loading

 Rika Sensor is a weather sensor manufacturer and environmental monitoring solution provider since 2010

The Necessity of Online Turbidity Monitoring in Liquid-Cooled CDU Systems: A Technical Analysis and Engineering Practice

Online Turbidity Monitoring in Liquid-Cooled CDU Systems Meta Description

LIQUID COOLING TECHNICAL REPORT

An engineering review of coolant cleanliness, suspended contamination, filtration performance, alarm strategy and online turbidity monitoring for liquid-cooled data center infrastructure.

Coolant Quality Monitoring CDU Systems Data Center Liquid Cooling Engineering Practice

Executive Summary

REPORT SUMMARY

In a liquid-cooled data center, coolant cleanliness is an operating condition that can affect the hydraulic reliability and long-term performance of coolant distribution units, manifolds, filters, heat exchangers and cold plates.

Turbidity is an optical indicator of suspended or dispersed matter in a liquid. A change in turbidity may indicate construction debris, corrosion products, precipitated chemicals, biological material, component wear particles or contamination introduced during maintenance.

Online turbidity monitoring provides continuous trend data and can identify coolant-condition changes that may be missed by periodic manual sampling. However, turbidity is not a direct measurement of total suspended solids, particle count, corrosion rate or microbial concentration.

Direct engineering answer:

Online turbidity monitoring is valuable in liquid-cooled CDU systems because narrow coolant passages and sensitive hydraulic components can be affected by suspended contamination. Continuous monitoring helps operators identify abnormal changes, evaluate filtration performance and determine when laboratory investigation or maintenance is required.

1. Scope of This Technical Report

This report focuses primarily on closed-loop, water-based Technology Cooling Systems, commonly referred to as TCS loops, operating between a coolant distribution unit and liquid-cooled IT equipment.

Typical applications include:

  • Single-phase direct-to-chip cold-plate cooling systems
  • Liquid-to-liquid CDU secondary loops
  • Rack manifolds and row manifolds
  • Water-based or inhibited-water coolant circuits
  • Compatible water-glycol coolant systems
  • High-density computing and AI data center cooling infrastructure

The principles discussed in this report should not automatically be applied to dielectric immersion fluids or proprietary coolant formulations. Operating limits must be confirmed with the CDU manufacturer, server or cold-plate supplier, coolant supplier and project engineering team.

2. What Is Turbidity in a Liquid Cooling System?

Turbidity describes the optical clarity of a liquid. When light passes through clean coolant, most of the light follows a predictable optical path. When suspended particles or other dispersed material are present, the light may be scattered, absorbed or attenuated.

The measured optical change is converted into a turbidity value, commonly expressed in Nephelometric Turbidity Units, or NTU.

ISO 7027-1 describes quantitative methods for determining turbidity using optical measurements. The standard distinguishes between measurements based on scattered radiation and measurements based on the attenuation of transmitted radiation.

What can increasing turbidity indicate?

  • Construction or commissioning debris
  • Metallic corrosion products
  • Scale or precipitated treatment chemicals
  • Polymer, hose, gasket or seal wear particles
  • External contamination introduced during maintenance
  • Biological material or detached biofilm
  • Material released after component degradation

What can turbidity not determine?

  • The chemical composition of suspended particles
  • The number and size distribution of particles
  • The concentration of a specific dissolved metal
  • The actual corrosion rate of system components
  • Whether microorganisms are active
  • The exact concentration of total suspended solids

Important distinction: Turbidity is an optical measurement, while total suspended solids, or TSS, is normally determined by separating and weighing suspended material. A correlation may exist in a stable fluid, but the relationship depends on particle size, shape, color, concentration and refractive index.

3. Why Coolant Cleanliness Matters in CDU Systems

3.1 Narrow coolant passages in cold plates

Modern cold plates use narrow internal flow passages to increase heat-transfer area and cooling performance. Open Compute Project pre-commissioning guidance notes that flow channels in current single-phase cold-plate designs may be approximately 100 micrometers wide.

Suspended material that can circulate through distribution piping may still accumulate or become trapped in:

  • Cold-plate microchannels
  • Quick-disconnect passages
  • Control valves and balancing devices
  • Strainers and filter elements
  • Heat-exchanger surfaces
  • Low-velocity sections and dead zones

A turbidity reading cannot directly predict whether a cold plate will become blocked. Nevertheless, a sustained increase provides evidence that the suspended-material condition of the coolant has changed and should be investigated.

