Fluidized Bed Combustor — Fluidization Monitor

Detects and confirms when a fluidized bed combustor transitions from a packed state to a sustained fluidized state, using live AVEVA InTouch historian data.

File	Purpose
tui.py	Terminal monitor — queries historian, shows state and sustain timer
fluidization_detector.py	Pure detection logic, no UI dependency
historian_client.py	AVEVA InTouch pyodbc query wrapper
fbc_tags.json	Tag names, thresholds, connection config

---

Quick Start

pip install -r requirements.txt
python tui.py

Edit fbc_tags.json to set your historian address, tag names, and detection parameters before first run.

---

Terminal Monitor

The TUI connects to the AVEVA InTouch SQL historian and polls every 30 seconds. It displays:

• Live values — all seven tags (three dP, four temperatures)
• Total Bed dP trend — ASCII sparkline of the last 90 minutes with a transition marker
• Detection state — PACKED / PULSING / FLUIDIZED / DE-FLUIDIZATION ALARM
• Sustain timer — progress bar filling toward the 5-minute confirmation window
• Auto-calculated baseline and threshold — derived from the trend data on every refresh

Key Bindings

Key	Action
R	Manual refresh
T	Set lookback window and resolution
C	Open fbc_tags.json in default editor
D	Toggle debug log panel
H	Help screen
Q	Quit

---

Detection Algorithm

The bed passes through three states during startup:

State	Total Bed dP	Description
PACKED	~40 inWC	Media settled, air below minimum fluidization velocity
PULSING	> threshold (brief)	Transient air blast — not sustained
FLUIDIZED	> threshold for 5 min	Confirmed fluidization — latched

Dynamic threshold — on every refresh the detector takes the 10th percentile of all valid Total Bed dP readings in the trend window as the packed-bed baseline, then adds 2.5 inWC. The threshold adapts automatically to changes in bed media weight.

Confirmation — dP must stay above the threshold continuously for sustain_seconds (default 300 s). Any dip below threshold resets the timer. Transient pulses never confirm.

De-fluidization guard — after confirmation, if dP drops below baseline − 2 inWC for more than 120 s a DE-FLUIDIZATION ALARM is raised.

---

Configuration — fbc_tags.json

{
    "connection": {
        "server":   "128.1.3.54",
        "database": "Runtime",
        "uid":      "wwUser",
        "pwd":      "wwUser",
        "driver":   "SQL Server"
    },
    "tags": {
        "total_bed_dp":  "eEI0122PT",
        "nozzle_dp":     "eEI0123PT",
        "bed_media_dp":  "eEI0124PT",
        "zone1_temp":    "eEI0164TE",
        "zone2_temp":    "eEI0165TE",
        "zone3_temp":    "eEI0166TE",
        "zone4_temp":    "eEI0167TE"
    },
    "detection": {
        "sustain_seconds":     300,
        "threshold_margin":    2.5,
        "packed_baseline_pct": 10,
        "defluid_margin":      2.0,
        "defluid_seconds":     120,
        "valid_low":           5.0,
        "valid_high":          150.0
    },
    "trend": {
        "lookback_minutes":   90,
        "resolution_seconds": 10,
        "chart_width":        60
    }
}

All changes take effect on the next refresh — no restart required.

---

Requirements

• Python 3.9+
• textual >= 0.47.0
• pyodbc >= 5.0.0
• pandas >= 2.0.0
• Microsoft ODBC Driver for SQL Server (or legacy "SQL Server" driver) — pip install pyodbc alone is not sufficient

---

File Structure

fluidized_bed/
├── tui.py                    Terminal monitor (run this)
├── historian_client.py       AVEVA InTouch pyodbc query wrapper
├── fluidization_detector.py  Pure detection logic
├── fbc_tags.json             Tag names, thresholds, connection config
├── requirements.txt
├── Screenshot.jpg
└── FBC_MONITOR.md            Detailed operating guide (also via H key in TUI)