Equipment Scaling, Hydrodynamics and Dispersion

Scope: Area; residence time; dispersion; experimental coefficients. Four lessons connect microscopic transport or process operating principles to practical calculations. Relevant laboratory models: 18.

Central engineering ideas

Equipment scale-up should distinguish physical similarity from equal nominal equipment size. Interfacial area, energy dissipation, liquid circulation and residence-time distribution often change even when fluid composition and bulk temperature are preserved. A prototype can fit mass-transfer coefficients over a narrow regime yet fail at scale because gas bubbles coalesce or flow bypasses wetted surfaces.

Use carefully selected nondimensional criteria, not indiscriminate matching of every group: matching Reynolds and Froude numbers simultaneously may be impossible when fluid and length scale are fixed. Tracer response and transient uptake reveal different aspects of equipment behavior. Quantify uncertainty on coefficients and validate outside the calibration points; otherwise excellent curve fitting can conceal a weak model.

Four-lesson reading path

Lesson 1: Interfacial area and geometric scaling

Equal tank volume does not imply equal interfacial area or mass-transfer performance, especially when bubble sizes change. The working equation a=6 epsilon_g/d_b for spherical bubbles makes a limited, testable claim about the dependence of transfer on specified properties. A larger bioreactor with coarse bubbles may lose kLa even at comparable superficial gas speed. The validity check is especially important here: Foam, coalescence and rheology change the bubble population. Before moving on, state in words what a higher driving force means physically, whether the response is proportional to area, and which temperature or composition must be used to obtain properties.

[Read Lesson 1](../lessons/unit-08-lesson-01.html).

Lesson 2: Residence-time distribution and axial dispersion

Departures from plug flow broaden residence-time distribution and alter concentration profiles and conversion/selectivity. The working equation Pe=u L/D_ax; sigma_theta^2≈2/Pe for high Pe makes a limited, testable claim about the dependence of transfer on specified properties. A tracer pulse reveals channeling that a mean residence-time measurement cannot detect. The validity check is especially important here: The high-Pe variance approximation fails in highly dispersed systems. Before moving on, state in words what a higher driving force means physically, whether the response is proportional to area, and which temperature or composition must be used to obtain properties.

[Read Lesson 2](../lessons/unit-08-lesson-02.html).

Lesson 3: Estimating coefficients from experiments

Transfer coefficients are inferred from transient measurements while separating sensor lag, mixing and physical property changes. The working equation ln[(c-c)/(c-c0)]=-kLa t makes a limited, testable claim about the dependence of transfer on specified properties. A dissolved oxygen step test estimates kLa only when oxygen uptake and probe response are accounted for. The validity check is especially important here: Regression uncertainty is not the same as model-structure uncertainty. Before moving on, state in words what a higher driving force means physically, whether the response is proportional to area, and which temperature or composition must be used to obtain properties.

[Read Lesson 3](../lessons/unit-08-lesson-03.html).

Lesson 4: Scale-up decisions, uncertainty and validation

Scale-up requires specifying preserved similarity groups and tolerable changes in pressure drop, transfer area and residence time. The working equation Report estimated coefficients with applicability range and confidence interval makes a limited, testable claim about the dependence of transfer on specified properties. Pilot trials should intentionally vary the parameters most likely to break the chosen correlation. The validity check is especially important here: A small pilot with uncontrolled chemistry cannot validate an industrial scale-up. Before moving on, state in words what a higher driving force means physically, whether the response is proportional to area, and which temperature or composition must be used to obtain properties.

[Read Lesson 4](../lessons/unit-08-lesson-04.html).

Analysis, synthesis and laboratory practice

First create a one-page property table: list fluid phase, bulk composition, surface or interface state, temperature, pressure, and the reference from which viscosity or diffusivity was obtained. Next sketch the principal control volume and write a component balance even when a transfer correlation is already known. Conservation closure should be computed using independent inventory data, not simply by substituting the computed flux back into the same equation.

