Crystallised AI

The Crystallised AI Predictive Maintenance Dashboard is an industrial-grade monitoring and control system designed for real-time operational intelligence across distributed equipment fleets. The application provides operators, maintenance engineers, and facility managers with actionable insights into equipment health, alarm lifecycle management, and predictive maintenance workflows.

Purpose

The dashboard serves four primary functions:

Real-time Alarm Lifecycle Management — Operators can monitor active alarms through a complete lifecycle from NEW → UNACKNOWLEDGED → ACKNOWLEDGED → ASSIGNED → ACTIONED → RESOLVED. The system automatically prioritizes critical unacknowledged alarms and provides instant visibility into owner assignment, next actions, and escalation status. Delta values show exactly how far readings exceed thresholds (e.g., "+22 PSI over critical"), eliminating guesswork during high-pressure situations.

Asset Health Monitoring — Engineers gain visibility into device health scores, sensor readings, and degradation trends through interactive visualizations including geographic maps, digital twin floor plans with clickable equipment units, and tabular data views. Each device displays real-time metrics with threshold indicators, stale data warnings, and pattern detection explaining why alarms were triggered (drift, cross-sensor inconsistency, recurrence, trend accumulation, or spike patterns).

Configuration & Threshold Management — Administrators can configure global sensor thresholds with visual zone-based sliders, set device-specific overrides, manage user permissions, upload custom deployment photos, and configure scheduled alarm reports. All configuration changes take effect immediately without requiring system restarts.

Operational Intelligence & Reporting — The system provides clickable KPI cards that instantly filter to affected items, CSV export for offline analysis, bulk resolve workflows with resolution tracking, alarm storm suppression to prevent notification fatigue, and scheduled email reports summarizing critical trends and resolution rates.

Design Philosophy

The interface follows a Neo-Brutalist Industrial Functionalism aesthetic where icons serve as the primary language and text provides confirmation. This design reduces cognitive load during high-pressure operational scenarios by making critical information immediately recognizable through color-coded lifecycle state badges, sensor modality icons, directional trend arrows with explicit references, and delta values showing exact threshold exceedance.

System Requirements

The dashboard is a web-based application accessible through modern browsers including Chrome, Firefox, Safari, and Edge. The interface is fully responsive, adapting to desktop, tablet, and mobile screen sizes. No installation is required—simply navigate to the application URL and log in with your credentials.

First Login

Upon first accessing the dashboard, you will see the Operations Dashboard displaying real-time alarm data. The interface consists of five main areas:

Left Sidebar — Contains navigation icons for Dashboard, Devices, and Settings pages. The sidebar is persistent across all pages, providing consistent navigation.

Top Header — Displays the page title, operational control buttons (Group, Storm Control, Theme Toggle), view mode controls (List, Map, Twin), and system scale indicator (FLEET > SITE > ASSET > SENSOR).

Header KPI Cards — Seven interactive metric cards showing Active Alarms, Devices Online, Action Queue, Fleet Health, Data Quality, Unassigned Critical, and Overdue Actions. Clicking any card instantly filters the alarm table to show relevant items.

Main Content Area — Shows the primary data view based on your selected mode (alarm table with pinned priority section, device map, or interactive digital twin floor plan).

Right Panel (Desktop) — A resizable detail panel that appears when you click on an alarm or device, showing comprehensive information including lifecycle state, owner, next action, why flagged narrative, sensor readings, deployment photos, GPS location, action history audit trail, and resolve workflow.

Bottom Sheet (Mobile) — On mobile devices, tapping an alarm opens a full-screen bottom sheet with the same comprehensive context information available on desktop.

Navigation

The application consists of three main pages accessible via the left sidebar:

Interface Layout

The dashboard uses a consistent layout across all pages with the following components:

Header KPI Cards — Seven interactive metric cards at the top display key performance indicators with sparkline trend charts. These cards are operational—clicking them instantly filters the alarm table below to show affected items:

Clicking the same card again clears all filters and returns to the full alarm list.

Pinned Priority Section — The top 3 critical unacknowledged unassigned alarms are automatically pinned to the top of the alarm table with a visual separator. This ensures operators immediately see the most urgent items requiring assignment and action.

