ThermalDIAG

ThermalDIAG – Programmable measurement, monitoring and diagnostics system

Define workflows with rule lists – no programming knowledge required.

Complete ThermalDIAG system

ThermalDIAG is a local on-premise system for industrial plants where thermal condition, electrical load, mechanical vibration and machine signals need to be monitored, evaluated and documented together.

ThermalDIAG runs locally on suitable Linux IPC hardware, complemented by ThermalDIAG Analytics on Windows. Deployment and integration are agreed for each project.

Benefits

Individual measurements become a traceable operating context.

Detect changes

ThermalDIAG can continuously evaluate temperature rises, current curves, vibration metrics as well as digital and analog signals against project-specific rules. This makes trends, recurring deviations and gradual changes visible.

Investigate faults traceably

An incident collects rule context, measurements, timestamps, actions and evidence so operators and technicians do not need to reconstruct events from separate sources.

Document clearly

Tickets, snapshots, work notes, measurements and reports support a structured workflow from the first alarm to the completed inspection.

ThermalDIAG Rules

Program workflows with rule lists – no programming knowledge required.

Rules are a core ThermalDIAG function. In a rule list, you define conditions, combine signals and specify automatic responses. Conditions → Filters → Actions: the rule list determines how project-specific monitoring and auxiliary procedures behave.

1. Conditions

Compare temperatures, current, flow, digital states or numeric variables with thresholds and reference values. Combine conditions with AND or OR; delta and temperature changes complement direct comparisons.

2. Filters

A schedule, previous incident and activation count within a time window define when a rule may trigger. All enabled filters must pass.

3. Actions

Document incidents, save snapshots, issue messages, control approved outputs or update values, text and display colours. Targets and permissions are configured for each project.

Example: monitor temperature during operation

The run signal is active AND a camera temperature exceeds its threshold for longer than the configured minimum duration. A schedule restricts triggering to the intended operating hours. A rule configured for this purpose documents an incident, saves a snapshot and marks the display status red. Hysteresis can prevent frequent state changes near the threshold.

Further applications

Rule-based functional and endurance tests.

ThermalDIAG rules connect measurements, states and approved reactions without classic program code. In addition to alarms, this supports project-specific test and auxiliary procedures that are not safety-related.

A prepared digital output can, for example, actuate a moving component. One or two limit switches provide position feedback while a numeric variable counts cycles and temperature, current or vibration is monitored. This keeps it traceable when and under which conditions a deviation occurs.

An anomaly is assessed in the next applicable rule evaluation cycle. Permitted actions are then executed through their action processors. Calculated numeric variables are available for subsequent rule evaluations. Evaluation and processing are separate sequences on a seconds timescale.

Technical scope:

  • 3 condition types: direct value, delta and temperature rate, using digital input and output states, analog values, camera temperatures, sensor values and numeric variables.
  • 3 filters: schedule, previous incident within a time window, and activation count within a time window.
  • 19 actions: create event, e-mail, voice, snapshot, set output, cancel pending output, set impulse, stop impulse, set Torch/LED, cancel pending Torch, calculate a numeric variable, reset a DI counter, write OPC UA, start an N6 inspection or request its result, set text, change display colours, and start or stop a TUF measurement.

ThermalDIAG supports supervised test and auxiliary sequences on the Linux IPC on a seconds timescale. Time-critical machine functions, emergency stops and protective interlocks are handled by the independent PLC or safety controller.

Physical outputs may only be enabled by qualified personnel after a project-specific review of the electrical, mechanical and safety-related conditions.

On narrow screens, scroll the tables sideways to read all columns.

Rules in detail: conditions, filters and actions

This description explains the Rules functions of ThermalDIAG V1.402, dated 2026-11-14.

Each rule has a name, a severity of 1–10, a cooldown, an order and an enabled state. Conditions use AND (all must be true) or OR (at least one must be true). A rule list defines the required behaviour without classic program code.

1. Conditions: sources and comparisons

Available sources depend on the configured devices and project variables.