Then use the simulation models to build a two-dimensional experiment: perform one sweep in the primary adjustable parameter and repeat at two distinct fixed values of a secondary parameter. Compare the resulting curves and explain whether the ranking changes. When model rankings do change, identify the competing effects. Where the equation predicts no change with a parameter that matters experimentally, the discrepancy is useful evidence of a missing mechanism rather than a reason to manipulate the fit.

For technical reporting, preserve the input CSV, baseline calculation, units and full list of assumptions. Include a short paragraph that distinguishes model sensitivity from actual process uncertainty, and identify one input that should be measured more precisely before design decisions. Link the interpretation to a measurable concentration, flow, pressure or inventory instead of treating the plotted output as a stand-alone achievement.

Suggested learning evidence: a control-volume sketch, two parameter sweeps, a worked balance, one falsification experiment and a paragraph about uncertainty.

Unit 8, Lesson 1: Interfacial area and geometric scaling

Unit: Equipment scaling. Core model: a=6 epsilon_g/d_b for spherical bubbles. Interactive laboratory: [Laboratory 18](../labs/lab-18.html).

What the phenomenon means

Equal tank volume does not imply equal interfacial area or mass-transfer performance, especially when bubble sizes change. The governing relationship is a=6 epsilon_g/d_b for spherical bubbles. In the calculation, every coefficient and concentration must be tied to a specific thermodynamic or hydraulic state. When a textbook draws a flux arrow, that arrow represents a signed rate through a stated area; it does not by itself prove that the assumed regime exists. The objective of the lesson is to connect the measurable laboratory signal to the correct transfer mechanism and to identify when one mechanism is no longer dominant.

Equipment scale-up should distinguish physical similarity from equal nominal equipment size. Interfacial area, energy dissipation, liquid circulation and residence-time distribution often change even when fluid composition and bulk temperature are preserved. A prototype can fit mass-transfer coefficients over a narrow regime yet fail at scale because gas bubbles coalesce or flow bypasses wetted surfaces.

A process example and a defensible workflow

A larger bioreactor with coarse bubbles may lose kLa even at comparable superficial gas speed. Start with a physical sketch showing control volumes, phase boundaries, sensor positions and the relevant flow rates. Identify the driving potential, compute a result on a consistent mole, mass, time and area basis, and translate the result into a quantity that a laboratory could measure. Confirm conservation independently. A predicted high flux is not automatically useful: it may require an unacceptable pressure, cause overheating, or imply an unrealistic solvent inventory. The preferred answer balances transfer rate with credible operating limits.

An experiment should separately record an initial value, a late value and at least one intermediate point. Compare the rate inferred from each interval. Under a steady, linear, constant-property model the slopes should be similar after excluding instrument transients; a systematic change suggests the claimed steady regime was not reached or the physical properties are changing.

Interpretation and validity

Use carefully selected nondimensional criteria, not indiscriminate matching of every group: matching Reynolds and Froude numbers simultaneously may be impossible when fluid and length scale are fixed. Tracer response and transient uptake reveal different aspects of equipment behavior. Quantify uncertainty on coefficients and validate outside the calibration points; otherwise excellent curve fitting can conceal a weak model.

Practical measurements have time constants. A concentration electrode or humidity sensor may respond more slowly than the process, causing a fitted diffusion time or mass-transfer coefficient to absorb instrument delay. Record probe location and sampling interval, and compare apparent rates in the first, middle and final periods of the experiment. If the response changes with stirring intensity, external transfer may be important.

Failure mode to recognize: Foam, coalescence and rheology change the bubble population. Document at least one observation that would falsify the current model. If a theoretical response is monotonic, ask what physical effect could make an observed curve flatten or reverse. The simulation provides a controlled numerical experiment; only measured property data and an appropriate scale-up study can establish real equipment performance.