3.2 Filter loading and hydraulic performance

Filters are a primary barrier against circulating contamination. As suspended material accumulates in a filter, differential pressure may increase and available coolant flow may decrease.

Monitoring Parameter Primary Information Provided
Turbidity Changes in suspended or optically active material
Filter differential pressure Hydraulic resistance and accumulation across the filter
Coolant flow Actual hydraulic performance of the cooling circuit
Laboratory TSS Gravimetric concentration of suspended material
Particle count Number and size distribution of particles
Metals analysis Potential evidence of corrosion or material release

Turbidity and filter differential pressure provide complementary information. Turbidity indicates a change in suspended optical material, while differential pressure indicates hydraulic resistance across the filter.

4. Principal Sources of Increasing Coolant Turbidity

4.1 Construction and commissioning debris

New piping and cooling systems may contain:

  • Metal chips and machining residue
  • Welding debris and oxides
  • Thread-sealant fragments
  • Dust, dirt and assembly contamination
  • Cleaning-agent residue
  • Packaging or transportation contamination

Cleaning, flushing and filtration should therefore be completed before the final coolant is placed into normal service. Online turbidity monitoring during commissioning can help the engineering team observe whether coolant clarity is improving as circulation and filtration continue.

4.2 Corrosion products

Liquid-cooling circuits may contain copper, stainless steel, brass, brazing alloys and other wetted materials. Incompatible materials, unsuitable coolant chemistry, oxygen ingress, depleted inhibitors or contaminated makeup water may contribute to corrosion.

Corrosion can release both dissolved metal ions and insoluble corrosion products. Insoluble material may increase turbidity, while dissolved metal ions may produce little or no proportional change in optical clarity.

A normal turbidity value does not prove that corrosion is absent. Corrosion assessment should also include pH, conductivity, inhibitor condition, dissolved and total metals, visual inspection and system-performance data.

4.3 Precipitation and coolant incompatibility

Incorrect chemical dosing, incompatible coolant formulations, hardness contamination or treatment chemical degradation may produce precipitated material.

  • Mixing incompatible coolants
  • Using unsuitable makeup water
  • Incorrect corrosion-inhibitor dosing
  • Changes in pH or ionic concentration
  • Contamination from cleaning chemicals
  • Chemical interaction with system materials

4.4 Microbial contamination and detached biofilm

Closed cooling systems reduce external contamination but are not automatically sterile. Microorganisms may enter through fill water, hydrotest water, contaminated hoses, open maintenance procedures or inadequately cleaned equipment.

Microbial cells and detached biofilm may influence turbidity, but turbidity cannot identify microbial species or determine whether microorganisms are active. Microbiological testing remains necessary when biological contamination is suspected.

4.5 Maintenance-related contamination

Opening a cooling loop to replace a pump, cold plate, filter, valve or quick disconnect creates an opportunity for contaminants and air to enter the system.

  • Residue from a replacement component
  • Dirt introduced during installation
  • Air bubbles following system refilling
  • Disturbance of settled material
  • Incomplete flushing
  • A change in coolant formulation or concentration

5. Industry Guidance for Turbidity and Suspended Solids

The Open Compute Project publication Guidelines for Using Water-Based Transfer Fluids in Single-Phase Cold Plate-Based Liquid-Cooled Racks provides example coolant-quality criteria for the water-based system covered by that document.

The listed criteria include:

  • Total suspended solids below 5 ppm
  • Initial-fill total suspended solids below 1 ppm
  • No visible solids
  • Turbidity below 5 NTU
  • No visible discoloration or opacity

These values should not be interpreted as universal limits for every CDU, coolant or data center. Final operating criteria must reflect the actual coolant formulation, filtration system, cold-plate design, wetted materials and equipment-manufacturer requirements.

Project-specific criteria should be agreed among:

  • The data center owner or operator
  • The CDU manufacturer
  • The server and cold-plate supplier
  • The coolant manufacturer
  • The water-treatment provider
  • The commissioning and engineering teams

6. Why Periodic Sampling Alone Is Not Sufficient

Laboratory analysis is essential because it can identify parameters that an online turbidity sensor cannot measure, including total suspended solids, particle distribution, metal concentration, inhibitor condition and microbial activity.

The main limitation of periodic sampling is measurement frequency. A laboratory sample represents the coolant condition at a specific location and time. A short-duration event occurring between two sampling dates may not be recorded.