Alarm Table — The main content area displays all active alarms with the following columns:

• Severity — Color-coded badge (Critical/Warning/Info)
• Entity — Device name with deployment photo thumbnail
• Signal — Sensor type with modality icon
• Value — Current reading with delta threshold exceedance (e.g., "+22 PSI over critical")
• Trend — Direction arrow with tooltip showing explicit reference (e.g., "↑ 15% vs last hour")
• Status — Lifecycle state badge (NEW, UNACKNOWLEDGED, ACKNOWLEDGED, ASSIGNED, ACTIONED, RESOLVED)
• Owner — Assigned operator (highlighted in red if unassigned for critical alarms)
• Persistence — Time since alarm first triggered
• Actions — Quick action buttons

Context Panel — Clicking any alarm row opens a detailed panel on the right (desktop) or bottom sheet (mobile) showing:

• Operational Status — Lifecycle state, escalating flag, owner, next action
• Why Flagged — Pattern detection (drift, inconsistency, recurrence, trend, spike) with auto-generated narrative
• Sensor Readings — Current value, threshold zones, delta exceedance, trend references
• Action History — Complete audit trail with timestamps and actors
• Resolve Workflow — State-aware action buttons (Acknowledge, Assign, Mark as Actioned, Resolve & Close Loop, Escalate)

Theme Toggle

The dashboard supports both light and dark themes. Click the theme toggle button in the top header to switch between modes. Your preference is saved and persists across sessions.

The Operations Dashboard is the primary workspace for monitoring and managing active alarms across your equipment fleet.

Alarm Lifecycle States

Every alarm progresses through a defined lifecycle with six states:

Escalating Flag

Alarms can be marked as ESCALATING (separate from lifecycle state) when they require supervisor attention. Escalating alarms display a red "ESCALATING" badge next to the lifecycle state badge.

Delta Threshold Values

The Value column shows the current sensor reading followed by delta threshold exceedance in parentheses. Examples:

• "142 PSI (+22 PSI over critical)" — Pressure is 22 PSI above the critical threshold
• "78% (+8% over warning)" — Reading is 8% above the warning threshold
• "45°C (within safe)" — Reading is within safe operating range

This eliminates the need to mentally calculate how far a reading exceeds thresholds during high-pressure situations.

Trend References

Hover over the trend arrow in the Trend column to see explicit references:

• "↑ 15% vs last hour" — Reading increased 15% compared to one hour ago
• "↓ 8% vs baseline" — Reading decreased 8% compared to baseline average
• "↑ 22% vs 24h trend" — Reading increased 22% compared to 24-hour trend

Pattern Detection

The "Why Flagged" section in the context panel explains why each alarm was triggered using pattern detection:

• DRIFT — Gradual deviation from baseline over time
• INCONSISTENCY — Cross-sensor readings don't match expected correlations
• RECURRENCE — Same alarm keeps triggering repeatedly
• TREND — Sustained directional movement toward threshold
• SPIKE — Sudden sharp increase or decrease

Stale Data Indicators

Alarms with stale sensor data (not reporting for extended period) display:

• Clock icon next to the value
• "Last updated: 2h 15m ago" timestamp in the context panel
• "STALE" indicator in the Data Quality KPI card

Bulk Actions

Select multiple alarms using the checkboxes in the leftmost column, then use the bulk action buttons:

• Bulk Acknowledge — Acknowledge all selected alarms
• Bulk Assign — Assign all selected alarms to an operator
• Bulk Resolve — Resolve all selected alarms with a resolution form

The bulk resolve workflow prompts for:

• Resolution Reason — Dropdown with predefined options (Sensor Recalibrated, Component Replaced, False Alarm, Scheduled Maintenance, etc.)
• Notes — Free-text field for additional details
• Follow-up Task — Optional task description for future action

CSV Export

Click the Export CSV button to download all visible alarms (after filters applied) with 17 data fields:

• Alarm ID, Severity, Device Name, Device Type, Sensor Type, Current Value, Threshold Exceeded, Delta Value, Trend Direction, Trend Reference, Lifecycle State, Escalating Flag, Owner, Next Action, Pattern Type, First Triggered, Last Updated

Group View

Click the Group button to toggle grouped view. Alarms are grouped by device with collapsible headers showing:

• Device name with severity summary (e.g., "3 Critical • 5 alarms")
• Worst horizon — Longest persistence time among alarms in the group
• Suppressed count — Number of duplicate alarms hidden by storm suppression
• Owner — Highlighted in red if unassigned for critical alarms

Click the chevron icon to expand/collapse each group.

The Devices page provides asset-centric health monitoring with three view modes accessible via the view switcher in the top header.