Guided investigation

• Write each symbol and its SI units for a=6 epsilon_g/d_b for spherical bubbles. Distinguish a total rate from a flux and any concentration ratio from a dimensional concentration.
• Run the linked laboratory at its baseline settings and export the parameter-response curve. Identify a portion where the physical response is strongest.
• Predict the effect of raising the principal independent parameter by 10%, then compare the simulated value with your prediction. Explain any nonlinear response using the original equation.
• Identify a measurement that would estimate the most uncertain coefficient. Describe sensor position, replicate count and whether a control experiment is needed.

Questions for self-check

1. Which assumption permits the governing equation to be used here, and what real observation might contradict it?
2. If the primary driving-force measurement has a +5% systematic bias, what direction of bias would appear in the inferred coefficient?
3. Which is more useful for scaling this model: matching the total mass-transfer rate or matching the dimensionless groups, and why?

Unit 8, Lesson 2: Residence-time distribution and axial dispersion

Unit: Equipment scaling. Core model: Pe=u L/D_ax; sigma_theta^2≈2/Pe for high Pe. Interactive laboratory: [Laboratory 18](../labs/lab-18.html).

What the phenomenon means

Departures from plug flow broaden residence-time distribution and alter concentration profiles and conversion/selectivity. The governing relationship is Pe=u L/D_ax; sigma_theta^2≈2/Pe for high Pe. In the calculation, every coefficient and concentration must be tied to a specific thermodynamic or hydraulic state. When a textbook draws a flux arrow, that arrow represents a signed rate through a stated area; it does not by itself prove that the assumed regime exists. The objective of the lesson is to connect the measurable laboratory signal to the correct transfer mechanism and to identify when one mechanism is no longer dominant.

Equipment scale-up should distinguish physical similarity from equal nominal equipment size. Interfacial area, energy dissipation, liquid circulation and residence-time distribution often change even when fluid composition and bulk temperature are preserved. A prototype can fit mass-transfer coefficients over a narrow regime yet fail at scale because gas bubbles coalesce or flow bypasses wetted surfaces.

A process example and a defensible workflow

A tracer pulse reveals channeling that a mean residence-time measurement cannot detect. Start with a physical sketch showing control volumes, phase boundaries, sensor positions and the relevant flow rates. Identify the driving potential, compute a result on a consistent mole, mass, time and area basis, and translate the result into a quantity that a laboratory could measure. Confirm conservation independently. A predicted high flux is not automatically useful: it may require an unacceptable pressure, cause overheating, or imply an unrealistic solvent inventory. The preferred answer balances transfer rate with credible operating limits.

Before a design estimate becomes a purchasing specification, attach a measurement plan: the temperature and composition range, the reference for transport properties, the proposed independent validation case and an acceptable error tolerance. If the coefficient comes from a chart, record the relevant fluid regime, dimensionless groups and length definition so that another engineer can reproduce it.

Interpretation and validity

Use carefully selected nondimensional criteria, not indiscriminate matching of every group: matching Reynolds and Froude numbers simultaneously may be impossible when fluid and length scale are fixed. Tracer response and transient uptake reveal different aspects of equipment behavior. Quantify uncertainty on coefficients and validate outside the calibration points; otherwise excellent curve fitting can conceal a weak model.

A dimensionally plausible equation can still be wrong for the equipment. Confirm the area to which the flux refers, the concentration basis (mass, molar, dry or wet), and whether the flow is per total area or open pore area. In scaled equipment, area-to-volume ratio varies with size, so equal concentration gradients do not automatically imply equal total processing time.

Failure mode to recognize: The high-Pe variance approximation fails in highly dispersed systems. Document at least one observation that would falsify the current model. If a theoretical response is monotonic, ask what physical effect could make an observed curve flatten or reverse. The simulation provides a controlled numerical experiment; only measured property data and an appropriate scale-up study can establish real equipment performance.

