PID Control for Viscosity

PID control for viscosity in industrial processes

PID Control for Viscosity: What Are You Really Controlling?

Viscosity affects how a material pumps, mixes, sprays, coats, doses and moves through production equipment. In paints, resins, adhesives, lubricants, syrups, polymers and many chemical products, maintaining a repeatable viscosity can be directly related to processability and finished-product quality.

A basic viscosity-control loop may look like this:

Viscosity setpoint → Viscosity measurement → PI/PID → Temperature, dilution or another manipulated variable → New viscosity

The important engineering point is that a PID controller cannot physically command “lower viscosity”. It must manipulate something in the plant that actually changes the material.

Depending on the process, this may be:

product temperature,

solvent or water addition,

concentrate flow,

blend ratio,

or another process input.

Before tuning Kp or integral action, define exactly which physical variable can move viscosity in the required direction.

Viscosity Is Not Just “How Thick the Product Looks”

Viscosity describes resistance to deformation and flow.

Dynamic viscosity is commonly expressed in Pa·s, mPa·s or cP.

Kinematic viscosity is commonly expressed in mm²/s or cSt and includes the relationship between viscosity and fluid density.

For process control, it is important to know which viscosity definition the instrument and product specification use. A setpoint in cP and a laboratory result in cSt should not be compared as though they were the same measurement.

Temperature Is Often the Dominant Process Variable

For many liquids, increasing temperature reduces viscosity. Cooling generally increases resistance to flow.

This is why temperature control is one of the most common ways of influencing viscosity in oils, resins and other process liquids.

A typical loop may therefore be:

Inline viscosity measurement → Viscosity PI/PID → Temperature demand → Heater or cooling system → Product viscosity

This means viscosity control can be closely connected to PID temperature control.

However, temperature may also affect reaction rate, curing, product stability or other quality properties. The controller cannot simply increase temperature without respecting these process limits.

A reliable temperature measurement is therefore often part of a good viscosity-control strategy.

A Cascade Structure Can Be Better Than Driving the Heater Directly

Viscosity is often slower than temperature.

Instead of allowing the viscosity controller to drive heater power directly, two loops can be used:

Outer loop: Viscosity PI/PID → Temperature setpoint

Inner loop: Temperature PID → Heater or cooling output

This cascade control structure allows the faster temperature loop to handle local thermal disturbances while the slower viscosity loop focuses on the final material property.

The inner loop should normally be significantly faster than the outer loop for cascade control to provide a real advantage.

Dilution Control Is a Different Viscosity Strategy

In paints, coatings and some blending processes, viscosity may be controlled by adding a solvent, water or another diluent.

The architecture may then become:

Viscosity measurement → PI/PID → Dosing pump or valve → Mixing → New viscosity

This looks straightforward, but the process may contain substantial mixing delay.

The controller adds a small amount of diluent, but the effect is not visible at the sensor until that material has mixed and travelled to the measurement point.

If integral action continues building during this delay, the system can add too much diluent and drive the product well below its target viscosity.

This makes time delay and process dynamics especially important in viscosity tuning.

Newtonian and Non-Newtonian Fluids Must Be Treated Differently

This distinction is essential in real viscosity work.

For a Newtonian fluid, viscosity at a given temperature remains essentially independent of *shear rate*.

For a non-Newtonian fluid, the apparent viscosity can change as *shear rate* changes.

Many industrial materials including paints, creams, adhesives, polymers and sauces can show non-Newtonian behaviour.

This means the viscosity measured while a material is gently moving may not be the same as the apparent viscosity while it is:

being pumped at high speed,

mixed aggressively,

forced through a narrow nozzle,

or applied as a coating.

For a non-Newtonian material, a viscosity value without temperature and shear conditions may be incomplete process information.

Shear Rate Can Change the Number the Controller Sees

*Shear rate* describes how rapidly adjacent layers of fluid move relative to one another.

In shear-thinning products, apparent viscosity decreases as shear rate increases.

Other materials may show more complex behaviour, including yield stress or time-dependent effects.

This has a practical consequence for PID control:

If production rate, pump speed or mixer speed changes, the measured viscosity of a non-Newtonian product can change even when the formulation has not changed.

In such a process, flow control, mixer speed and viscosity measurement may be coupled rather than completely independent.

Why Can Laboratory and Inline Viscosity Be Different?

A laboratory sample may be tested using a defined spindle, geometry, speed and temperature.

The production line is different.

The product may be flowing continuously, experiencing pump shear, mixing, different temperatures and different mechanical history.

For a non-Newtonian material, these differences matter.

An inline process viscometer and a laboratory rheometer or viscometer therefore do not necessarily have to produce identical numbers.

The important question is whether the process measurement is:

repeatable, stable, sensitive to meaningful product changes and correlated with accepted product quality.

Where required, the plant can establish a product-specific relationship between laboratory results and the inline process value.

Inline Measurement Changes the Control Problem

Traditional laboratory viscosity measurement is usually intermittent.

A sample is removed, transported and tested later.

During that time, the process may already have changed.

An inline or on-line viscometer can provide continuous or near-continuous process feedback, allowing a PI/PID controller to react before a large amount of off-specification material is produced.

Depending on the application, process viscosity measurement may use vibrating, rotational or other suitable measurement principles.

The measurement principle should be selected for the actual material, viscosity range, temperature and process conditions.

