Salmon Thawing — Water & Chiller Redesign
A thaw process fixed with a bucket, a stopwatch and a thermal model
01 — Results
At spec
Bath held at setpoint — stops thermal degradation of the fish
Millions of gal
Annual water draw quantified and targeted for reduction
~85% less
Redesigned batch flow vs. continuous overflow today
Sized
Peak chiller duty modeled for summer conditions
| Metric | Before | After | Delta |
|---|---|---|---|
| Bath temperature control | Drifts above setpoint on air agitation | Setpoint held by pre-cooled supply + probe shutoff | Quality protected |
| Water per tote | Initial flood plus hours of overflow | No-overflow batch | Overflow eliminated |
| Thaw flow rate | Continuous overflow across all stations | Batch flooding only | −86% |
| Annual water draw | Millions of gallons per year | Targeted reduction, future state capped | Quantified |
| Chiller basis | Unsized, winter assumption | Duty range modeled across seasonal ΔT | Summer case carried |
| Capacity basis | Current line rate | Design rate with added stations and totes/day | Headroom built in |
02 — Challenge
Salmon thaws in insulated totes flooded to a chilled water bath. Air agitation was pushing the bath above its target temperature and degrading the fish, and the only lever operators had was letting city water overflow continuously for the whole thaw.
Nobody knew the real numbers. Hose flow, water volume per tote and the actual cooling duty were all assumed, and incoming city water swings sharply between winter and summer — so any chiller sized off winter conditions would fail in the season that matters.
03 — My role
Project lead — measurement, thermal sizing, phased capital plan
Timeline: Phased program
04 — Scope
- Baselined the process end to end: tote sizes, batch counts, shift pattern, inbound frozen condition and target bath temperature
- Measured hose-station flow with staged bucket tests rather than nameplate data, proving the system is supply-limited and that per-station flow collapses as stations are added
- Quantified water consumption per tote, per day and per year — establishing the real cost of continuous overflow control
- Sized the chiller against summer city-water conditions, carrying both current and future duty cases
- Engineered flow-reduction options: orifice restrictors, tote temperature probes with solenoid shutoff at setpoint, a recirculation loop with plate heat exchanger and controlled bleed, and tote covers with a standardized fill procedure
- Redesigned the batch method to eliminate overflow entirely, with staggered infeed and flooding compressed into a shorter window
- Phased the capital: pre-cooler validation and reuse of existing refrigeration first, purpose-built thawer budgeted in a later year
- Future state sized for planned growth, including the seasonal warm-water case
05 — Execution
- Ran staged bucket tests instead of trusting nameplate flow, proving the header is supply-limited and per-station flow collapses as stations open
- Built water usage bottom-up: initial flood plus continuous overflow, per tote, per day and annualized
- Sized the chiller against the summer city-water case, bracketing both a moderate and an aggressive supply target
- Laid out five flow-reduction options and ranked them: hose-station orifice restrictors, tote probe with solenoid shutoff at setpoint, recirculation loop with plate heat exchanger and controlled bleed, batch recirc skids, and tote covers plus a standardized fill procedure
- Redesigned the batch to a fixed tote count with no overflow and staggered infeed, compressing flooding into a single shift
- Phased the spend: validate a pre-cooler on the existing chiller first, then budget a purpose-built thawer in a later year
06 — Systems & technology
- Thermal load modeling
- Industrial thawing systems
- Plate heat exchanger / recirc
- Temperature-based solenoid control
07 — Artifacts & evidence
- Measured flow data at increasing station counts, cross-checked against a known reference
- Water balance per tote, per day and annualized for both current and future state
- Chiller duty table across the seasonal ΔT range for current and future flow
- Pre-cooler validation test plan including flood time and existing chiller reuse check
08 — Key takeaways
The whole project turned on a five-gallon bucket. Measuring flow with every station open — not one — is what showed the system was supply-limited, and it changed both the chiller size and the batch design.
Phasing the capital mattered as much as the engineering: a pre-cooler proves the thermal case now and earns the full thawer a budget line later.