SourceAvailable values
AX8 thermal measurementsSpot temperature; Box minimum, maximum and average temperature.
Camera and Modbus I/OConfigured numeric measurements, digital states and supported rising/falling edges; counter-capable Modbus inputs also expose counter value and increment.
MQTT devicesNumeric and Boolean signals, including configured OpenMV and Shelly values; communication-lost status is also available.
OPC UA variablesReadable Boolean, Int32 and Double variables; String variables support text equality and inequality.
TUF-2000M flow meter channelsConfigured Modbus channels for flow rate, velocity, net/positive/negative total volume, measurement volume, measurement-running and measurement-valid states.
Numeric variablesProject variables Num1–Num100, usable as values, references, thresholds and calculation results; an OpenMV-N6 camera result stored in NumX can be checked by another rule.
System variablestimer_sec (Unix seconds), local_minute_of_day (0–1439), local_day_of_week (1 = Monday, 7 = Sunday); CPU usage, CPU temperature and RAM usage.
ConstantsNumeric thresholds and references, including a left-hand constant for Delta; text constants for OPC UA String comparisons.
ConditionEvaluation
Direct comparisonCompare compatible values with >, ≥, <, ≤, = or ≠; digital-state and OPC UA text comparisons use = or ≠.
DeltaCompare the absolute difference, abs(A − B), with a constant or NumX threshold using >, ≥, < or ≤. Supported sources include numeric variables, system variables, constants, measurements, counters and digital values as 0/1.
Temperature changeDetect a rise or fall in an AX8 Spot/Box over 1–899 seconds using live samples; no threshold time or hysteresis.
Digital edgeDetect a supported camera/Modbus rising or falling transition; the sampling interval is selected to suit the pulse duration.
Communication lostFor MQTT sources: direct =/≠ comparison with 0 (healthy) or 1 (stale/offline); no hysteresis.

2. Filters: when may the rule trigger?

Filters decide whether true conditions may open or reactivate an incident. There is one filter of each type per rule. Disabled filters are ignored, and all enabled filters must pass. Evaluation order is Schedule → Previous incident window → Count window.

FilterPurpose
ScheduleAllow triggering on selected weekdays within daily local-time windows, using the project time zone. Unchecked weekdays are blocked while the filter is enabled.
Previous incident windowRequire another rule in the same project to have opened/reactivated an incident, or not to have done so, within the selected time window of up to 24 hours.
Count windowRequire a selected number of separate false-to-true condition activations within a time window. Continuously true conditions count once; the count resets after triggering.

Condition stability and freshness: Threshold time requires a supported condition to stay true for a configured duration. A NumX duration is rounded down and frozen when the true candidate starts; NumX hysteresis is read again at each evaluation. Hysteresis is available only for supported numeric comparisons. MQTT and OPC UA sources use direct comparisons and Delta.

Evaluation uses fresh, valid readings. IPC CPU/RAM values have a freshness threshold of 15 seconds. Local time values use the project timezone. Project text variables serve the Set text action, while OPC UA String sources serve text conditions.

3. Numeric calculation and NumX variables

A calculation action can assign a value or calculate with +, −, ×, ÷, MOD and FMOD and store the result in NumX. Sources include constants, NumX, system variables, AX8 temperatures, compatible numeric/digital Modbus values and counters, numeric OPC UA values and numeric or Boolean MQTT signals. Boolean values use 0/1. System variables are read as sources; results are stored in numeric project variables.

MOD uses rounded integer magnitudes; FMOD uses unrounded magnitudes. A zero divisor is rejected. Constants support up to six decimal places. The final result is rounded to the target precision using ROUND_HALF_UP. Measurement operands are captured for the queued action job; an earlier calculation action can update a variable used by a later action.

4. All 19 actions and their execution modes

The three execution modes of an action

The execution mode determines whether an action starts when the rule becomes active, while it remains active or when its active state ends.

Here, “active” means that the state monitored by the rule is present. This is separate from whether the rule is enabled in the settings.

A – When the rule becomes active

The action starts when the conditions are met, the filters allow activation and the rule changes from inactive to active.

The action is started once for that activation. While the rule remains continuously active, mode A does not start it again. If the active state later ends and the rule subsequently becomes active again, the action starts again.

Example: When an overtemperature alarm begins, save a snapshot and switch on an indicator light.

W – While the rule is active

The action can execute repeatedly while the rule remains active. This mode is suitable for operations that need to be repeated regularly during the monitored state.

Repetition in mode W continues while the rule is active. After the active state ends, this mode does not start any further repetitions.

Example: While the alarm is active, a calculation repeatedly updates a numeric variable.

C – When the active state of the rule ends

The action starts when a previously active rule returns to the inactive state, for example because the condition that caused the alarm is no longer met.

This is a single start associated with the end of that active state. The required closing action must be configured separately: mode C does not automatically undo the actions executed when the rule became active.

Example: When the overtemperature alarm ends, switch off the indicator light.

Different actions in the same rule can use different execution modes. Only the modes supported by an action can be selected for it.

