Freeze dryer

A slow leak or a tired vacuum pump stretches every cycle. The twin sees it cycle by cycle.

A freeze dryer takes ice out of the product as vapour, at low pressure, over many hours. Shelf heat drives it, the condenser catches it, and the vacuum pump keeps the pressure down. The twin solves the heat and mass balance for each cycle, so a leak, a weak pump or a heater fault shows as a number, not a long night.

Production freeze dryer, 20 m² shelf area, condenser at −70 °C, chamber 50–200 µbar

Start with one utility See how we prove results

Primary drying: 12.6 h
Peak sublimation: 9.5 kg/h
Warmest product: −27.8 °C
Condenser at peak: −68.5 °C

Anatomy and nameplate

1Chamber pressure. Capacitance gauge reads true pressure. Pirani reads water vapour high. When they agree, the ice is gone.
2Condenser temperature. Must stay cold enough to pull vapour from the chamber.
3Shelf inlet and outlet temperature. Heat into the product. The heater or chiller sets it.
4Vacuum pump current. Pump health and the gas load it is carrying.
5Refrigeration compressor current. Condenser capacity under load.
6Product probes. Product temperature in a few vials, kept below collapse.

Freeze dryer nameplate
Chamber and shelves
Shelf area	20 m², 16 m² of vials
Chamber volume	10 m³
Chamber pressure range	50–200 µbar, controlled with a nitrogen bleed
Reference recipe, primary	Shelf −10 °C, chamber 100 µbar, 0.8 cm fill, 91 kg of ice
Heat transfer to vials	Kv 18 W/(m²·K) at 100 µbar
Product collapse temperature	−25 °C
Condenser and vacuum
Condenser	−70 °C at no load, 300 kg ice capacity
Refrigeration	8 kW at −70 °C
Ice vapour pressure at the condenser	0.26 Pa at −70 °C
Vacuum pump	500 m³/h, rotary vane with booster
Pump-down to 50 µbar	11.9 min

The clean utilities behind a batch

Three paths run side by side. Water goes from the still to the hot loop. Steam sterilizes parts, and product moves on to drying and cold storage. Air goes through HEPA filters into the graded rooms.

Water
Air
Steam
Product
Has its own twin page

The physics

The equations behind every curve on this page.

Sublimation rate

dm/dt = Kv·Av·(T_shelf − T_product) / ΔH_s

Heat from the shelf, divided by the heat it takes to turn ice to vapour.

Vapour out of the vial

dm/dt = Ap·(P_ice(T_product) − P_c) / Rp

The dried layer resists the vapour. Heat in and vapour out must match, which sets the product temperature.

Pressure rise test

Q_L = V·Δp / Δt

Isolate the empty chamber and watch the pressure climb. The rate is the leak.

End of primary drying

P_Pirani / P_capacitance → 1

Pirani reads water vapour high. When the two gauges agree, the ice is gone.

Condenser balance

Q_refrigeration(T_c) = ṁ·ΔH_s + Q_leak

Every kilogram of ice that lands on the condenser is heat the refrigeration has to remove.

Leaks and vacuum

A tight chamber holds its vacuum. The pressure-rise test, read across cycles, shows a seal or valve going before it fails.

Pressure-rise leak test, cycle by cycle

Empty 10 m³ chamber isolated for 30 minutes

Test result
Trend

Rise in 30 min at the limit: 3.6 µbar
Pump-down to 50 µbar, new pump: 11.9 min
Same, tired pump and leak at the limit: 17.0 min

A tight chamber holds its vacuum. A rising result, test after test, finds the seal or valve before the batch record shows it.

Q_L = V·Δp/Δt | t = (V/S)·ln((p0 − p_u)/(p1 − p_u))

The condenser under load

At peak sublimation the condenser takes the most heat. It must stay cold enough to keep pulling vapour.

Condenser temperature against sublimation load

Chamber at 100 µbar

Refrigeration healthy
Compressor worn, 75% capacity

Peak sublimation, this recipe: 9.5 kg/h
Condenser at peak, healthy: −68.5 °C
Condenser at peak, worn: −59.6 °C

Every kilogram of ice that leaves the vials lands on the condenser as heat. A weak compressor lets the condenser warm, and the vapour stops flowing to it.

Q(T_c) = ṁ·ΔH_s + Q_leak | P_ice(T_c) < 0.2·P_c

Cycle by cycle

Same recipe, longer cycles. The twin works out how much heat reached the product and points at the part that fell short.

Primary drying time across cycles

Shelf fluid heater losing output, same recipe

Primary drying time
Model at the shelf temperature reached
Shelf temperature reached (right axis)

Primary drying, healthy: 12.6 h
Cycle 40: 15.1 h at −14.5 °C shelf

Less heat into the shelf means less ice leaving per hour. The twin ties the longer cycle to the heater, not to the product.