Guided investigation

• Write each symbol and its SI units for Pe=u L/D_ax; sigma_theta^2≈2/Pe for high Pe. Distinguish a total rate from a flux and any concentration ratio from a dimensional concentration.
• Run the linked laboratory at its baseline settings and export the parameter-response curve. Identify a portion where the physical response is strongest.
• Predict the effect of raising the principal independent parameter by 10%, then compare the simulated value with your prediction. Explain any nonlinear response using the original equation.
• Identify a measurement that would estimate the most uncertain coefficient. Describe sensor position, replicate count and whether a control experiment is needed.

Questions for self-check

1. Which assumption permits the governing equation to be used here, and what real observation might contradict it?
2. If the primary driving-force measurement has a +5% systematic bias, what direction of bias would appear in the inferred coefficient?
3. Which is more useful for scaling this model: matching the total mass-transfer rate or matching the dimensionless groups, and why?

Unit 8, Lesson 3: Estimating coefficients from experiments

Unit: Equipment scaling. Core model: ln[(c-c)/(c-c0)]=-kLa t. Interactive laboratory: [Laboratory 18](../labs/lab-18.html).

What the phenomenon means

Transfer coefficients are inferred from transient measurements while separating sensor lag, mixing and physical property changes. The governing relationship is ln[(c-c)/(c-c0)]=-kLa t. In the calculation, every coefficient and concentration must be tied to a specific thermodynamic or hydraulic state. When a textbook draws a flux arrow, that arrow represents a signed rate through a stated area; it does not by itself prove that the assumed regime exists. The objective of the lesson is to connect the measurable laboratory signal to the correct transfer mechanism and to identify when one mechanism is no longer dominant.

Equipment scale-up should distinguish physical similarity from equal nominal equipment size. Interfacial area, energy dissipation, liquid circulation and residence-time distribution often change even when fluid composition and bulk temperature are preserved. A prototype can fit mass-transfer coefficients over a narrow regime yet fail at scale because gas bubbles coalesce or flow bypasses wetted surfaces.

A process example and a defensible workflow

A dissolved oxygen step test estimates kLa only when oxygen uptake and probe response are accounted for. Start with a physical sketch showing control volumes, phase boundaries, sensor positions and the relevant flow rates. Identify the driving potential, compute a result on a consistent mole, mass, time and area basis, and translate the result into a quantity that a laboratory could measure. Confirm conservation independently. A predicted high flux is not automatically useful: it may require an unacceptable pressure, cause overheating, or imply an unrealistic solvent inventory. The preferred answer balances transfer rate with credible operating limits.

Make a small mass-balance closure table with inflow, outflow, transfer rate and accumulation. Flux integrated over area must agree with the change in inventory after any time integration. When a calculation satisfies a local transfer law but violates total inventory, the likely issue is a basis conversion, interface area or hidden phase holdup.

Interpretation and validity

Use carefully selected nondimensional criteria, not indiscriminate matching of every group: matching Reynolds and Froude numbers simultaneously may be impossible when fluid and length scale are fixed. Tracer response and transient uptake reveal different aspects of equipment behavior. Quantify uncertainty on coefficients and validate outside the calibration points; otherwise excellent curve fitting can conceal a weak model.

Model assumptions should be tested against data rather than copied uncritically from a standard derivation. Vary one independent parameter at a time when possible; then vary a second parameter to reveal interaction. Report whether a discrepancy has the sign expected for fouling, film resistance, transient storage or a changed equilibrium. Use a simple residual plot instead of relying exclusively on R².

Failure mode to recognize: Regression uncertainty is not the same as model-structure uncertainty. Document at least one observation that would falsify the current model. If a theoretical response is monotonic, ask what physical effect could make an observed curve flatten or reverse. The simulation provides a controlled numerical experiment; only measured property data and an appropriate scale-up study can establish real equipment performance.

