membrane-troubleshooting

What this skill does

# RO/NF Troubleshooting & Diagnostic Methodology

Scope: spiral-wound thin-film-composite (TFC) polyamide RO/NF elements (8" standard + 4" variants). Numeric thresholds carry vendor + year of source.

Convention markers:
- `⚠️` — likely bug or incorrect assumption in calling code; highest priority
- `💡` — enhancement opportunity (new alert, new state, refined threshold)

---

## §0 — Diagnostic Frame of Reference

### 0.1 Five orthogonal failure modes

Every observed deviation must be classified into one (or more) of:

1. **Fouling** — reversible / partially reversible accumulation (colloidal, biological, organic, particulate)
2. **Scaling** — supersaturation-driven mineral precipitation, mostly in tail/concentrate stages
3. **Integrity loss** — physical breach of active polyamide layer or seal (O-ring, glue line, telescoping, pinhole)
4. **Mechanical compaction / chemical hydrolysis** — pressure-driven densification (largely irreversible); polyamide hydrolysis outside pH 2–11
5. **Instrumental / operational error** — out-of-cal flow meter, conductivity probe drift, dosing pump miscount, baseline never re-established after element swap

### 0.2 ASTM D4516 — the three observables

| Variable | Symbol | Direction of concern | Typical noise band |
|---|---|---|---|
| Normalized permeate flow | NPF | decreasing | ±3 % |
| Normalized salt passage | NSP | increasing | ±5 % |
| Normalized ΔP (feed → concentrate) | ΔPn | increasing | ±5 % |

### 0.3 Canonical signature table

Combines DuPont 45-D01650 troubleshooting matrix + Hydranautics TSB 107 logic + WaterTechOnline normalization article:

| NPF | NSP | ΔPn | Stage affected | Most likely root cause |
|---|---|---|---|---|
| ↓ | → / slight ↑ | ↑ | First stage | Colloidal / particulate fouling (front end) |
| ↓ | ↑ | ↑↑ | Tail stage | Advanced scaling (CaCO₃ / CaSO₄ / BaSO₄) |
| ↓ | ↑ | ↑ | All stages | Biofouling (channel plugging by biofilm) |
| ↓ | → | → / slight ↑ | All stages | Compaction or organic NOM adsorption |
| → / slight ↓ | ↑↑ | → | Localised (one vessel) | O-ring leak / element breach / glue-line crack |
| ↑ | ↑↑ | → | All stages | Chlorine oxidation (PA degradation, RO → NF drift) |
| → | → | ↑ | First stage | Mechanical / feed-spacer plugging, no membrane impact |
| ↓↓ | ↑ | → | One vessel | Telescoping in that vessel |
| inconsistent | inconsistent | inconsistent | — | Suspect instrumentation first; run mass balance |

### 0.4 Mass balance sanity check

Before any diagnosis:
```
|Qf − (Qp + Qc)| / Qf  <  0.02      # within 2 %
```
A 5 % error in recovery shifts scale-potential calculations dramatically.

💡 No mass-balance closure check is the **#1 cause of phantom alerts** from instrument drift. Add this check before any rule evaluation; halt diagnosis if it fails.

---

## §1 — Fouling Taxonomy

### 1.1 Colloidal fouling

- **Definition** — sub-µm inorganic / mixed particles (silt, clays, Fe/Al hydroxides) too small to settle
- **Predictor** — SDI₁₅ (ASTM D4189-23): < 3 acceptable; 3–5 frequent CIP; > 5 needs MF/UF upstream. Vendors recommend SDI₁₅ < 4
- **Signature** — NPF↓ first stage, ΔP↑ lead element, NSP stable
- **Cleaning** — alkaline (pH 11–12, NaOH + EDTA + anionic surfactant) first; acid (HCl/citric, pH 2–3) only if Fe/Al hydroxide confirmed

### 1.2 Biofouling

- **Definition** — biofilm (EPS matrix + bacteria) on membrane and feed spacer
- **Predictors** — AOC (assimilable organic carbon): biofilm-free risk threshold ~1 µg/L; conventional treatment achieves 50–100 µg/L (Pereira et al. 2014, PMC4021920). ATP, BFR, rising SDI without particulate cause
- **Signature** — NPF↓, ΔP↑ over weeks, slight NSP↑; autopsy shows slime mat on lead-element feed face
- **Cleaning** — alkaline NaOH (pH 11–12) + EDTA + surfactant, then non-oxidizing biocide soak. **Never** free chlorine for in-service biocide on PA membranes
- **DBNPA dosing** (industry consensus, IWA AQUA 2022; Feedwater datasheet):
  - Online maintenance: 10–50 mg/L for ≥ 1 h, weekly
  - CIP additive: 50–200 mg/L circulated 1–3 h
  - Continuous preventive: ~1 mg/L (where allowed)
  - Bisulfite upstream consumes DBNPA → increase dose
- Other accepted non-oxidizers: isothiazolinone (CMIT/MIT), glutaraldehyde for storage. See Hydranautics TSB 110