Kv·Av·(T_s − T_p) = ΔH_s·Ap·(P_ice(T_p) − P_c) / Rp

Failure modes

What breaks, what it looks like, what the twin does about it.

Failure mode	The physics	Signals that show it	How the twin catches it
Chamber leak	Air comes in through a seal or valve. The pump has to carry it, and the bleed has less room to control pressure.	Pressure-rise test climbing. Nitrogen bleed lower at the same pressure.	Trends the test and flags it 22 cycles before the site limit.
Vacuum pump oil and wear	Wet or old oil and worn vanes lower pumping speed and raise the lowest pressure.	Longer pump-down. Higher base pressure. Pump current and temperature up.	Times pump-down against pump speed and leak rate: +5.1 min with a tired pump and a leak at the limit.
Refrigeration compressor wear	Less capacity lets the condenser warm under load.	Condenser temperature above its curve at the same sublimation rate. Compressor current changes.	Checks condenser temperature against load. At 75% capacity, −59.6 °C at peak.
Shelf fluid heater or chiller fault	The shelves do not reach their set point, so less heat reaches the product.	Shelf inlet below set point. Outlet-to-inlet difference changes. Primary drying longer.	Solves the heat balance and flags it in cycle 21. By cycle 40 primary drying is 2.5 h longer.
Condenser overload	Too much ice or too warm a condenser, and vapour stops flowing to it. Chamber pressure rises.	Condenser near −55 °C. Chamber pressure climbing with the bleed closed.	Tracks condenser temperature against the limit and the ice load against its 300 kg capacity.

What the twin does not do

The recipe and the batch record stay with your freeze dryer controller and QA. The twin watches the equipment across cycles.
We do not change a recipe or a qualified setting. We recommend. Your change control and QA decide.

How we catch it early

A physics twin and a PINN, working on one freeze dryer.

The twin is a set of equations for this machine, tuned to its nameplate. Give it the flow, speed and conditions and it returns what every reading should be. It always gives the same answer for the same inputs.

A physics-informed neural network, or PINN, sits on top. It is trained on the plant’s own data and it is penalized whenever it breaks a physics rule. So it can only learn what the equations leave out: wear, fouling, friction. It cannot invent an answer that the machine could never produce.

Step 1
Measure
Flow, pressure, current, temperature, vibration. Whatever the asset already has.
Step 2
Physics twin
The equations of this machine, set to its nameplate. It says what the readings should be right now.
Step 3
Residual
Measured minus expected. Demand, weather and speed are already taken out.
Step 4
PINN
A small neural network that learns the cause of the gap, under the physics rules.
Step 5
Finding
A named cause, a date, and the kWh or hours at stake. An engineer reviews it.

loss = (measured − predicted)² + λ · (physics residual)²

The second term is the rulebook. It is zero when the network obeys the equations.

Why this beats black-box AI

A black-box model learns patterns from history. It needs examples of each failure, and it goes quiet when the pump runs at a point it has not seen. The twin starts from physics, so it works on day one. The PINN adds what physics cannot know about this particular machine.

That is also what makes the savings countable. The baseline is a calculation, not a guess, so the gap between baseline and meter is a number a customer can check.

What the twin checks on this asset

Product temperature from the heat and mass balance against the probes: −27.8 °C at its warmest
Pressure-rise test as a trend, cycle by cycle
Condenser temperature against sublimation load: −68.5 °C at 9.5 kg/h
Primary drying time against the model for the shelf temperature reached: 12.6 h
Pirani and capacitance gauges agreeing at the end of primary drying

	Fixed alarm	Black-box AI	Twin + PINN
Needs failure history to learn	No	Yes, lots	No
Says what is wrong	No	Rarely	Yes, by name
Works at a flow it has never seen	No	Often fails	Yes, the equations still hold
Puts a number on the saving	No	No	Yes, in kWh and dollars
Lead time on slow wear	Late	Varies	Weeks to months
Every number can be traced	Yes	No	Yes, to an equation

What the twin puts a number on

Energy and cost, from the equations.

Shelf heater losing output

+2.5 h a cycle

124 hours a year at 50 cycles

Leak warning

22 cycles

From the first rise in the test to the 0.02 mbar·L/s site limit

Tired pump and a leak

+5.1 min

Added to every pump-down to 50 µbar

Collapse margin at 200 µbar

0.3 K

2.8 K at 100 µbar. A pump that cannot hold pressure takes the product close to collapse.

What we do for you

We watch it. We catch it. We prove it.

Our reliability engineers run the twin on your freeze dryer, review every finding, and help your crew fix it. After the repair we compare the meter to the baseline and write the result in the ledger.

Start with one utility See the 90-day pilot

More assets

More from pharmaceutical, then the lead asset of each other industry.