Background: The Original Problem
The plant had been operating since 2018 with a 4A molecular sieve bed sized for the original 4.5 MTPA nameplate. When a debottlenecking project brought throughput to 5 MTPA in 2024, the bed was undersized.
Symptoms: dehydration bed breakthrough occurred every 8 hours instead of the design 24 hours, regeneration cycles were running back-to-back, and sieve life dropped from 5 years to 22 months. Annual sieve replacement cost had hit $480,000.
The Diagnosis: Three Root Causes
A 2-week root cause analysis found three compounding issues:
- Cause 1: Original bed was 18% undersized for the new flow rate. The 2024 debottlenecking project did not include bed replacement.
- Cause 2: Regeneration gas was tapped from main process gas at 70 bar, with no pre-drying. Each regen cycle was loading the bed with 200-400 ppmv water. This 'self-poisoning' effect cut working capacity by 25%.
- Cause 3: Cycle time was fixed at 8 hours (legacy setting from 2018). With the new flow, this was far too long - sieve was 90% saturated at regen, accelerating degradation.
The Solution: Three Coordinated Changes
The plant implemented a coordinated upgrade package:
- Change 1: Replaced 4A bed with 22% larger high-density configuration. Custom 3.0-4.0mm bead size for higher bulk density (780 g/L vs 720 g/L standard).
- Change 2: Installed a small dedicated 4A guard bed (8% of main bed volume) to pre-dry regeneration gas. Regen gas water content dropped from 350 ppmv to <30 ppmv.
- Change 3: Upgraded control system for variable cycle time. Bed now regenerates based on outlet dew point sensor, not fixed timer. Average cycle extended from 8 to 14 hours.
Results: 12 Months of Operating Data
12 months after commissioning (May 2025 - May 2026):
- Sieve consumption: down 35% (from 48 MT/year to 31 MT/year)
- Sieve service life: 5+ years projected (up from 22 months)
- Regeneration energy: down 18% (fewer cycles, better regen)
- Unplanned downtime: zero in 12 months
- Total annual savings: $400,000 ($50K sieve + $180K regen energy + $170K deferred replacement)
Lessons That Apply to Any LNG or Natural Gas Dehydration Project
Lesson 1: When debottlenecking, always re-evaluate dehydration bed sizing. A 5% flow increase can cut sieve life by 30%+.
Lesson 2: Regeneration gas pre-drying is not optional at high throughput. The 8% guard bed paid back in 9 months.
Lesson 3: Fixed-time regeneration cycles are legacy thinking. Outlet dew point-based cycling extends bed life 40-60%.