The Mode column lists the available choices; it does not mean that all modes execute automatically at the same time.

ActionDescriptionMode
Create eventRecord an event with an optional message and tags.A
EmailSend a notification to configured recipients.A
VoiceSpeak a configured English, German or Italian message, with optional numeric placeholders and delay.A, C
SnapshotSave an image from a supported camera assigned to the selected location.A
Set outputSwitch a permitted camera, Modbus or MQTT/Shelly output ON/OFF, optionally after a delay.A, C
Cancel pending outputCancel a waiting output action for the selected target without forcing it OFF.A, C
Set impulseGenerate a timed pulse or pulse sequence with duration, pause, repetition and optional delay.A
Stop impulseRequest a running pulse sequence to finish after at most one further pulse and cancel older waiting work for the target.A, C
Camera torchSwitch supported camera illumination ON/OFF, optionally after a delay.A, C
Cancel pending torchCancel a waiting camera-light action without switching the light OFF.A, C
Numeric calculationAssign or calculate a value and store it in NumX; use numeric and Boolean MQTT signals as sources.A, W, C
Reset DI counterReset a digital-input counter that supports and permits resetting.A
OPC UA writeWrite a configured, correctly typed constant to a permitted OPC UA variable.A, C
Start OpenMV-N6 camera checkStart an inspection on the assigned OpenMV-N6 camera using configured result and send-time NumX variables.A
Get OpenMV-N6 camera check statusQuery the current OpenMV-N6 camera inspection and publish its validated result into the configured NumX variable.A, W, C
Set textReplace a project text variable with configured language-specific text values.A, W, C
Set display colorsChange the foreground and/or background color of a selected display cell.A, W, C
Start TUF-2000M flow meter measurementStart a measurement session using a fresh meter-counter boundary reading and a configured maximum permitted delay.A, C
Stop TUF-2000M flow meter measurementStop the measurement session using a fresh boundary reading and retain the measured volume and session status.A, C

OpenMV-N6 camera: The Start action begins the inspection; Get status subsequently retrieves the validated result.

TUF-2000M flow meter: The actions define the start and end of a software measurement session using the meter totalizer. Pump control and management of the device counter are configured separately.

5. Execution and observation

Cooldown defines the wait before reactivation. Actions are processed individually in the configured order. Delayed jobs are scheduled; cancellation removes waiting jobs. Stop impulse ends the sequence in a controlled way according to its pulse parameters.

Device capabilities, write permissions, enabled runtime and licensing still apply. The Incidents view shows activations, filter blocking and action results.

6. Time Delta and MQTT calculation

Time Delta: Delta uses the absolute difference, abs(A − B), including timer_sec. With timer_sec and NumX, the result is abs(timer_sec − NumX) regardless of operand order.

MQTT calculation operands: Numeric signals and Boolean signals as 0/1 are supported. The fresh, valid reading is matched to its configured source and retained for the queued action job.

Measurement service

Agree the measurement and evaluation

After technical consultation, you can commission a time-limited measurement with a documented evaluation. The measurement setup, duration and prerequisites are set out in an individual quotation.

1. Define the question
Describe the plant, observed condition, relevant operating phases and intended measurement objective together.
2. Check feasibility
Check available sensors, signals, interfaces and a suitable IPC.
3. Agree the measurement
Only if prerequisites are suitable, agree on the measurement setup, period, access rights and controlled procedure.
4. Assess results professionally
Document data quality, observations, limitations and next steps.

Discuss your application

Use and integration

Use ThermalDIAG locally.

ThermalDIAG is available as local software. Configuration, scope and delivery dates are set out in the individual quotation.

  • Local software for suitable Linux IPC hardware.
  • Hardware selection and sizing are project-specific; the UniPi Edge E410 is an integrated ARM64 platform.
  • ThermalDIAG runs on the UniPi Edge E410 ARM64 reference platform.
  • We can integrate further customer-owned IPCs, cameras and sensors into ThermalDIAG after technical consultation.
  • Setup and integration are agreed to suit the plant.

System requirements

Suitable IPC hardware for local operation.

ThermalDIAG runs locally on a suitable Linux industrial PC (IPC). The exact sizing depends on the number of cameras, measurement points, signals, recording duration and required integration.

Used environments include Ubuntu 22.04 with 4 GB RAM in WSL, on Linux PCs and on an OnLogic CL260, as well as the UniPi Edge E410 ARM64 platform.