List View

Displays all devices in a tabular format with the following columns:

• Device Name — Name with deployment photo thumbnail
• Type — Equipment type (Pump, Motor, Turbine, Compressor)
• Health Score — Percentage score with color-coded badge
• Active Alarms — Count of active alarms with severity breakdown
• Location — GPS coordinates
• Status — Online/Offline indicator

Click any row to open the device detail panel showing comprehensive information including sensor readings, alarm history, deployment photo, GPS location, and configuration settings.

Map View

Displays devices on an interactive geographic map with color-coded markers:

• Green — Healthy devices (no active alarms)
• Yellow — Devices with warning-level alarms
• Red — Devices with critical alarms

Click any marker to open the device detail panel. The map supports zoom, pan, and satellite/terrain view toggles.

Twin View (Digital Twin Floor Plan)

Displays an interactive isometric floor plan showing equipment layout with real-time status indicators. This is the default view when navigating to the Devices page.

Interactive Features:

• Clickable Equipment Units — Click any equipment unit (pumps, motors, turbines, compressors) to open the device detail panel
• Real-time Status Indicators — Color-coded dots with pulsing animations:
• Floor Plan Navigation — Zoom and pan to explore different areas of the facility

The digital twin provides spatial context for equipment health, making it easy to identify clusters of problematic devices or understand physical relationships between assets.

Custom Deployment Photos

In the device detail panel, click the camera icon to upload a custom deployment photo for any device. Supported formats include JPG, PNG, and WebP with a maximum file size of 5MB. Custom photos replace the generic fallback images and display in both the device list and alarm table thumbnails.

The Settings page provides configuration options organized into five tabs accessible via the left sidebar.

Global Thresholds

Configure sensor thresholds for all devices using visual zone-based sliders. Each sensor type displays three zones:

• SAFE zone — Bright emerald green
• WARNING zone — Vibrant yellow
• CRITICAL zone — Bright red

Drag the threshold handles to adjust the boundaries between zones. The range indicator shows the current threshold values with semi-transparent background. Zone labels display with white text overlays and text shadows for contrast.

Supported sensor types:

• Pressure (PSI), Temperature (°C), Vibration (mm/s), Flow Rate (L/min), RPM, Voltage (V), Current (A), Power (kW), Torque (Nm), Humidity (%), Oil Level (%), Gas Concentration (ppm)

All changes take effect immediately without requiring system restarts.

Device Setup

Manage individual devices with the following options:

• Add New Device — Create new device entries with name, type, location, and deployment photo
• Edit Device — Modify device properties including name, type, GPS coordinates
• Upload Deployment Photo — Click the camera icon to upload custom equipment photos (max 5MB)
• Delete Device — Remove devices from the system (requires confirmation)

User Management

Manage operator accounts and permissions:

• Add User — Create new operator accounts with email, name, and role
• Edit User — Modify user properties including name, email, role, and active status
• Delete User — Remove user accounts (requires confirmation)
• Role Assignment — Assign roles (Operator, Engineer, Administrator) with different permission levels

Scheduled Reports

Configure automated email reports summarizing alarm trends and resolution rates:

• Enable/Disable — Toggle scheduled reports on/off
• Frequency — Select Daily, Weekly, or Monthly delivery
• Email Recipients — Add multiple email addresses (comma-separated)
• Include Resolved Alarms — Toggle whether to include resolved alarms in reports
• Include Charts — Toggle whether to include trend charts and visualizations
• Report Preview — View sample report before enabling
• Test Report — Send test report immediately to verify configuration

Reports include:

• Active alarm summary by severity and device
• Resolution rate trends over time
• Top 10 devices by alarm count
• Average time to resolution by severity
• Escalation rate trends

Storm Control

Configure alarm storm suppression to prevent notification fatigue during cascading failures:

• Time Window — Duration in minutes for grouping related alarms (default: 5 minutes)
• Sensor Count Threshold — Minimum number of sensors triggering to classify as storm (default: 3)
• Grouping Rules — How to group related alarms:
• Show Suppressed Alarms — Toggle whether to display suppressed alarms in the table (with "Suppressed" badge)

When storm suppression is active, the group header displays the suppressed count (e.g., "3 suppressed").