### 1.3 Organic fouling (NOM)

- **Definition** — natural organic matter (humics, fulvics), hydrophobic adsorption on PA surface
- **Signature** — NPF↓, NSP stable or slight ↑, ΔP → (no spacer plugging until late)
- **Cleaning** — alkaline NaOH pH 11–12 + EDTA chelant + surfactant (DuPont 45-D01504 default rule for unknown / combined fouling)

### 1.4 Particulate fouling

- **Definition** — > 1 µm material from failed pretreatment (cartridge bypass, blown media filter)
- **Diff vs colloidal** — visible deposit on cartridge filter; rapid ΔP rise on lead element only
- **Cleaning** — flush forward at high crossflow; alkaline CIP if needed

### 1.5 Combined fouling (most common in practice)

DuPont 45-D01504 rule: **alkaline cleaning as the first step** for unknown / combined; acid only when CaCO₃ or Fe/Mn hydroxide is confirmed.

Sequence:
1. Alkaline (NaOH + EDTA + surfactant), pH 11–12, T ≤ 35 °C above pH 10
2. Flush to neutral pH with permeate-quality water
3. Acid (HCl, citric, or sulfamic), pH 2–3, T ≤ 35 °C
4. Optional biocide soak (DBNPA 50–200 mg/L, 1–3 h)

⚠️ Never mix caustic and acid in the same loop (Hydranautics TSB 107).

---

## §2 — Scaling by Mineral Species

Solubility products at 25 °C cross-checked against Hydranautics TAB-111, DuPont 45-D01552, and standard chemistry tables.

| Species | Ksp (25 °C) | Saturation metric | Antiscalant family | Cleaning chemistry |
|---|---|---|---|---|
| **CaCO₃** (calcite) | 3.4 × 10⁻⁹ | LSI < 0 (no AS); ≤ +2.5 with AS; S&DSI for TDS > 10 000 ppm | Phosphonate (HEDP, PBTC, ATMP) | HCl or citric, pH 2–3, ≤ 35 °C |
| **CaSO₄·2H₂O** (gypsum) | 4.93 × 10⁻⁵ | IP/Ksp < 230 % (no AS); < 400 % with AS | Phosphonate + polyacrylate / sulfonated copolymer | EDTA-tetra-Na pH ~11, T 30–35 °C; long soak |
| **BaSO₄** (barite) | 1.08 × 10⁻¹⁰ | IP/Ksp < 6 000 % with strong AS | Phosphonate (DTPMP) + sulfonate polymer | Practically irreversible — EDTA hot soak partial; usually replace |
| **SrSO₄** (celestite) | 3.44 × 10⁻⁷ | IP/Ksp < 800 % with AS | Phosphonate + polyacrylate | EDTA + Na-citrate hot, slow |
| **SiO₂** (amorphous) | — | Solubility ≈ 100–150 mg/L @ 25 °C neutral pH; up to 250–300 traditional, 400–600 with dispersants | Polymeric dispersants (PEG/PVA/PAM) | Alkaline pH > 11 + fluoride (NH₄F or NaF); often autopsy/replace |
| **CaF₂** | 3.45 × 10⁻¹¹ | rare except geothermal/industrial | Phosphonate | Strong acid HCl, hot |
| **Fe(OH)₃ / Mn(OH)x** | low | Fe feed < 0.05 mg/L; Mn < 0.02 mg/L | Phosphonate + dispersant; better prevent via reduced Fe upstream | Citric acid + ammonia (pH 4), or Na₂S₂O₄ reductive clean |

### 2.1 Saturation calculations

**LSI** (low TDS): `LSI = pH − pH_s`; pH_s depends on Ca²⁺, alkalinity, TDS, T. LSI > 0 ⇒ CaCO₃ supersaturated.

**S&DSI** (TDS > 10 000 ppm): same logic, ionic-strength corrected.

**Sulfate scales / silica**: use ion-product / Ksp ratio: `S = IP/Ksp`. Vendor projection tools (ROSA, IMSdesign, WAVE, ROProMax) compute these per stage.

⚠️ Compute always at the **concentrate** (last element). LSI at feed is meaningless for scaling prediction.

---

## §3 — Membrane Integrity Loss

### 3.1 O-ring failure

- **Location** — interconnector between elements; permeate-tube to vessel end-adapter
- **Signature** — sudden NSP↑ localized to **one vessel**, NPF and ΔP unchanged
- **Detection** — conductivity profiling (Toray "Probing" white-paper); Rhodamine WT dye test
- **Repair** — replace O-ring; rebuild dry, light glycerin to seat. Do **not** use petroleum grease (attacks EPDM and PA)

### 3.2 Breach / pinhole

- **Detection methods**:
  - **VDT** (Vacuum Decay Test, ASTM D6908) — primary for RO/NF
  - **PDT** (Pressure Decay Test) — for MF/UF
  - Bubble test at 3–5 psig with submerged element