  • A sudden release of corrosion products
  • Filter breakthrough or bypass
  • Contamination introduced during maintenance
  • Particle release during pump startup
  • Detachment of internal deposits
  • A temporary air-bubble event
  • Precipitation after chemical addition

The main value of online turbidity monitoring is not simply producing more NTU readings. It is showing when the coolant condition changed, how quickly it changed and whether it returned to the established baseline.

7. Recommended Coolant Monitoring Architecture

Online turbidity monitoring should complement laboratory analysis and routine engineering inspection. A practical monitoring program can be divided into three layers.

Layer 1: Continuous operational monitoring

  • Turbidity
  • Conductivity
  • pH where required by the coolant specification
  • Coolant temperature
  • Pressure and flow
  • Filter differential pressure
  • Coolant level and leak status

Layer 2: Scheduled field checks

  • Visual inspection of coolant appearance
  • Portable turbidity verification
  • Filter and strainer inspection
  • Leak and connection inspection
  • Basic coolant chemistry checks
  • Sensor cleaning and functional verification

Layer 3: Laboratory analysis

  • Total suspended solids
  • Particle count and size distribution where required
  • Dissolved and total metals
  • Corrosion-inhibitor concentration
  • Glycol concentration
  • Microbiological analysis
  • Coolant-specific chemical parameters

8. Establishing a Baseline and Alarm Strategy

There is no single turbidity alarm value suitable for every liquid-cooling installation. Alarm settings should be based on the clean-system baseline, coolant specification, measurement uncertainty and equipment-manufacturer requirements.

8.1 Establish the commissioning baseline

  1. Cleaning and flushing have been completed.
  2. The required filtration criteria have been achieved.
  3. The approved coolant has been added.
  4. Air has been removed from the circulation loop.
  5. The system has reached stable flow and temperature.
  6. A representative laboratory sample has been collected.

The initial online reading should be documented together with:

  • Coolant type, concentration and batch information
  • Operating temperature and flow rate
  • Sensor installation location
  • Filter type and operating condition
  • Laboratory turbidity and TSS results
  • Relevant metal concentrations
  • Date, operating status and commissioning conditions

8.2 Use multiple alarm dimensions

Alarm Dimension Engineering Purpose
Absolute turbidity Identifies exceedance of the approved maximum value
Deviation from baseline Detects deterioration before the maximum limit is exceeded
Rate of increase Identifies rapid contamination or disturbance events
Persistence time Reduces alarms caused by brief bubble or startup events
Cross-parameter confirmation Compares turbidity with flow, pressure and filter condition

8.3 Interpret the trend pattern

  • Gradual increase: possible corrosion-product accumulation, biological growth, precipitation or ongoing contamination.
  • Sudden step increase: possible maintenance contamination, deposit release, filter failure or chemical addition.
  • Intermittent spikes: possible bubbles, unstable flow or periodically resuspended material.
  • Increase after restart: possible sediment disturbance, outgassing or material released during pump startup.

9. Sensor Selection Requirements for CDU Applications

9.1 Low-range measurement capability

A properly maintained coolant loop is normally expected to operate at relatively low turbidity. The selected sensor should therefore provide useful resolution and repeatability close to the expected clean-system baseline.

9.2 Coolant and material compatibility

  • Deionized water
  • Inhibited water
  • Propylene-glycol mixtures
  • Ethylene-glycol mixtures
  • Corrosion inhibitors
  • Approved biocides and cleaning chemicals

9.3 Pressure and temperature rating

The sensor body, optical window, seals, cable connection and process fitting should be rated for the system's maximum normal operating conditions and any applicable pressure-test conditions.

9.4 Bubble control and installation position

Air bubbles scatter light and may produce false turbidity readings. The sensing area should remain completely filled with coolant during normal operation.

  • Air should not be routinely trapped around the optical window.
  • The measured coolant should be representative of the main loop.
  • Flow should remain sufficiently stable for repeatable measurement.
  • The probe should be accessible for inspection and cleaning.

9.5 Optical-window condition

Deposits, scratches or films on the optical window can alter the measurement. The maintenance plan should define inspection intervals, cleaning procedures, verification methods and replacement criteria.

