ParticleTensorAI® PTA1

ParticleTensorAI® PTA1 is the compact single-channel configuration of the ParticleTensorAI® measurement platform. It combines the LSS1 optical probe with the ZEON control unit for AI-based analysis of time-resolved light-scattering signals.

Unlike SprayQuantAI® SQA1, PTA1 does not primarily evaluate each droplet or particle event separately. Instead, the time-resolved detector signal is recorded over a defined acquisition interval, retained as buffered measurement data, and transformed into a multidimensional representation for ParticleTensorAI® analysis.

This approach allows the AI model to analyze not only individual signal shapes but also the temporal structure, signal density, statistical patterns, and changes occurring across many droplets or particles within the measurement interval.

PTA1 consists of the LSS1 optical measurement probe, the ZEON control and acquisition unit, and the ParticleTensorAI® analysis software.

Component	PTA1 Configuration	Function
Optical probe	LSS1	Generates the measurement region and detects time-resolved scattered light
Detector channels	1	Records one continuous time-resolved optical signal
Control unit	ZEON	Signal acquisition, buffering, processing and system communication
Evaluation	ParticleTensorAI®	AI-based analysis of buffered signal sequences

The LSS1 optical probe generates a specifically shaped illumination field. Droplets or particles passing through the measurement region scatter light toward the optical detector.

Because the particles move through the spatial light distribution, their interaction with the illumination produces characteristic time-dependent light-scattering signals. The LSS1 detector records these signals continuously with high temporal resolution.

In PTA1, ZEON retains the recorded signal over a defined acquisition interval instead of reducing the measurement immediately to individual particle parameters.

PTA1 uses one active optical detector channel and is therefore the most compact standard configuration within the ParticleTensorAI® platform.

The single detector channel records a continuous sequence of light-scattering information from droplets or particles passing through the measurement region. Instead of considering only isolated signal peaks, ParticleTensorAI® can analyze the structure of the complete buffered sequence.

This means that information can be contained not only in the waveform of an individual event but also in:

• signal amplitudes
• signal shapes
• temporal signal sequences
• particle or droplet event density
• overlapping signal structures
• statistical patterns within the acquisition interval
• changes in the signal distribution over time

The defining difference between PTA1 and conventional event-based particle measurement is the way the acquired signal is handled.

In classical counting systems, individual particle events are normally detected first and then evaluated separately. PTA1 can instead retain a complete signal segment over a defined time interval and use this buffered information as the basis for AI analysis.

The signal sequence is transformed into the multidimensional data representation required by the ParticleTensorAI® model. The AI model then evaluates characteristic patterns within the complete measurement interval.

PTA1 does not necessarily require every droplet or particle event to be detected individually before the AI analysis begins.

This is particularly useful when signal amplitudes vary strongly, particle events occur in rapid succession, or individual light-scattering signals overlap.

Because ParticleTensorAI® can analyze the complete buffered signal sequence, information from dense or complex particle populations can remain available to the AI model even when individual-event separation becomes difficult.

PTA1 is designed exclusively for AI-based evaluation. The relationship between the recorded optical signal structure and the desired measurement result is learned from suitable training and validation data.

Depending on the application, the AI model can be trained to distinguish between different spray conditions, particle populations, materials, or process states.

The exact output is therefore application-specific and depends on the optical configuration, available reference data, training range, and AI model.

Depending on the trained model and measurement task, PTA1 can be used for applications such as:

• classification of spray conditions
• classification of particles or droplets
• detection of process changes
• comparison of different operating conditions
• identification of material-related signal changes
• monitoring of stable and unstable spray conditions
• trend analysis
• AI-based process monitoring

Additional material-related or optical properties can be investigated when suitable reference data are available and the property of interest produces a measurable change in the recorded light-scattering signals.

Light-scattering signals are influenced not only by particle size and velocity. Optical and material properties can also change the measured signal structure.

ParticleTensorAI® investigates how these complex changes can be recognized directly from the recorded optical data. PTA1 therefore provides a compact platform for developing AI models that detect material-related or process-related changes within a spray or particle flow.

The achievable result depends on the application and must be demonstrated using appropriate reference measurements and independent validation data.

A ParticleTensorAI® model requires measurement data representative of the intended application. During development, optical signals are recorded under defined reference conditions and assigned to known classes or reference quantities.

The resulting dataset is divided into training and validation data. The AI model is trained using one part of the data and subsequently evaluated using independent measurements that were not used during training.

Important factors include:

• quality of the reference measurements
• range of operating conditions represented in the dataset
• stability of the optical configuration
• material properties
• number and diversity of measurement cases
• selected AI model architecture
• similarity between training and later application conditions

Customer-specific PTA1 applications may therefore require an individual measurement, training, and validation campaign.

The LSS1 is the single-channel optical probe used by PTA1. It creates the illumination field, defines the measurement region, and detects the scattered light generated by droplets or particles passing through the optical field.

