A Universal Method for the Evaluation of In Situ Process Monitoring Systems in the Laser Powder Bed Fusion Process

Abstract

In situ process monitoring systems (IPMSs) are rapidly gaining importance in quality assurance of laser powder bed fusion (L-PBF) parts, yet standardized methods for their objective evaluation are lacking. This study introduces a novel, system-independent assessment method for IPMSs based on a specially designed Energy Step Cube (ESC) test specimen. The ESC enables systematic variation in volumetric energy density (VED) by adjusting laser scan speed, without disclosing complete process parameters. Two industrially relevant IPMSs—PrintRite3D by Divergent and Trumpf’s integrated system—were evaluated using the ESC approach with AlSi10Mg as the test material. System performance was assessed based on sensitivity to VED changes and correlation with actual porosity, determined by metallographic analysis. Results revealed significant differences in sensitivity and effective observation windows between the systems. PrintRite3D demonstrated higher sensitivity to small VED changes, while the Trumpf system showed a broader stable observation range. The study highlights the challenges in establishing relationships between IPMS signals and resulting part properties, currently restricting their standalone use for quality assurance. This work establishes a foundation for standardized IPMS evaluation in additive manufacturing, offering valuable insights for technology advancement and enabling objective comparisons between various IPMSs, thereby promoting innovation in this field.

1. Introduction

Additive manufacturing (AM) has evolved over the past few decades from a niche technology to an integral part of modern production processes. In particular, L-PBF has gained importance in industries such as aerospace, automotive and medical technology due to its ability to produce complex geometries with high precision. With increasing industrial application, there is also a growing need for reliable quality assurance methods that meet the specific challenges of additive manufacturing [1,2,3,4,5,6,7,8,9,10,11].

The transition from prototyping to series production in additive manufacturing has intensified the demand for robust quality assurance frameworks. Traditional post-process inspection methods, while effective, are time-consuming and cannot provide real-time feedback for process correction [3,12,13,14]. This limitation becomes particularly critical when manufacturing high-value components for aerospace and medical applications, where material waste due to defective parts represents significant economic losses [4,15,16]. Furthermore, the layer-by-layer nature of AM processes creates unique opportunities for real-time monitoring that are not available in conventional manufacturing methods.

IPMSs have proven to be a promising approach to real-time monitoring and quality control in additive manufacturing [2,3,15,17]. These systems use various sensor technologies to capture critical influencing parameters such as melt pool geometry, temperature distribution and surface finish during the build process [18,19,20]. Despite the great potential of IPMSs, there is currently no uniform standard for their evaluation and comparability, which hinders their widespread acceptance and effective implementation in industrial environments [2,21,22,23].

Recent advances in sensor technology, computational power, and data analytics have accelerated the development of sophisticated IPMS solutions. However, the proliferation of different monitoring approaches has created a fragmented landscape where direct comparison between systems is challenging [21,23,24]. Current evaluation methods are often limited to single-system studies or vendor-specific metrics, making it difficult for manufacturers to make informed decisions about IPMS selection and implementation [19].

The scientific community remains divided on the optimal sensor technology and data analysis method for IPMSs. While some researchers emphasize the superiority of optical systems [19,25], others argue for the advantages of acoustic or thermal sensors [20,26]. This controversy is amplified by the fact that different AM applications have varying requirements for process monitoring [4].

Another point of contention is the extent to which IPMS can be used as the sole means of quality assurance. Some studies suggest that IPMSs have the potential to replace traditional post-build inspection methods [25,27,28,29], while other researchers advocate a hybrid approach that combines in situ monitoring with conventional testing methods [4,26].

In view of these challenges and controversies, the present study aims to develop and validate a standardized, system-independent method for evaluating IPMSs. By introducing a novel test specimen, the Energy Step Cube (ESC), this work enables an objective comparison of different IPMSs in terms of their sensitivity and reliability in detecting process deviations and component defects.

The present study addresses these critical gaps through several novel contributions:

• First standardized cross-system evaluation method: Unlike previous studies that focus on individual IPMS performance, this work introduces the first systematic methodology for comparing different IPMS technologies.
• Universal test specimen design: ESC represents a paradigm shift from specific test methods for each IPMS to a universal approach that can be applied across different AM systems with different IMPS installed and materials while protecting proprietary process parameters.
• Vendor-independent assessment framework: The methodology enables objective performance comparison without requiring disclosure of proprietary algorithms or system specifications, addressing a major barrier to IPMS adoption in industry.
• Foundation for standardization: This work provides the technical foundation for developing international standards for IPMS evaluation, potentially accelerating widespread adoption of in situ monitoring technologies.

The main findings of this study show significant differences in the performance of various IPMSs and highlight the need for a differentiated approach when selecting and implementing these systems. In addition, this work lays the foundation for the development of industry-wide standards for evaluating IPMSs, which in the long term can contribute to improved quality assurance and increased reliability in additive manufacturing.