9.6 System integration

  • 4–20 mA analog output
  • RS485 digital communication
  • PLC or CDU controller integration
  • BMS or DCIM data acquisition
  • Timestamped trend storage
  • Alarm delay and persistence logic
  • Communication and sensor-fault detection

10. Recommended Turbidity Sensor Installation Locations

Return line before filtration

This location can help detect suspended material returning from racks, manifolds and cold plates. It may provide an early indication of contamination generated within the IT cooling loop.

Downstream of filtration

This location can be used to evaluate the cleanliness of coolant supplied to the racks and to detect possible filter breakthrough, bypass or ineffective filtration.

Filter inlet and outlet

Monitoring both sides of a filter can provide information about particle-removal performance. However, two sensors increase installation cost, maintenance requirements and the need for measurement comparison.

CDU bypass or sample loop

A controlled bypass can provide stable flow and easier sensor access. The bypass must remain representative of the main circuit and should not become a stagnant dead leg.

For critical liquid-cooling systems, the monitoring location should be selected during the design stage. Adding a sensor after commissioning may result in poor flow conditions, trapped air or an unrepresentative measurement point.

11. Engineering Response to a Turbidity Alarm

A turbidity alarm should trigger a structured investigation. It should not immediately be assumed that the coolant requires replacement.

Step 1: Validate the measurement

  • Check the sensor operating status.
  • Review the latest calibration or verification record.
  • Inspect the optical measurement window.
  • Confirm that the sensing area is completely filled.
  • Check for air bubbles and unstable flow.
  • Inspect cable and communication conditions.
  • Determine whether the high reading is persistent.

Step 2: Review operating data

  • Pump starts and stops
  • Flow and pressure changes
  • Filter differential pressure
  • Coolant temperature
  • Chemical dosing
  • Filling or draining activity
  • Recent maintenance or component replacement

Step 3: Collect a representative coolant sample

  • Turbidity
  • Total suspended solids
  • Particle count and particle-size distribution
  • Dissolved and total metals
  • pH and conductivity
  • Glycol concentration
  • Microbiological analysis

Step 4: Inspect the filtration system

  • Filter retention rating
  • Current differential pressure
  • Position of any bypass valve
  • Filter-element integrity
  • Installation direction
  • Replacement and maintenance history

Step 5: Identify and correct the root cause

Corrective actions may include filter replacement, additional filtration, controlled flushing, removal of contaminated components, coolant replacement or correction of the approved treatment program.

12. RK500-07LC Liquid Cooling Turbidity Sensor

PRODUCT APPLICATION REFERENCE

The RK500-07LC is designed for online turbidity monitoring in liquid-cooling applications. According to its technical documentation, the sensor uses an optical correlation measurement principle.

A light emitter and receiver evaluate the change in transmitted light caused by scattering and absorption from suspended impurities in the coolant. The product documentation states that the optical design is intended to reduce interference caused by reflections from stainless-steel piping.

The sensor uses a sapphire measurement window and an integrated structure without requiring a separate external measurement module.

Main technical specifications

Parameter Technical Specification
Measurement range 0–10 NTU or 0–100 NTU
Accuracy ±2% of reading or ±0.1 NTU, whichever is greater
Resolution 0.1 NTU
Response time 1 second
Supply voltage 7–30 VDC
Signal outputs Simultaneous 4–20 mA and RS485
Power consumption Less than 0.2 W
Operating temperature 0 to 60°C
Pressure resistance 1 MPa / 10 bar
Probe protection IP68
Main material 316L stainless steel
O-ring material EPDM
Listed applicable media Deionized water, PG25 and EG25
Process connections G3/4, NPT3/4 or 50.5 mm sanitary clamp
Cable options M16 connector or direct cable

Recommended engineering role

The RK500-07LC can provide a continuous turbidity signal for integration with:

  • CDU control systems
  • Industrial PLC platforms
  • Building management systems
  • Data center infrastructure management platforms
  • Independent coolant-quality monitoring panels

The sensor should be described as detecting changes in coolant optical clarity. It should not be described as directly measuring particle size, TSS, corrosion rate or microbial concentration.

13. Measurement Limitations

Turbidity response is fluid-specific

The same particle concentration may produce different NTU readings in deionized water, propylene-glycol coolant and ethylene-glycol coolant because the optical properties of the continuous fluid phase are different.

Different particles produce different optical responses

Metal oxides, polymer fragments, biological material and mineral precipitates may scatter or absorb light differently. A previously established relationship between NTU and TSS may no longer be valid if the contamination type changes.