System requirements for small installations

  • CPU: Linux-based x86_64 systems and the UniPi Edge E410 ARM64 reference platform have been technically tested. Other hardware requires a project-specific suitability check.
  • Operating system: Ubuntu 22.04 LTS or a currently supported Ubuntu LTS version is recommended. The operating system environment is agreed during setup.
  • Memory: at least 4 GB RAM; 8 GB or more is recommended.
  • Storage: at least 32 GB total SSD/eMMC capacity for small installations. The operating system, application, database and evidence all occupy this space.
  • Network: at least one Ethernet interface with 100 Mbit/s. Required bandwidth depends on camera count, data rate and other devices. For larger installations or higher data volumes, 1 Gbit/s is recommended, not a minimum.
  • Storage and hardware sizing depend on the number of signals, recording duration and operating mode. 24/7 operation requires suitable industrial hardware.

Select I/O for the measurement task

  • Digital inputs (DI): define the number and electrical characteristics according to the required status and enable signals.
  • Digital outputs (DO): provide only for required, authorized actions; assess the number and characteristics for each project.
  • Analog inputs (AI): select the number and ranges to suit the sensors, for example 0–10 V or 4–20 mA.
  • RS-485: check a suitable interface and required protective measures when using relevant devices.
  • I/O connection: directly on the IPC or through compatible external modules. There is no general minimum I/O channel count.

ThermalDIAG

Measurements, inspections and displays in one project.

ThermalDIAG extends shared monitoring with liquid measurement, OPC UA, N6 inspection workflows and configurable displays.

Displays for operation

Users assemble displays from measurements, digital states, text and images. Content updates automatically. The editor offers live preview and full-screen view; decimal places, text styles and colours are adjustable. Rules can change text and field colours.

TUF-2000M: flow and volume

The ultrasonic flow meter is integrated through Modbus RTU. Flow rate, velocity, totalizer values and measurement state are available for Visualization and rules. Displays also offer litres and l/min. Start/Stop measurements can be requested manually or through rules.

OPC UA: connect machine signals

Multiple OPC UA connections can be configured per project. Selected Boolean, Int32, Double and String variables can be read and used in rules. Numeric and Boolean values support history; approved writable targets can be set manually or through rules.

OpenMV N6: inspections and images

Rule actions start a configured N6 inspection and explicitly request its result. Results can be assigned to numeric variables. The camera can be assigned to a location; a snapshot can retrieve the JPEG of the latest available inspection.

Text and image variables

Each project provides 100 text variables and 100 image slots alongside numeric variables. Texts have separate German, English and Italian values. Rules can set texts; snapshots can update image slots. The same content can appear on multiple displays.

Snapshots, voice and system state

One capture can be archived, sent to an image slot or used for both purposes. Voice actions support Piper and browser speech, with up to ten ordered values. CPU usage, RAM usage and CPU temperature can also be used in rules and calculations.

A successful start begins measurement at zero; Stop retains the result. Measured volume and total volume since IPC start add validated amounts in both flow directions. The display identifies the measurement state; selected TUF fields can retain the last valid value in grey. The documented water tests illustrate data transfer and Start/Stop. Installation and calibration are defined for the application.

OPC UA data types and server profiles are matched to the installation. Analytics, backtesting and publishing use the feature profile agreed for each task.

A custom data display for the hall and workstation.

Users assemble their own displays from measurements, states, text and images. Content updates automatically; fixed numbers and text can be added through manually maintained project variables. Target values, shift notes and work instructions can therefore appear alongside current data.

The grid fills the available full-screen area; font sizes adapt to the space and images retain their aspect ratio. 8–10 large values suit a clear hall display, while 40–60 smaller values suit a detailed workstation. These are layout examples; monitor size, text length and viewing distance determine the arrangement.

A basic PC with a current browser and a LAN or WLAN connection to the Linux IPC is enough at the hall monitor. ThermalDIAG acquires and processes data on the IPC; the display PC opens the selected view. Authorized users can also change text and values from another workstation.

Explore integration

Interface

The key operating states at a glance.

The images illustrate the ThermalDIAG interface.

ThermalDIAG Home dashboard with project status, active incidents and health warnings
Home dashboard
ThermalDIAG visualization with temperature history, incident timeline and digital signals
Visualization
ThermalDIAG rules page with conditions, filters and actions
Rules
ThermalDIAG incidents page with workflow status, rule context and actions
Incidents

Shared signal context

Thermal, electrical and mechanical changes on the same time base.

Visualization separates temperature history, analog inputs, vibration, digital states and numeric variables into clear bands. CPU usage, memory usage and CPU temperature can also be viewed for the latest 15 minutes. Signal aliases and technical names remain visible together.

