reason, operator, time, permitted amount and expiry; verify actual fluid concentration before returning to automatic mode. Quaker Houghton’s product sheet documents a time-limited override, but your site’s authorization and limits need their own risk review. Where multiple products are used, prevent wrong-fluid connections and identify every line. A shared supply header also needs a documented changeover, residual-fluid and flushing procedure.

What to put in a cutting-fluid monitoring RFP

An RFP should elicit comparable answers from vendors, not just list sensor model numbers. Give each bidder machine IDs, sump capacities, fluid products, supplier-approved ranges, Brix factors, mixing-water conditions, existing piping and power, and network constraints. Identify facts to be confirmed by a paid site survey instead of letting each bidder make hidden assumptions.

For measurement, request the sampling diagram, pretreatment, principle, measurement cycle, calibration method, comparison with the handheld refractometer, and explicit handling of stops, low flow, foam and fouling. Ask where product-specific factors are stored, who can change them, and how an old factor is prevented from remaining active after a product change. If pH or conductivity is included, state the operational decision tied to each. More channels alone do not establish value.

For top-up, specify preparation of makeup mix, water and concentrate piping, backflow prevention, pumps, valves, proof of actual flow, high-level stop, dose ceilings, fault stop, manual switch and restoration. Ask for machine-builder approval conditions, local spares, in-country calibration support and planned downtime for maintenance. For data, define time zone, missing-data representation, machine and fluid IDs, factor revisions, dose events, audit logs, CSV or API export, retention and access rights. Involve plant IT if a cloud connection is proposed.

Set acceptance criteria when issuing the RFP. Measurement tolerance and response time cannot be copied from this article as universal values; agree them for the actual fluid, sampling method and process risk. Name the reference instrument, number of sample points and repeats, fluid-change conditions, and pass/fail calculation. Without this, the vendor may show catalogue accuracy while the plant questions real-fluid disagreement during acceptance.

Separate FAT function tests from SAT tests with the real coolant

Factory acceptance testing (FAT) checks functions before shipment using permitted test fluid or simulated inputs. Verify factor calculations, changing fluid master data, missing and stale data, alarms, override, dose limits, access permissions, logs and behavior on communication loss. Observe the actual pump and valve outputs as well as the HMI. Record input, expected result, observed result, verdict and responsible person. Define safe containment for any test discharge.

Site acceptance testing (SAT) uses the real machine and working fluid. Check installation and leaks, sample representativeness, filter clogging, level input, independent refractometer comparison, mixing time after top-up, variations across shifts and restart after power loss. Use the sampling and manual method agreed with the fluid supplier as the comparison basis. If readings differ, investigate sampling point and circulation state as well as calibration. Schedule simulated faults and any machine stop with production and safety staff.

“A number appears on screen” is not a sufficient SAT result. Verify that (1) the number has the right meaning, (2) invalid readings are visibly invalid, (3) dosing stops as designed and local staff can intervene, (4) events can be reconstructed from logs, and (5) maintenance staff can clean filters, check the reference instrument and change the fluid master safely. Transfer operation only after those points are demonstrated.

Cutting Fluid Concentration Monitoring: Measurement to Safe Top-Up - figure 3

Calculate costs and return from your own factory baseline

Vendor cases demonstrate that an arrangement is possible. Quaker Houghton describes one North American aerospace application with six CNC machines and fluid-level and concentration control. Its reported workload and outcome apply to that customer and fluid. They do not establish a savings rate or payback period for a different factory in Thailand. Start with observed work and spending in your own plant.

An illustrative formula is: annual benefit = verified reduction in manual measuring and replenishment hours × actual loaded labour cost + verified changes in fluid, waste, defects and downtime attributable to the project. Annual net benefit = annual benefit − calibration, filters, consumables, connectivity, service and additional operating cost. Simple payback = initial cost ÷ annual net benefit, only when the latter is positive. These are evaluation equations, not forecasts or performance guarantees. Avoid counting the same freed labour both as saved payroll and as additional production unless the use of that time is evidenced.

Split initial cost into site survey, sample and return piping, sensor and pretreatment, control panel, top-up equipment, installation, data integration, FAT and SAT, and training. Include warranty approval and production-stop coordination when applicable. Recurring cost includes reference instruments, calibration material, filters, waste treatment, power and support. Baseline one machine or one circuit for manual-check time, replenishment events, out-of-range duration, fluid use, filter replacement, interventions and quality outcomes. Compare a later pilot period with identical definitions. Tool life and uptime have other causes, so document changes in tools, material and cutting conditions. Our CNC production preparation guide explains how to organize such machining conditions.

Make the procedure workable across shifts and languages

Choose one pilot machine through a site survey

Do not automatically choose the machine with the most problems. Severe contamination or complex piping can obscure whether the monitoring method works. Compare candidates by stable fluid product, safe sump access, existing manual records and trackable top-up volume. Record machine identification, sump capacity, fill and drain points, pump operating state, oil skimmer, chip type, existing filtration and available stop windows. Trace undocumented piping on the actual machine. Even two machines in one plant can have different water entry and chip removal, so a sample point cannot simply be copied.

Preserve when, which fluid, and which measurement state

Put machine stops, circulation restarts, concentrate-container changes and filter cleaning on the same time axis as concentration. Agree whether raw timestamps are stored in ICT or UTC and how they appear on screen. Never silently replace raw readings; flag a correction and record its reason. Quality states may include circulation stopped, mixing after replenishment, inadequate sample flow, blurred handheld reading and factor revision unconfirmed. Set which states trigger an alarm and which are only stored for later review. The mixing wait must be verified on the actual circuit.