Guided investigation

• Write each symbol and its SI units for ln[(c-c)/(c-c0)]=-kLa t. Distinguish a total rate from a flux and any concentration ratio from a dimensional concentration.
• Run the linked laboratory at its baseline settings and export the parameter-response curve. Identify a portion where the physical response is strongest.
• Predict the effect of raising the principal independent parameter by 10%, then compare the simulated value with your prediction. Explain any nonlinear response using the original equation.
• Identify a measurement that would estimate the most uncertain coefficient. Describe sensor position, replicate count and whether a control experiment is needed.

Questions for self-check

1. Which assumption permits the governing equation to be used here, and what real observation might contradict it?
2. If the primary driving-force measurement has a +5% systematic bias, what direction of bias would appear in the inferred coefficient?
3. Which is more useful for scaling this model: matching the total mass-transfer rate or matching the dimensionless groups, and why?

Unit 8, Lesson 4: Scale-up decisions, uncertainty and validation

Unit: Equipment scaling. Core model: Report estimated coefficients with applicability range and confidence interval. Interactive laboratory: [Laboratory 18](../labs/lab-18.html).

What the phenomenon means

Scale-up requires specifying preserved similarity groups and tolerable changes in pressure drop, transfer area and residence time. The governing relationship is Report estimated coefficients with applicability range and confidence interval. In the calculation, every coefficient and concentration must be tied to a specific thermodynamic or hydraulic state. When a textbook draws a flux arrow, that arrow represents a signed rate through a stated area; it does not by itself prove that the assumed regime exists. The objective of the lesson is to connect the measurable laboratory signal to the correct transfer mechanism and to identify when one mechanism is no longer dominant.

Equipment scale-up should distinguish physical similarity from equal nominal equipment size. Interfacial area, energy dissipation, liquid circulation and residence-time distribution often change even when fluid composition and bulk temperature are preserved. A prototype can fit mass-transfer coefficients over a narrow regime yet fail at scale because gas bubbles coalesce or flow bypasses wetted surfaces.

A process example and a defensible workflow

Pilot trials should intentionally vary the parameters most likely to break the chosen correlation. Start with a physical sketch showing control volumes, phase boundaries, sensor positions and the relevant flow rates. Identify the driving potential, compute a result on a consistent mole, mass, time and area basis, and translate the result into a quantity that a laboratory could measure. Confirm conservation independently. A predicted high flux is not automatically useful: it may require an unacceptable pressure, cause overheating, or imply an unrealistic solvent inventory. The preferred answer balances transfer rate with credible operating limits.

To check the engineering consequence, compute the predicted response at the nominal case, then repeat after increasing the principal resistance by 25%. For a truly inverse dependence on resistance, the answer should fall to 80% of its original value. If your result rises instead, revisit the boundary definitions and the sign convention. This is a sensitivity test, not a substitute for experimental validation.

Interpretation and validity

Use carefully selected nondimensional criteria, not indiscriminate matching of every group: matching Reynolds and Froude numbers simultaneously may be impossible when fluid and length scale are fixed. Tracer response and transient uptake reveal different aspects of equipment behavior. Quantify uncertainty on coefficients and validate outside the calibration points; otherwise excellent curve fitting can conceal a weak model.

A reliable material balance begins by choosing a control surface and defining what passes through it. If the measured concentration belongs to the bulk while the governing law requires the interfacial value, the missing boundary condition must be solved or independently measured. Draw the transfer path before deciding which concentration difference is available. This small discipline prevents fitting a physically impossible flux to a convenient linear graph.

Failure mode to recognize: A small pilot with uncontrolled chemistry cannot validate an industrial scale-up. Document at least one observation that would falsify the current model. If a theoretical response is monotonic, ask what physical effect could make an observed curve flatten or reverse. The simulation provides a controlled numerical experiment; only measured property data and an appropriate scale-up study can establish real equipment performance.

Guided investigation

• Write each symbol and its SI units for Report estimated coefficients with applicability range and confidence interval. Distinguish a total rate from a flux and any concentration ratio from a dimensional concentration.
• Run the linked laboratory at its baseline settings and export the parameter-response curve. Identify a portion where the physical response is strongest.
• Predict the effect of raising the principal independent parameter by 10%, then compare the simulated value with your prediction. Explain any nonlinear response using the original equation.
• Identify a measurement that would estimate the most uncertain coefficient. Describe sensor position, replicate count and whether a control experiment is needed.