Clickable KPI Cards

All seven header KPI cards are interactive filters. Clicking a card instantly filters the alarm table to show only relevant items:

• Active Alarms → Critical + Unacknowledged
• Devices Online → Offline devices
• Action Queue → Assigned + Actioned items
• Fleet Health → Escalating alarms
• Data Quality → Stale data alarms
• Unassigned Critical → Critical alarms without owners
• Overdue Actions → Assigned/Actioned items past due time

Click the same card again to clear the filter and return to the full alarm list.

Priority Sorting

The alarm table uses intelligent priority sorting that considers multiple factors:

1. Critical severity — Critical alarms sort above warnings
2. Unacknowledged state — Unacknowledged alarms sort above acknowledged
3. Unassigned owner — Unassigned alarms sort above assigned
4. Longest persistence — Older alarms sort above newer

The top 3 alarms meeting all criteria (Critical + Unacknowledged + Unassigned) are automatically pinned to the top with a visual separator.

Action History Audit Trail

Every alarm maintains a complete audit trail of all actions taken. The Action History section in the context panel displays:

• Timestamp — When the action occurred
• Actor — Who performed the action (operator name)
• Action — What was done (Acknowledged, Assigned to X, Marked as Actioned, Resolved, Escalated)
• Notes — Any additional details provided

This provides full accountability and traceability for compliance and post-incident analysis.

Resolution Workflow

The context panel displays state-aware action buttons based on the current lifecycle state:

The Resolve & Close Loop button opens a resolution form prompting for:

• Resolution Reason — Dropdown with predefined options
• Notes — Free-text field for additional details
• Follow-up Task — Optional task description

All actions are recorded in the action history audit trail.

Mobile Bottom Sheet

On mobile devices (smartphones and tablets), tapping an alarm row opens a full-screen bottom sheet instead of a side panel. The bottom sheet provides the same comprehensive context information available on desktop:

• Operational Status (lifecycle state, owner, next action)
• Why Flagged (pattern detection and narrative)
• Sensor Readings (current value, thresholds, delta, trend)
• Action History (complete audit trail)
• Resolve Workflow (state-aware action buttons)

Swipe down or tap the X button to close the bottom sheet.

Acknowledging a New Alarm

1. Locate the alarm in the table (NEW or UNACKNOWLEDGED state)
2. Click the alarm row to open the context panel
3. Review the "Why Flagged" section to understand the pattern
4. Click Acknowledge Alarm button
5. The lifecycle state changes to ACKNOWLEDGED
6. Action is recorded in the audit trail

Assigning an Alarm to an Operator

1. Click the alarm row to open the context panel
2. Click Assign to Operator button
3. Select operator from the dropdown list
4. The lifecycle state changes to ASSIGNED
5. The Owner column updates with the operator name
6. Action is recorded in the audit trail

Resolving an Alarm

1. Ensure the alarm is in ACTIONED state (if not, click Mark as Actioned first)
2. Click Resolve & Close Loop button
3. Fill out the resolution form:
4. Click Submit to close the alarm
5. The lifecycle state changes to RESOLVED
6. Resolution details are recorded in the audit trail

Escalating an Alarm

1. Click the alarm row to open the context panel
2. Click Escalate to Supervisor button
3. The ESCALATING flag is set (displays red badge)
4. The alarm appears when clicking the Fleet Health KPI card
5. Action is recorded in the audit trail

Bulk Resolving Multiple Alarms

1. Select multiple alarms using the checkboxes
2. Click Bulk Resolve button in the header
3. Fill out the resolution form (applies to all selected alarms)
4. Click Submit to close all selected alarms
5. All alarms transition to RESOLVED state
6. Actions are recorded in each alarm's audit trail

Filtering by KPI Card

1. Click any KPI card in the header (e.g., Unassigned Critical)
2. The alarm table instantly filters to show only relevant items
3. The card displays a highlighted state indicating active filter
4. Click the same card again to clear the filter

Exporting Alarm Data

1. Apply any desired filters (severity, device, owner, etc.)
2. Click Export CSV button in the header
3. A CSV file downloads with all visible alarms
4. Open in Excel, Google Sheets, or other spreadsheet software
5. The export includes 17 comprehensive data fields

Uploading Custom Device Photo

1. Navigate to Settings > Device Setup
2. Locate the device in the list
3. Click the camera icon
4. Choose Upload button
5. Select image file from your computer (max 5MB)
6. The custom photo replaces the generic fallback image
7. Photo displays in device list and alarm table thumbnails