Temperature

Thermal-camera ROIs and other temperature measurements show local thermal changes.

Current and analog values

Current, voltage and other analog channels are grouped by compatible unit for meaningful comparison.

Vibration

Velocity, acceleration, frequency, kurtosis and crest factor support mechanical condition monitoring.

Digital states

Camera I/O and multiple Modbus gateways provide time-aligned inputs, outputs and relay states. DI counter values appear alongside the digital signals.

Reference sensors

Banner QM30VT2 and Seneca T201DCH100-OPEN are supported reference configurations. Naming, units and scaling are verified for each project.

Shared interaction

A shared cursor links the selected charts. The tooltip shows the nearest value for every visible curve, with a colour marker and an individual timestamp when sample times differ. The cursor shows the time difference from the marked position. Each series keeps its source visible.

In operation

What ThermalDIAG makes visible for operators.

  • Organize plant areas, cameras, measurement points and I/O signals by project.
  • Capture and visualize temperature, current, voltage, vibration, flow, volume, and digital, analog and OPC UA signals.
  • Connect integrated AX8 and OpenMV cameras, Shelly devices and multiple Modbus gateways through Modbus TCP, Modbus RTU and MQTT. Discovery checks AX8 cameras on the network and prepared OpenMV/Shelly devices through MQTT; Shelly devices can be assigned to locations.
  • Record OpenMV Spot and Box temperatures in history and use them in rules. Edit measurement areas graphically and open AX8 ROIs from the I/O device selector.
  • Use AX8, OpenMV and N6 images as snapshots: archive them, update image slots or do both. OpenMV captures offer original raw data and clean or annotated colour views.
  • Record DI counter history, use counter values in rules and reset supported counters under controlled conditions.
  • Use rules with conditions, filters and actions for alarms and approved outputs; verify digital output writes by readback.
  • Document incidents with measurements, action diagnostics and image evidence, and create reports.
  • Manage tickets for inspection, measurement, repair or feedback.
  • Show the state of runtime, cameras, I/O, rules and jobs, acknowledge previous failures, and as Supervisor/Admin stop runtime or additionally shut down the IPC after confirmation.
  • Pause runtime in a controlled way and resume with fresh measurement and rule state.
  • Deliver voice actions through Piper, browser audio or browser speech with up to ten values; available voices depend on the selected backend.
  • Review storage usage, request local recovery backups and configure retention for completed action execution logs.
  • Store data locally without mandatory cloud operation.

Licence status and resumption

A clear state instead of silent processing.

The licence page displays the current status in colour and warns when no more than 14 days remain.

  • Data acquisition and normal automation operate during the valid licence period.
  • Jobs start within the valid licence period. A pulse already in progress completes its pulse duration and returns to the configured state.
  • After a valid licence is installed, acquisition and rule evaluation resume automatically with fresh data and newly established timing and count windows.
  • Existing measurement, incident and audit data remain available.

Typical use areas

For plants where temperature alone does not explain enough.

ThermalDIAG combines the sensors, signals and operating states suited to the measurement task.

Motors, bearings and drives

Temperature, phase current, vibration metrics and machine state are viewed together so load-dependent or mechanical anomalies become easier to classify.

Control cabinets and connection points

Thermal image data can be combined with current, switching or enable information for terminals, relays, load circuits and distributions.

PV, inverters and battery storage

ThermalDIAG supports combined views of temperature, signal state and measurements in plants with changing operating states.

Cooling sections, pumps and fans

When temperature rise, flow, enable or switching state must be interpreted together, ThermalDIAG keeps the relevant data in one history.

Temporary measurements and diagnostics

Not every project needs to become a permanent system immediately. ThermalDIAG can also be used as a time-limited measurement and diagnostics setup.

Structured incident work

Incidents and tickets help turn an alarm into a traceable task with responsibility, measurements, notes, evidence and completion.

Project workflow

From observed condition to reliable evidence.

The following sequence describes project-specific setup after technical consultation and agreement on the order.

1. Clarify the starting pointWhich component is abnormal? Which signals matter? Are there shifts, cycles, load changes or recurring events?
2. Define measurement and signal conceptCameras, measurement points, current and vibration sensors, digital signals, analog values and required evidence are matched to the concrete question.
3. Set up the system locallyThermalDIAG is installed as a local web-based application. The exact setup depends on plant, roles, measurement task and safety framework.
4. Support operation and evaluationData is monitored, incidents are handled and results are documented. If needed, the rule strategy is adjusted or checked with Analytics.