Questions for self-check

1. Which assumption permits the governing equation to be used here, and what real observation might contradict it?
2. If the primary driving-force measurement has a +5% systematic bias, what direction of bias would appear in the inferred coefficient?
3. Which is more useful for scaling this model: matching the total mass-transfer rate or matching the dimensionless groups, and why?

Applied project — Unit 8: Equipment scaling

Engineering decision

Equipment scale-up should distinguish physical similarity from equal nominal equipment size. Interfacial area, energy dissipation, liquid circulation and residence-time distribution often change even when fluid composition and bulk temperature are preserved. A prototype can fit mass-transfer coefficients over a narrow regime yet fail at scale because gas bubbles coalesce or flow bypasses wetted surfaces.

Project brief: Design a small experimental method to compare two candidate operating conditions related to interfacial area and geometric scaling; residence-time distribution and axial dispersion; estimating coefficients from experiments; scale-up decisions, uncertainty and validation. The purpose is not to produce a procurement specification, but to identify which input most affects mass transfer and whether an economical operating change could plausibly meet the desired duty.

Data and deliverables

• Draw a process flow schematic showing flow paths, relevant interfacial area and material inventory.
• Specify the two parameter sets using the bounds in the interactive laboratories; preserve the original and modified CSV files.
• Derive the governing model and predict relative response without touching the controls. Then compare with calculated responses.
• Determine the physical property that would contribute most to uncertainty. Plan at least three replicate measurements at two operating points.
• Document material compatibility, potential exposure or emissions, hydraulic constraints and credible failure mechanisms.
• Write a one-page decision statement recommending either a pilot trial, a revised analytical model, or rejection of the proposed change.

Independent validation

Use carefully selected nondimensional criteria, not indiscriminate matching of every group: matching Reynolds and Froude numbers simultaneously may be impossible when fluid and length scale are fixed. Tracer response and transient uptake reveal different aspects of equipment behavior. Quantify uncertainty on coefficients and validate outside the calibration points; otherwise excellent curve fitting can conceal a weak model.

A strong answer explicitly separates a screening assumption from a measured process characteristic. If two mechanisms can lead to the same observed transfer rate, propose a controlled test that discriminates between them. Scale-up should retain the causal mechanism, not just the appearance of a laboratory curve.

Evaluation: method and assumptions 25%; balance closure and units 20%; experiment design 20%; quantitative interpretation 20%; candid limitations and safety 15%.

Laboratory 15: Concentration polarization ratio

Relevant unit: Membrane transport. Governing relationship: c_m/c_b=exp(J_v/k).

Physical question

What happens to the predicted concentration polarization ratio when its principal adjustable parameter changes, and how does that sensitivity relate to the mechanism described in this unit? The baseline parameter values are synthetic teaching inputs, not measurements from a named plant. This experiment is designed to expose a mathematical trend and make its units visible before real property data are substituted.

Experimental setup and protocol

Perfect rejection, film theory and equal velocity units. List the assumptions on paper before changing inputs. Treat the first control as the independent variable; keep the remaining parameters fixed for the first sweep. Read the numerical output at the minimum, baseline and maximum setting, then export the 121-point curve. Repeat at a second value of the parameter most strongly coupled to the driving force. Use the on-screen schematic to identify the direction of transport and the conceptual location of the limiting resistance. It is not a scale drawing.

Parameters

• Permeate velocity: default 4 μm/s, adjustable range 0.2–12 μm/s. Check the measurement source and correlation domain before substitution.
• Boundary-layer coefficient: default 35 μm/s, adjustable range 8–90 μm/s. Check the measurement source and correlation domain before substitution.