Configuring Scheduled Reports

1. Navigate to Settings > Scheduled Reports
2. Toggle Enable Scheduled Reports to ON
3. Select frequency (Daily, Weekly, Monthly)
4. Add email recipients (comma-separated)
5. Toggle Include Resolved Alarms and Include Charts as desired
6. Click Test Report to verify configuration
7. Reports will be sent automatically at the scheduled time

Configuring Alarm Storm Suppression

1. Navigate to Settings > Storm Control
2. Set Time Window (duration for grouping related alarms)
3. Set Sensor Count Threshold (minimum sensors to classify as storm)
4. Select Grouping Rules (By Device, By Sensor Type, By Location)
5. Toggle Show Suppressed Alarms to control visibility
6. Changes take effect immediately
7. Suppressed alarms display in group headers when storm is detected

The dashboard supports keyboard shortcuts for common operations:

Keyboard shortcuts work in both desktop and mobile views (when using external keyboard).

The dashboard is fully responsive and optimized for mobile devices (smartphones and tablets).

Mobile Table View

When viewing the alarm table on mobile, you can switch between:

• Card View — Each alarm displays as a card with key information (severity, device, sensor, value, status)
• Table View — Compact table with essential columns (Severity, Entity, Signal, Value, Status)

Use the view mode switcher in the top header to toggle between views.

Mobile Bottom Sheet

Tapping an alarm row opens a full-screen bottom sheet with comprehensive context information:

• Operational Status
• Why Flagged
• Sensor Readings
• Action History
• Resolve Workflow

Swipe down or tap the X button to close the bottom sheet.

Mobile Navigation

The left sidebar collapses into a hamburger menu on mobile devices. Tap the menu icon to reveal navigation options (Dashboard, Devices, Settings).

Touch Gestures

• Tap — Select alarm or device
• Long press — Open context menu (multi-select mode)
• Swipe down — Close bottom sheet
• Pinch to zoom — Zoom map or digital twin floor plan

Mobile Optimization

The mobile experience includes:

• Larger touch targets — Buttons and interactive elements sized for finger taps
• Simplified layouts — Reduced clutter for smaller screens
• Optimized images — Compressed deployment photos for faster loading
• Responsive tables — Horizontal scroll for wide tables
• Bottom sheet panels — Full-screen detail views instead of side panels

Common Issues and Solutions

Alarm table is empty

• Check if filters are applied (click KPI cards to clear filters)
• Verify devices are online and reporting sensor data
• Check browser console for JavaScript errors

Context panel won't open

• Ensure you're clicking the alarm row (not the checkbox)
• Try refreshing the page
• Check if browser popup blocker is interfering

Custom deployment photo not uploading

• Verify image file size is under 5MB
• Ensure file format is JPG, PNG, or WebP
• Check browser console for upload errors
• Try a different image file

KPI cards not filtering correctly

• Click the card again to clear the filter
• Refresh the page to reset all filters
• Check if multiple filters are applied simultaneously

Bulk actions not working

• Ensure you've selected at least one alarm using checkboxes
• Verify you have permission to perform the action
• Check if alarms are in the correct lifecycle state for the action

Scheduled reports not sending

• Verify email addresses are correct (comma-separated)
• Check spam/junk folder for test reports
• Ensure "Enable Scheduled Reports" toggle is ON
• Verify SMTP settings are configured (contact administrator)

Theme toggle not working

• Clear browser cache and cookies
• Try a different browser
• Check if localStorage is enabled in browser settings

Mobile bottom sheet not opening

• Ensure you're tapping the alarm row (not swiping)
• Try refreshing the page
• Check if touch events are blocked by browser extension

Browser Compatibility

The dashboard is tested and supported on:

• Chrome 90+
• Firefox 88+
• Safari 14+
• Edge 90+

Older browser versions may experience compatibility issues. Update to the latest version for best performance.

Performance Optimization

For optimal performance with large datasets:

• Use filters to reduce visible alarm count
• Enable alarm storm suppression to group related alarms
• Close the context panel when not in use
• Disable "Include Charts" in scheduled reports if not needed
• Use List View instead of Map/Twin view for faster rendering

Getting Help

If you encounter issues not covered in this documentation:

1. Check the browser console for error messages
2. Take a screenshot of the issue
3. Note the steps to reproduce the problem
4. Contact your system administrator or support team with the above information

System Status

Check the system status indicator in the top header:

• Green — All systems operational
• Yellow — Degraded performance or partial outage
• Red — System unavailable or critical error

If the status is yellow or red, contact your system administrator.