The central hypothesis of this work is that a controlled variation in volumetric energy density within a standardized specimen enables a system-independent evaluation of in situ process monitoring systems. By inducing reproducible process-signature changes without altering proprietary machine settings, the ESC provides a neutral basis for quantitative comparison between different IPMS technologies.

2. Methodology and Comparative Analysis of In Situ Process Monitoring Systems (IPMSs)

The most common defects in L-PBF components are pores, which can be divided into gas pores and bonding defects [26,27,30]:

• Gas pores are cavities caused by trapped gases and have a spherical structure. They are often caused by the evaporation of impurities or insufficient degassing during the melting process [31,32,33,34].
• Binding defect pores are amorphous cavities filled with unmelted powder and process gas and range in size from ten to several hundred micrometers. These pores occur when the process energy is insufficient to completely melt the material. This can occur stochastically when splashes with a larger diameter than the powder particles fall onto the powder bed, or systematically when the process parameters are set incorrectly [27].

These defects are particularly relevant for safety-critical applications, as they cannot be detected by visual inspection because they lie beneath the surface of the part and can act as stress concentrators, impairing the mechanical properties [1,26]. The industry expects IPMSs to be suitable for detecting process deviations that lead to the formation of these defects [15].

A special test specimen was developed for the systematic evaluation and comparability of different IPMSs, following approaches established in standardization literature [22,23]. This enables the sensitivity of the IPMS to fluctuations in volume energy density (VED) to be evaluated without having to disclose the complete process parameters [35,36].

2.1. Design of the Test Specimen “Energy Step Cube” (ESC) [37,38]

The test specimen, as shown in Figure 1, consists of a rectangular prism with dimensions of 30 mm × 10 mm × 10 mm. It is divided into nine vertical zones, each characterized by different volume energy densities. This variation is achieved by systematically adjusting the laser scanning speed (v) while keeping all other process parameters constant.

The middle zone of the test body represents the optimal construction process with a nominal volume energy density (VED_nom) of 100%. The selected scan-speed variation factors (0.25 × v to 2 × v) were derived from preliminary experiments conducted on AlSi10Mg using the baseline parameter set recommended by the machine manufacturer. This range was found to reliably induce both lack-of-fusion and keyhole-type porosity, thereby covering the full defect formation spectrum required for sensitivity assessment. Starting from this reference zone VED_nom, the scanning speed is adjusted factorially in both directions:

• The speeds range from 0.25 × v to 2 × v.
• This results in relative VED deviations of 400% (far left) to 50% (far right) compared to the nominal value.

The volume energy density is calculated using the following formula [39]:

$$\mathrm{V}\mathrm{E}\mathrm{D}={\displaystyle \frac{P_L}{h\times d\times v}}$$

(1)

where

• VED: Volume energy density [J/mm³];
• P_L: Laser power [W];
• h: Track spacing [mm];
• d: Layer thickness [mm];
• v: Scanning speed [mm/s].

2.2. Evaluation Methodology

The test specimens are evaluated using detailed graphical analysis. The nine zones of the test specimen are plotted along the X-axis of a diagram. Figure 2. Example (schematic) sensitivity curve illustrating the evaluation principle of the ESC for an IPMS (not experimental data). The proprietary anomaly detection metric of the respective IPMS is plotted on the left Y-axis for the corresponding zones [%]. At the same time, the relative porosity determined from the metallographic analysis or X-ray microcomputed tomography is displayed on the right Y-axis [%] [40].

The two curves are compared based on their slope behavior and shape. A high degree of consistency in the slope changes between the two curves indicates a reliable IPMS that can detect process deviations. High sensitivity of the IPMS in the range of 100% VED is particularly important, as this is where the optimal process window is located.

The more similar the two curves are in terms of their slope and the more sharply the IPMS anomaly curve approaches 0 at VED_nom, the more reliable the IPMS is in detecting defect occurrence probabilities. This detailed evaluation allows the IPMS’s ability to reliably detect defect occurrence probabilities to be validated and compared, enabling a well-founded assessment of the performance of different IPMSs in terms of their sensitivity to process deviations.