Interpretation and model boundaries

Membrane permeance is a module-scale property that summarizes the driving-force to flux response. Osmotic pressure and hydraulic pressure oppose each other in pressure-driven desalination, while concentration polarization increases the solute concentration next to the membrane. Because membrane transport depends on chemistry, fouling and mechanical compression, a clean-water straight-line relation should be labeled a screening model.

The response curve is a local sensitivity plot in which other controls remain fixed. It does not automatically represent a physical trajectory through a reactor or a separation column. A reported ''outside model validity'' result indicates a mathematical or physical constraint such as an inverted concentration difference, invalid pinch, an incompatible moisture ordering, or a correlation beyond its specified regime. The laboratory deliberately leaves these cases visible so they can be diagnosed instead of silently clipped.

Investigation questions

1. Calculate the baseline result independently using c_m/c_b=exp(J_v/k). Convert all parameter units first. What is the largest source of numerical uncertainty?
2. Explain the curvature or straightness of the sweep. Does doubling the first parameter double, halve, or otherwise change the response?
3. Which plausible operating effect is excluded by the model: property variation, pressure drop, fouling, chemical reaction, unsteady accumulation, or imperfect contact? Describe a test for that effect.
4. If a customer specification required the output to improve by 15%, identify two possible changes and discuss an operational cost or safety implication of each.

Use restriction: Educational sensitivity model; it is not a substitute for verified correlations, material data, equipment design standards or engineering sign-off.

Laboratory 16: Single-contact leaching fraction

Relevant unit: Solid-fluid transfer. Governing relationship: f=K V/(M+K V).

Physical question

What happens to the predicted single-contact leaching fraction when its principal adjustable parameter changes, and how does that sensitivity relate to the mechanism described in this unit? The baseline parameter values are synthetic teaching inputs, not measurements from a named plant. This experiment is designed to expose a mathematical trend and make its units visible before real property data are substituted.

Experimental setup and protocol

Linear distribution model using consistently defined solid-equivalent units. List the assumptions on paper before changing inputs. Treat the first control as the independent variable; keep the remaining parameters fixed for the first sweep. Read the numerical output at the minimum, baseline and maximum setting, then export the 121-point curve. Repeat at a second value of the parameter most strongly coupled to the driving force. Use the on-screen schematic to identify the direction of transport and the conceptual location of the limiting resistance. It is not a scale drawing.

Parameters

• Fresh solvent-equivalent amount: default 90 kg, adjustable range 15–250 kg. Check the measurement source and correlation domain before substitution.
• Solid-equivalent inventory: default 180 kg, adjustable range 30–350 kg. Check the measurement source and correlation domain before substitution.
• Partition factor: default 2 —, adjustable range 0.1–6 —. Check the measurement source and correlation domain before substitution.

Interpretation and model boundaries

Liquid–solid extraction spans dissolution, porous diffusion and equilibrium partitioning. Available transfer area changes as particles shrink or swell; smaller particles transfer faster but complicate filtration. Adsorption involves finite sorption capacity and often competitive binding, so a linear distribution factor is useful only within an appropriate concentration interval.

The response curve is a local sensitivity plot in which other controls remain fixed. It does not automatically represent a physical trajectory through a reactor or a separation column. A reported ''outside model validity'' result indicates a mathematical or physical constraint such as an inverted concentration difference, invalid pinch, an incompatible moisture ordering, or a correlation beyond its specified regime. The laboratory deliberately leaves these cases visible so they can be diagnosed instead of silently clipped.

Investigation questions

1. Calculate the baseline result independently using f=K V/(M+K V). Convert all parameter units first. What is the largest source of numerical uncertainty?
2. Explain the curvature or straightness of the sweep. Does doubling the first parameter double, halve, or otherwise change the response?
3. Which plausible operating effect is excluded by the model: property variation, pressure drop, fouling, chemical reaction, unsteady accumulation, or imperfect contact? Describe a test for that effect.
4. If a customer specification required the output to improve by 15%, identify two possible changes and discuss an operational cost or safety implication of each.