Sensor Location Can Make or Break the Loop

A viscosity sensor installed immediately after solvent injection may see material that has not mixed properly.

A sensor installed much farther downstream may see a more homogeneous product but introduce significant transport delay.

The best location therefore needs to balance:

mixing quality,

measurement delay,

flow profile,

product temperature,

air or gas entrainment,

cleaning and maintenance access.

Sensor placement is part of the control design, not merely an instrumentation installation detail.

What Do P, I and D Do in a Viscosity Loop?

Proportional action reacts to the present viscosity error.

Integral action removes a persistent difference between measured and target viscosity.

Derivative action can respond to the rate of change, but it can also react strongly to measurement noise and short-term process fluctuations.

For this reason, a PI controller may be entirely adequate in many viscosity applications.

Full PID should be selected because D provides a real process benefit, not merely because the controller offers three terms.

The detailed behaviour of the terms is explained in P, I and D components, while PID gain and tuning covers the parameter side.

Feedforward Can Handle Predictable Production Changes

Suppose product throughput doubles.

If required heating or diluent demand also changes predictably, a feedback-only viscosity controller has to wait for viscosity to move before correcting the process.

A feedforward term can use the measured production flow to provide an initial correction immediately.

The viscosity controller then trims the remaining error.

For example:

Main product flow → Feedforward heating or dilution demand

combined with:

Viscosity PI/PID → Fine correction

This can be particularly valuable where the viscosity loop contains significant mixing or thermal delay.

The relationship between feedback, feedforward and cascade structures is discussed in PID and other control strategies.

PLC Implementation

An inline viscosity transmitter can provide feedback to a PLC, DCS or dedicated process controller.

Where an analog output is used, 4-20 mA or 0-10 V scaling must correctly represent the engineering viscosity value.

The controller may then manage:

viscosity SP and PV,

PI/PID calculation,

temperature or dilution output,

Auto / Manual mode,

output limits,

instrument faults,

recipe-dependent parameters.

These implementation details are part of PID control with PLCs, not just viscosity control itself.

Output Saturation and Integral Windup Still Matter

Suppose the heater is already at 100% output but viscosity remains above target.

Or the solvent-dosing pump is already at its maximum flow.

The PID controller has no additional physical correction available.

If integral action keeps accumulating error, integral windup can develop.

When process conditions later change, the stored integral contribution may cause serious overshoot.

Before increasing Ki, check whether the actuator actually has remaining control authority.

For diagnosis of saturation, windup and related symptoms, see PID troubleshooting.

SCADA Trends Should Show More Than Viscosity

A useful SCADA trend should not display only viscosity PV.

Where possible, trend:

Viscosity SP

Viscosity PV

PID output

Product temperature

Main product flow

Diluent or additive flow

Heater or cooling demand

Auto / Manual status

If viscosity moves whenever temperature changes, that pattern is immediately visible.

If PID output changes but actual dilution flow does not, the fault is downstream of the PID controller.

Common Viscosity-Control Problems

PV is noisy: inspect sensor location, vibration, bubbles, flow conditions and the measurement signal.

Viscosity repeatedly crosses the target: Kp or integral action may be too aggressive for the thermal or mixing delay.

Diluent is added but viscosity changes much later: mixing and transport delay are likely important.

Viscosity drifts with temperature: temperature control or temperature compensation may need attention.

Laboratory and inline values disagree: compare temperature, *shear rate*, method, geometry and non-Newtonian behaviour.

One recipe controls well but another does not: the viscosity-temperature or viscosity-composition relationship may be product dependent.

PID output remains at maximum: check heater, cooling or dosing capacity before retuning.

Where Is PID Viscosity Control Used?

Paints and coatings: controlling spray, coating and application behaviour.

Resins, adhesives and polymers: maintaining processing, pumping and dosing consistency.

Oils and lubricants: monitoring and controlling material behaviour during transfer or processing.

Food processing: sauces, syrups, creams and other products where rheological consistency matters.

Chemical processing: applications where viscosity provides meaningful information about formulation or process state.

In every case, the relationship between the measured viscosity and actual product quality should be validated for the real process.

Commissioning Check for a Viscosity PID Loop

1. Is the inline viscosity measurement stable and repeatable?

2. Is measurement temperature known and controlled?

3. Is the material Newtonian or non-Newtonian?

4. What process variable actually changes viscosity?

5. Has the control direction been proven in Manual mode?

6. Does the heater, cooler, valve or dosing pump have enough capacity?

7. Is mixing or transport delay understood?

8. Is the sensor located where the material is representative and homogeneous?

9. Are output saturation and integral windup managed?

10. Are viscosity, temperature and flow variables trended together?

11. Has the inline measurement been correlated with the plant's accepted laboratory method?

12. Has tuning been checked across different recipes and production rates?

Related PID and Process-Control Topics

PID fundamentals: What Is PID? - P, I and D Components - PID Gain and Tuning

Process dynamics: Time Delay and Process Dynamics - Different Process Types - PID Troubleshooting

Measurement and control infrastructure: Temperature Sensors - Analog Signals - PID with PLCs - PID with SCADA

Related process control: PID Flow Control - PID Density Control - PID Temperature Control - PID and Other Control Strategies

Viscosity depends strongly on measurement conditions. Temperature, measurement method and, for non-Newtonian materials, shear rate and mechanical history can all affect the reported value. Inline and laboratory measurements should therefore be correlated under defined product and process conditions rather than assumed to be numerically interchangeable.