Air bubbles may resemble suspended particles

Entrained or trapped air can produce positive turbidity spikes. Correct installation, stable flow and alarm persistence logic are therefore necessary.

Optical deposits can cause measurement drift

A film or deposit on the measurement window may gradually change the indicated turbidity even when the bulk coolant condition remains stable.

Clear coolant is not necessarily healthy coolant

Dissolved ions, depleted corrosion inhibitors and some corrosion mechanisms may not create a significant turbidity increase. Turbidity should be interpreted together with conductivity, pH, metals analysis, inhibitor concentration and microbiological results.

14. Frequently Asked Questions

What is an acceptable turbidity level for CDU coolant?

There is no universal turbidity limit for every CDU system. OCP guidance for a specific water-based heat-transfer fluid lists turbidity below 5 NTU and no visible discoloration or opacity. The final limit should be confirmed with the CDU manufacturer, server supplier, coolant supplier and project engineering team.

Is turbidity the same as total suspended solids?

No. Turbidity is an optical measurement, while TSS is normally determined by separating and weighing suspended material. A correlation may be developed for a stable coolant and particle type, but it should not be assumed to remain valid when the contamination changes.

Can an online turbidity sensor detect corrosion?

It may detect insoluble corrosion products that affect optical clarity, but it cannot directly measure corrosion rate or dissolved metal concentration. Metals analysis and other corrosion assessment methods are still required.

Can turbidity monitoring replace filter differential-pressure monitoring?

No. Turbidity indicates changes in suspended optical material, while differential pressure indicates hydraulic resistance across the filter. The two parameters provide complementary information.

Can turbidity be measured in glycol-based coolant?

Yes, provided that the sensor materials and optical design are compatible with the specific glycol formulation. A separate operating baseline should be established for each coolant type and concentration.

Where should a turbidity sensor be installed in a CDU system?

It should be installed in a representative and continuously wetted location with stable flow and minimal bubbles. Depending on the monitoring objective, it may be installed in the return line, downstream of filtration or in a properly designed sample bypass.

How should a turbidity alarm be confirmed?

First inspect the sensor, optical window, flow condition and possible air bubbles. Then review filter differential pressure, pump status and recent maintenance. A representative coolant sample should be collected for laboratory analysis when the increase is persistent or significant.

15. Conclusion

Coolant cleanliness is a system-level reliability requirement in liquid-cooled data centers. Narrow cold-plate passages and sensitive hydraulic components make contamination control necessary during construction, commissioning and long-term operation.

Turbidity provides a rapid and practical indication of changes in suspended or optically active material. Continuous online measurement can reveal contamination events and deterioration trends that periodic sampling may miss.

Turbidity must nevertheless be interpreted correctly. It is not a direct substitute for TSS, particle count, metals analysis, microbiological testing or coolant chemistry measurements.

A reliable engineering strategy combines:

  • Online turbidity trending
  • Filter differential-pressure monitoring
  • Coolant flow, pressure and temperature data
  • Conductivity and pH monitoring where required
  • Scheduled laboratory analysis
  • Documented maintenance and coolant-treatment procedures

Within this framework, an online turbidity sensor becomes part of an evidence-based condition-monitoring system. It supports earlier investigation, more informed maintenance decisions and more controlled operation of the CDU and Technology Cooling System loop.

References

  1. Open Compute Project, Guidelines for Using Water-Based Transfer Fluids in Single-Phase Cold Plate-Based Liquid-Cooled Racks . View source
  2. Open Compute Project, Guidelines for Pre-Commission Preparation of Technology Cooling System Row Manifolds in Liquid-Cooled Data Centers . View source
  3. ISO 7027-1:2016, Water Quality—Determination of Turbidity—Part 1: Quantitative Methods . View source
  4. U.S. Environmental Protection Agency, Turbidity Parameter Factsheet. View source
  5. ASHRAE, Water-Cooled Servers: Common Designs, Components, and Processes . View source
  6. Rika Sensor, RK500-07LC Liquid Cooling Turbidity Sensor Technical Datasheet, Version V5.1.1 .

prev
The Necessity of Online Conductivity Monitoring in Liquid-Cooled CDU Systems – A Technical Analysis and Engineering Practice
recommended for you

LEAVE A MESSAGE

Just tell us your requirements, we can do more than you can imagine.
RIKA Sensor
Copyright © 2026 Hunan Rika Electronic Tech Co.,Ltd | Sitemap | Privacy Policy  
Customer service
detect