Treat fluid changes and calibration as normal work

A fluid change can follow a new workpiece, purchasing decision or supplier revision. Its procedure covers old-fluid drainage or residue, initial mix, pipe and sensor cleaning, new factor and range, approval, and first independent handheld check. Disable automatic dosing during changeover so it cannot run under the old recipe. Specify the reference instrument, its own maintenance, and which past readings must be reviewed if a tolerance is exceeded. After online sensor cleaning, confirm stable sample flow and agreement with the independent reading before dosing resumes. Link maintenance records with concentration history to distinguish fluid drift from sensor fouling.

Decide whether the pilot earns a wider rollout

At the pilot review, production, maintenance, quality, EHS, IT, fluid supplier and system vendor should examine representativeness, manual-versus-online differences, invalid readings, post-top-up mixing time, interventions and cleaning effort. Keep failed and interrupted tests in the evidence. Check whether night and holiday shifts can follow the manual response. Classify the outcome as adopt, continue with a named correction and retest date, or stop because measurement meaning or safe dosing is unproven. For rollout, recheck each machine’s product, volume, circulation, sample route and load. Standardize data fields, alarm codes and test procedure, while documenting the differences in fluid and piping per machine.

Compare the actual boundary of vendor quotations

Do not compare a monitoring-only proposal as if it included replenishment control. Ask each bidder to price survey, sample piping, filters, cleaning valves, control cabinet, installation, downtime coordination, commissioning, calibration, procedures and training against the same required/optional/excluded scope. Recurring cost includes filter changes, recipe changes after fluid replacement, local response time, spares and software maintenance. If operators must keep doing every manual check because they distrust the instrument, promised labour savings do not materialize. Ask what training and adjustment are included after acceptance and state assumptions such as imported-part pricing.

Specify what the alarm recipient must do

For a low or high reading, first confirm data validity, circulation, and whether the machine is still mixing after top-up. Compare with the approved product range and previous reading, then use the handheld reference before choosing a correction. Display first checks and escalation contact. Separate an early-warning management band from the supplier’s formal operating range. Log alarm acknowledgement, repeat reading, approval, actual dose and post-mixing check. A vanished alarm alone does not prove a proper correction. Give no-flow, communication loss, missing factor and overdue calibration distinct data-quality alarms; they call for manual checking, not automatic dosing.

Judge whether alarm handling works by time to first check, the share of alarms independently rechecked, and the share with a documented post-action result. Set local targets from the plant’s shift pattern and required response time rather than borrowing another site’s number. Before widening an alarm threshold merely to reduce alerts, look for invalid data, poor sampling or unnecessary notifications during the verified mixing wait.

Assign named roles by event: shift-start manual check, alarm review, top-up approval, concentrate replacement, fluid-product change, filter cleaning, calibration, monthly trend review and response to a machine stop. In a Thai plant reporting to a Japanese head office, translate more than the dashboard headline. Use the same equipment ID, fluid ID and alarm code in Thai and Japanese procedures, followed by first checks, prohibited actions and escalation contacts.

At a monthly review, ask why one machine needs unusually frequent replenishment, whether a product change altered readings, and whether clogged filters created missing data. A CSV is not a meaningful history if product, factor and calibration revision are absent. Standardize machine and fluid identifiers before expanding. Plan for ordinary operation: a sensor that works only after an unusually thorough pilot cleanout is not an operating solution. Contract for local spares, realistic cleaning time and the post-maintenance check. The real workflow is detecting bad data and assigning its recovery, not adding more charts.

Frequently asked questions

Can a Brix reading be used directly as cutting-fluid concentration?

Some products have a factor close to one, but that cannot be generalized. Check the current product data sheet, multiply by its factor, and follow any product-specific measurement exception. Reconfirm the factor after a fluid change.

Is a level sensor enough for automatic coolant top-up?

It detects falling level, not the balance of evaporation and carryout. An automatic design needs actual concentration, volume, makeup composition and a post-mixing recheck. Level remains essential to avoid overflow and pump dry-running, but it is not a substitute concentration signal.

Do pH and conductivity remove the need for refractive measurement?

They describe different properties. Choose signals according to the fluid supplier’s method and the plant decision. If measurements conflict, check sampling, calibration, tramp oil and contamination rather than declaring one channel automatically correct.

Where should an estimate for a monitoring system begin?

Survey the machines, fluid products, sump and sample points, water and concentrate feeds, communications and current manual work. Clarify whether the scope includes monitoring only or automatic dosing, how FAT and SAT will pass, and ongoing maintenance. Device price alone misses piping, calibration and operating differences.

How should a blurred refractometer boundary be treated?

Flag the reading as uncertain and record sample condition. Review sampling and investigate tramp oil or instability. If pretreatment removes oil before online measurement, compare against a defined manual reference to ensure the sample still represents the working fluid.

Conclusion

The sequence for introducing concentration monitoring is to confirm the product and its approved range, establish a handheld reference, choose representative sampling, manage the quality of online measurements, design dosing safety, and prove the complete arrangement at FAT and SAT. Level, pH, conductivity and tramp oil provide useful context but do not replace a concentration value derived by the product-approved method. Judge value against your own plant baseline rather than another vendor customer’s savings claim.

If your Thailand plant is standardizing checks for water-miscible cutting fluid, identifying sample points on existing CNC machines, or preparing an RFP and acceptance plan for replenishment equipment, you can discuss the scope with TOMAS TECH while options are still being compared. The fluid data sheet and machine drawing are useful starting inputs.