The probe is connected to ZEON by optical fiber, allowing the sensitive acquisition and processing electronics to be positioned separately from the measurement location.

Parameter	LSS1
Measurement principle	Time-resolved light scattering
Active detector channels	1
Wavelength	405 nm
Dimensions	182 × 74.9 × 32 mm
Mechanical mounting	M3
Control unit	ZEON

Technical details for LSS1

The ZEON control unit provides signal acquisition, buffering, processing, host communication, and interfaces for integration into laboratory or industrial systems.

In PTA1 operation, ZEON is configured to retain time-resolved light-scattering data for ParticleTensorAI® processing rather than reducing the measurement immediately to individual particle parameters.

Parameter	ZEON
Optical signal inputs	1
Host communication	Ethernet
Digital inputs	4 × 24 V
Digital outputs	4 × 24 V
Analog inputs	4 × 4–20 mA
Analog outputs	4 × 4–20 mA
Trigger	Trigger input and output
Dimensions	251 × 211 × 109 mm

Technical details for ZEON

PTA1 can evaluate consecutive measurement intervals during an operating spray or particle process. Each interval can be analyzed by the trained AI model and converted into a classification, predicted parameter, probability value, or other application-specific result.

The resulting values can be tracked over time to detect changes in the process or deviations from previously recorded reference conditions.

Depending on the configured system, selected results can also be transferred to higher-level monitoring and automation systems.

PTA1 can be integrated into laboratory test rigs, development installations, spray equipment, and industrial processes.

The ZEON control unit provides interfaces for communication with external equipment and process infrastructure.

• Ethernet communication
• digital inputs and outputs
• 4–20 mA analog inputs and outputs
• external trigger input and output
• transfer of AI-generated process values
• integration into monitoring and automation systems

• AI-based spray classification
• process change detection
• material-related spray characterization
• paint and coating process monitoring
• comparison of atomizer operating conditions
• quality monitoring
• research on optical particle signatures
• development of application-specific AI measurement models
• industrial process monitoring

PTA1 can be adapted to the measurement task. The appropriate configuration depends on the particle or spray system, material, optical properties, concentration, process conditions, installation environment, available reference data, and required AI output.

Possible adaptations include:

• optical probe geometry
• working distance
• measurement region
• mechanical mounting
• optical configuration
• acquisition interval
• buffer configuration
• AI model
• output parameters
• industrial interfaces

PTA1 and SprayQuantAI® SQA1 use the same LSS1 optical probe and ZEON control-unit hardware. The difference lies in the acquisition and evaluation concept.

Parameter	SQA1	PTA1
Optical probe	LSS1	LSS1
Control unit	ZEON	ZEON
Detector channels	1	1
Primary data unit	Individual particle event	Buffered signal sequence
Evaluation concept	Individual-event evaluation	Tensor-based AI analysis
Evaluation	Classical or AI-based depending on configuration	AI-based
Typical focus	Quantitative droplet measurement and spray monitoring	Complex signal, material and process characterization

The shared hardware platform makes the difference between the two systems particularly clear: the optical signal is generated by the same physical measurement hardware, while firmware, acquisition workflow, data representation, and evaluation software determine how the information is used.

The three standard ParticleTensorAI® configurations differ primarily in the number of synchronized optical detector channels.

System	Optical Probe	Control Unit	Detector Channels
PTA1	LSS1	ZEON	1
PTA2	LSS2	IMEA	2
PTA4	LSS4	CLEON	4

PTA1 provides the most compact standard configuration. PTA2 adds a second synchronized detector channel, while PTA4 provides four optical channels for the most extensive standard ParticleTensorAI® signal representation.

ParticleTensorAI® builds on research in time-resolved light-scattering measurements, machine-learning-based particle characterization, and the analysis of complex optical signal sequences.

The underlying research investigates how information about particles, droplets, materials, and processes can be extracted from measured light-scattering signals using artificial intelligence.

[1] Schaefer, W.; Li, L. Particle characterization by analyzing light scattering signals with a machine learning approach. Applied Optics 2024, 63(29), 7701–7707. https://doi.org/10.1364/AO.531346

[2] Schaefer, W.; Fleck, S.; Haas, M.; Jakobs, T. Optical measurement method for monitoring high-mass-concentration slurry sprays: An experimental study. Photonics 2025, 12(7), 673. https://doi.org/10.3390/photonics12070673

[3] Schäfer, W.; Jakobs, T.; Stegmann, P. AI-assisted monitoring of coal particles within droplets in a coal monoethylene glycol slurry spray. 41st International Symposium on Combustion, Kyoto, Japan, 2026. KITopen ID 1000196272. https://doi.org/10.5445/IR/1000196272

• ParticleTensorAI® Overview
• ParticleTensorAI® PTA2
• ParticleTensorAI® PTA4
• LSS1 Optical Probe
• ZEON Control Unit
• SprayQuantAI® SQA1

For an application-specific PTA1 configuration, contact ai-quanton with information about the spray or particle system, material, process conditions, available reference data, and required analysis result.