2.3. Advantages of the Methodology

This methodology, which uses the ESC as a test body, offers several significant advantages:

• Comparability: The standardized ESC test body geometry and sensitivity evaluation method allow different IPMSs to be compared directly with each other, addressing needs identified in standardization literature [2,19,41].
• Flexibility: Varying the VED by adjusting the scan speed and other process parameters allows the IPMS to be adjusted more precisely to the process window in order to optimize detection probability recognition [42,43].
• Protection of proprietary process parameter settings: Since only the relative change in scan speed is required, no sensitive process parameters need to be disclosed, which is crucial for industrial implementation [15,23].
• Universal applicability: The approach is applicable to every emission-based sensor signature that aims to detect deviations in the printing process. This enables cross-sensor comparisons, which is particularly relevant as there is still no consensus on universal test methods for assessing the sensitivity of individual IPMSs [2,3,15].
• Evaluation of anomaly detection capability: The method allows a universally applicable comparison between IPMS signals and material density, enabling a fundamental evaluation of detection capability [26,27,44].
• Flexibility: The ESC approach can be easily adapted to different materials and process parameters, making it relevant for a wide range of additive manufacturing applications, supporting the versatility requirements identified in quality engineering literature [4,41].

3. Experimental Investigation and Comparison of IPMSs

3.1. Experiment

This study examined two industrially relevant IPMSs:

• The integrated IPMS from the system manufacturer Trumpf;
• The equipment-independent PrintRite3D system from Divergent (formerly SigmaLabs or Sigma Additive Solutions).

ESC test specimens made of AlSi10Mg aluminum alloy were printed for both systems. Sensitivity curves were then determined to represent the response of the IPMSs to variations in volume energy density (VED) (see figures below). All builds were conducted under an argon atmosphere with oxygen concentration below 0.01%. The build chamber temperature was maintained at ~200 °C and monitored by the machine control system, ensuring stable environmental conditions throughout the build process. The TRUMPF Melt Pool Monitoring system records process-emission signals at approximately 100 kHz with two on-axis photodiodes covering the visible and infrared range. The effective spatial resolution is about 80–120 µm, defined by the scan-vector mapping of the melt pool.

The PrintRite3D system (Divergent) employs three on-axis photodiodes with a sampling rate of roughly 200 kHz, generating layer-wise grayscale process maps with a spatial resolution of approximately 100 µm per pixel. Additional internal validation builds were performed with different in situ monitoring systems (e.g., SLM Solutions MPM v4.47, EOSTATE Exposure OT v1.5, and Colibrium Additive MSPC v1.0) and with alternative alloys such as CoCrMo and Ti-6Al-4V. Although these results cannot be published in detail due to confidentiality agreements, the general behavior observed confirms that the ESC approach is transferable across systems and materials.

The actual porosity of the samples was quantified using micrograph analysis. Porosity evaluation was performed in-house by optical microscopy and image analysis on metallographic cross-sections. All measurements were repeated multiple times and analyzed by a single operator to ensure consistency. Although a dedicated uncertainty analysis was not explicitly conducted, the internal evaluation procedure used in this work typically achieves a measurement uncertainty of about ±0.05–0.1% absolute porosity.

3.2. Results Analysis

The specific observations from Figure 3a,b are evaluated in detail below.

Trumpf IPMS:

• Lowest porosity at 123% VED_nom, VED_nom and 80% VED_nom, corresponding to the lowest IPMS anomalies (37%, 21% and 28%).
• Porosity increases above 2% at >160% VED_nom and 67% VED_nom.
• Increased anomalies at VED > 123% VED_nom or <80% VED_nom.
• At VED_nom, the anomaly level was 20%, indicating a configuration for detecting slight changes in a wide process window.

PrintRite3D IPMS:

• Minimal porosity (<0.2%) was observed in VED_nom, 80% VED_nom and 67% VED_nom.
• Significant increase in porosity to 2.85% at 229% VED_nom, slight increase to 0.08% at 57% VED_nom with a high dispersion of anomaly values.
• The IPMS showed the lowest percentage of anomalies (2.45%) at 80%VED.
• Strong increase in anomalies to 93.83% at 123% VED_nom and a dispersion of the anomaly values with a reduction in VED.
• Low correlation between porosity and anomalies, as porosity only fluctuates with high energy input: this indicates a narrow process window.

3.3. Comparative Analysis

• Sensitivity:
  • PrintRite3D shows higher sensitivity in the nominal range with stronger deflections in the event of VED deviations.
  • Trumpf IPMS has a stable process window for high energy input, but is less sensitive to VED changes, as sensitivity in the nominal range is relatively inaccurate (see Figure 3a).
• Process window observation:
  • PrintRite3D is optimized for a narrower process window, resulting in faster responses to deviations.
  • Trumpf IPMS allows for greater VED variations before significant anomalies are detected. It is important to question whether the sensitivity in the nominal range is sufficiently accurate.
• Correlation with porosity:
  • Both systems show a general correlation between detected anomalies and actual porosity.
  • PrintRite3D shows a stronger response to small changes in VED (Figure 3b), which could potentially lead to a higher false positive rate.

The comparative behavior shown in Figure 3 demonstrates that the ESC methodology is capable of revealing distinct sensitivity profiles between different IPMS implementations. This confirms its suitability as a diagnostic and benchmarking tool rather than a system-specific calibration. These results underscore the importance of application-specific selection and configuration of IPMSs in additive manufacturing.