Use restriction: Educational sensitivity model; it is not a substitute for verified correlations, material data, equipment design standards or engineering sign-off.

Unit 8 assessment — Equipment scaling

The assessment distinguishes memorization, calculation, interpretation and model criticism. Use SI conversions explicitly, give dimensional units for computed transfer rates and justify equilibrium assumptions.

Problem 1: Interfacial area and geometric scaling

A larger bioreactor with coarse bubbles may lose kLa even at comparable superficial gas speed. Write a component balance and model the transfer using a=6 epsilon_g/d_b for spherical bubbles. Repeat the result after a 20% change in the dominant physical coefficient, if the chosen model is linear in that coefficient. Explain what physically changes when the input varies, specify one observation that would invalidate the equation, and identify the minimum additional measurement required to calculate an engineering uncertainty interval.

Instructor guidance: A complete response identifies units and basis, draws the interfacial or control-volume boundary, makes the rate/flux distinction, and explains why a favorable model result may still be infeasible. The reported model limitation should include: Foam, coalescence and rheology change the bubble population.

Problem 2: Residence-time distribution and axial dispersion

A tracer pulse reveals channeling that a mean residence-time measurement cannot detect. Write a component balance and model the transfer using Pe=u L/D_ax; sigma_theta^2≈2/Pe for high Pe. Repeat the result after a 20% change in the dominant physical coefficient, if the chosen model is linear in that coefficient. Explain what physically changes when the input varies, specify one observation that would invalidate the equation, and identify the minimum additional measurement required to calculate an engineering uncertainty interval.

Instructor guidance: A complete response identifies units and basis, draws the interfacial or control-volume boundary, makes the rate/flux distinction, and explains why a favorable model result may still be infeasible. The reported model limitation should include: The high-Pe variance approximation fails in highly dispersed systems.

Problem 3: Estimating coefficients from experiments

A dissolved oxygen step test estimates kLa only when oxygen uptake and probe response are accounted for. Write a component balance and model the transfer using ln[(c-c)/(c-c0)]=-kLa t. Repeat the result after a 20% change in the dominant physical coefficient, if the chosen model is linear in that coefficient. Explain what physically changes when the input varies, specify one observation that would invalidate the equation, and identify the minimum additional measurement required to calculate an engineering uncertainty interval.

Instructor guidance: A complete response identifies units and basis, draws the interfacial or control-volume boundary, makes the rate/flux distinction, and explains why a favorable model result may still be infeasible. The reported model limitation should include: Regression uncertainty is not the same as model-structure uncertainty.

Problem 4: Scale-up decisions, uncertainty and validation

Pilot trials should intentionally vary the parameters most likely to break the chosen correlation. Write a component balance and model the transfer using Report estimated coefficients with applicability range and confidence interval. Repeat the result after a 20% change in the dominant physical coefficient, if the chosen model is linear in that coefficient. Explain what physically changes when the input varies, specify one observation that would invalidate the equation, and identify the minimum additional measurement required to calculate an engineering uncertainty interval.

Instructor guidance: A complete response identifies units and basis, draws the interfacial or control-volume boundary, makes the rate/flux distinction, and explains why a favorable model result may still be infeasible. The reported model limitation should include: A small pilot with uncontrolled chemistry cannot validate an industrial scale-up.

Problem 5: Laboratory analysis

Select Laboratory 18 from the unit and export a sensitivity CSV. Report the largest and smallest finite result, calculate a percentage sensitivity at the baseline using two nearby points, and explain whether you expect the same sensitivity when the temperature or composition changes. Give at least two plausible sources of nonrepeatability in a physical test.

Rubric (100 points): physical model and assumptions 25; calculations and dimensional checks 25; clear diagrams and evidence 15; data interpretation 20; limitations and safety 15. Numbers produced by the simulation are not independent evidence that its selected correlation applies to a real installation.