4. Critical Analysis and Outlook

This study presents a system-independent evaluation method for IPMSs in additive manufacturing. The ESC approach combined with the sensitivity curve is discussed in a direct comparison of two different IPMSs for detecting process signature deviations. The results underscore the importance of a holistic evaluation, addressing standardization needs identified in the recent literature [21,22,23,41].

4.1. Methodological Advances and Limitations

This study presents a proof-of-concept implementation of the ESC methodology, focusing on demonstrating its functional principle and comparability rather than full parameter-space validation. Further experimental campaigns are recommended to expand the data basis and confirm reproducibility across different alloys and process conditions.

The ESC methodology represents a significant advancement in IPMS evaluation by addressing several fundamental challenges in the field. The systematic VED variation approach allows for controlled experimentation that was previously difficult to achieve due to the proprietary nature of AM process parameters. However, several limitations must be acknowledged:

Scope of Defect Types: While this study focuses on porosity as the primary quality indicator, real-world AM processes generate various defect types including surface roughness variations, residual stress accumulations, and microstructural heterogeneities. Future iterations of the ESC methodology should incorporate multi-criteria evaluation frameworks to assess IPMS sensitivity to these diverse quality characteristics.

Material Dependency: Although the ESC approach is designed to be material-independent, the current validation is limited to AlSi10Mg. Different materials exhibit varying thermal properties, absorption characteristics, and defect formation mechanisms that may influence IPMS performance. Comprehensive validation across a broader material spectrum is necessary to establish universal applicability.

Process Parameter Interactions: The current methodology focuses on scan speed variation while keeping other parameters constant. However, industrial AM processes involve complex interactions between laser power, scan speed, hatch spacing, and layer thickness. Future developments should explore multi-parameter variation matrices to better represent industrial process conditions [45,46].

Despite the promising results and the recognizable sensitivity of both IPMSs to VED changes, it is important to emphasize that the detection probability of process signature deviations in their current form is not sufficient for the systems investigated to be used exclusively for quality assurance in additive manufacturing [2,4,15].

The following aspects should be taken into account:

• Data collection: More extensive and diverse data sets are needed to validate the performance of IPMSs under different conditions and for different materials [15,18,25,47].
• Adaptation of IPMS metrics: Current anomaly detection metrics need to be refined and possibly expanded to improve the accuracy and reliability of defect occurrence probability [25,26,27].
• Correlation with mechanical properties: Future studies should investigate the relationship between the anomalies detected by IPMSs and the resulting mechanical properties of the components [32,44,48].
• Machine learning and AI: The implementation of advanced algorithms could improve the interpretability of IPMS data and lead to a more accurate defect occurrence probability [25,28,49,50].

A decisive step for the future of IPMS technology is the development of standardized procedures for evaluating different systems. The procedure presented in this paper provides an initial basis for such a standardized evaluation procedure. By establishing a uniform evaluation standard, the following will be possible in the future:

• Developments in various IPMSs can be compared objectively.
• Advances in IPMS technology can be quantified more accurately.
• Manufacturer-independent performance comparisons can be carried out.
• The selection of suitable IPMSs for specific applications is made easier.
• A better basis for regulatory decisions and certifications is created.

Further development and refinement of this approach could lead to an industry-wide accepted standard. This would not only improve the comparability of IPMSs but also increase the speed of innovation in this area, as developers would have clear benchmarks against which to measure their progress.

The implementation of such a standardized evaluation procedure will be an important step towards unlocking the full potential of IPMS in quality assurance and process optimization in additive manufacturing and further advancing the technology.

4.2. Summary of Contributions

This study has successfully developed and validated a novel, system-independent methodology for evaluating IPMSs in laser powder bed fusion additive manufacturing. The ESC approach addresses a critical gap in the field by providing the first standardized method for objective IPMS comparison across different vendor platforms and technologies.

Key Scientific Contributions:

• Methodological Innovation: The ESC methodology represents a paradigm shift from vendor-specific evaluation approaches to a universal framework that enables direct comparison of IPMS performance while protecting proprietary process parameters.
• Empirical Validation: Comparative analysis of two industrially relevant IPMSs (PrintRite3D and Trumpf) revealed significant performance differences in sensitivity and observation windows, demonstrating the methodology’s effectiveness in identifying system-specific characteristics.
• Standardization: The proposed ESC methodology contributes to the ongoing efforts toward standardization of IPMS evaluation by providing a conceptual framework and initial experimental evidence that can inform future inter-laboratory and inter-material studies.

Future work will extend the ESC approach to additional materials and process parameter combinations, integrating multi-sensor data fusion and machine-learning-based anomaly classification. Furthermore, inter-laboratory validation studies are planned to establish reproducibility and support standardization efforts (e.g., within ISO/ASTM TC 261).