Printing information modeling (PIM) for additive manufacturing of concrete structures

,


Introduction
Digital manufacturing technologies, such as additive manufacturing (AM), have been at the foreground of efforts to automate and digitalize the architecture, engineering, and construction (AEC) industry in the context of Industry 4.0.AM is defined by the ISO/ASTM standards as a process of producing objects from 3D model data by joining materials layer upon layer [1].With the application of AM in construction, customization and freeforming may be obtained, and challenges in construction related to performance, productivity, and sustainability may be solved [2].By deploying printable construction materials, such as concrete, AM methods have been adapted to print large-scale components and structures, translating into a reduction of construction time, of cost, and of environmental impact.AM of concrete structures, also known as concrete printing, is the most common application of AM in construction.AM methods for concrete printing can be classified into material extrusion, particle-bed binding, and material jetting [3], where extrusion-based and particle-bed-binding-based methods are more mature in terms of technological development.In extrusion-based methods, printing material with self-hardening properties is deposited layer upon layer, through nozzles or printheads, in a continuous or discontinuous manner [4].In particle-bed-binding-based methods, the printing material (i.e., dry particles) is introduced into the printing platform in thin layers and selectively bound by applying a fluid phase [5].In material jetting methods, the printing material is sprayed through nozzles or printheads, usually from a distance, on a layer-by-layer basis [6].The combination of AM methods and cementitious materials has led to innovative AM systems for fabricating concrete elements and structures without using formwork.Paolini et al. [7] present a review of the application of AM in construction, where AM systems developed for concrete printing are discussed.Established AM systems, such as contour crafting [8] and D-Shape [9], have paved the way for further developments in AM of concrete structures.
Recent research has focused on developing AM systems [10] [11], on advancing printing strategies and optimizing process parameters [12], on material development and characterization [13] [14], on optimizing topologies to minimize material waste [15], on numerical modeling and simulations [16], and on including reinforcement to improve ductility [17].Buswell et al. [18] have reviewed technological issues that affect extrusion-based concrete printing, drawing attention to open research points.To increase the understanding of the underlying physics governing extrusion-based concrete printing, Mechtcherine et al. [19] have given an overview of material behavior of fresh concrete during the manufacturing process.Furthermore, Perrot et al. [16] have implemented analytical and numerical tools to assess processes in concrete printing as a function of the material properties, the geometry of the components, and the process parameters (e.g. machine settings).However, data modeling for AM of concrete structures has not kept pace with concrete printing research and development.
Current AM data modeling approaches are based on standardized data exchange considerations described in ISO/ASTM 52950.The digital workflow, described as a dataflow from 3D digital models to physical components, is specified together with the most common data exchange formats, such as the standard tessellation language (STL), the additive manufacturing format (AMF), and the 3D manufacturing format (3MF) [20].Additional data exchange formats are used for numerical control of AM systems based on G-code (ISO 6983-1) and on the standard for the exchange of product model data compliant numerical control (STEP-NC), which extends ISO 10303 standards in ISO 14649 [21].The standard data modeling approaches for AM may cause information breaks along the modeling process by decomposing digital computer models into several data formats, resulting in information loss and inconsistencies.Efforts to advance AM data modeling approaches have been carried out by developing AM-related formal descriptions, such as ontologies and semantic models, that encompass the current knowledge in the field (or subfields) of AM.In [21], a data model for AM technologies has been defined to improve the adoption of STEP-NC in AM systems by unifying the dataflow, from design to manufacturing, in a single file.Similarly, ontologies have been developed to support manufacturability analysis [22], interoperability for data management [23], lifecycle data management [24], and data provenance in metal-based AM [25].Furthermore, efforts to advance and couple data modeling in AM with digitalization approaches in the AEC industry, such as building information modeling (BIM), have started to gain attention.In [26], a framework based on fabrication information modeling is developed to integrate AM planning into BIM workflows considering material, machine, and process parameters.
In [27], a methodology for knowledge-driven decision support system based on BIM concepts is developed and represented as an application ontology.Integrating lean production principles into the concrete printing process has also been discussed as an approach to address needs regarding bidirectional information flows and databases for design-to-construction workflows in AM of concrete structures [28].
Current AM data models, however, assume materials that can be controlled by assuring constant process settings (e.g.print speed), well-defined material properties (e.g. grain size), and controlled environmental conditions (e.g.temperature), which differs from the reality of extrusion-based concrete printing [29].When considering cementitious materials, which present time-dependent rheological behavior, material-process interactions are crucial when defining process settings and toolpaths, as discussed in [18].Furthermore, material properties of concrete change over time, having a detrimental effect on the quality of the printed components when material property variations are unaccounted during manufacturing.Material-process interactions affecting the overall structural properties of printed components are still to be seamlessly integrated into data modeling approaches used in the digital workflow in AM of concrete structures, to ensure the success of the concrete printing project.Current data modeling approaches limit AM of concrete structures to a process with a long trial-and-error learning curve to determine the ideal process and material parameters.Therefore, semantic descriptions are necessary to advance data modeling approaches for AM of concrete structures, formalizing materialprocess interaction.Also, new data models are required, enabling smoother digital workflows, describing AM input parameters, and allowing monitoring of material-process interactions in real time.
This study aims at formally describing the digital workflow, the input parameters, as well as interactions between the parameters in concrete printing, following a semantic modeling approach, referred to as "printing information modeling" (PIM).The PIM approach represents a step necessary to standardize processes information, material information, and geometry information into a unified data model.In this regard, a "printing information model", i.e. a semantic model, for extrusion-based AM of concrete structures is proposed building on previous studies from the authors [30], [31].A software tool that deploys building information modeling (BIM) concepts is developed to test the printing information model as a formal basis for BIM-based concrete printing.This paper is structured as follows.Section 2 gives an overview of the system and process analysis and requirements, integrating the results of a survey conducted in this study among practitioners and researchers.In Section 3, the proposed printing information model is presented, describing parameters and fundamental material-process interactions for AM of concrete structures.The semantic model is validated via a software tool and an example case.In Section 4, the results are discussed in the context of BIM-based concrete printing.The paper concludes with a summary of this study and an outlook on potential future work.

System and process analysis
An analysis of the concrete printing process is conducted to trace the information along the digital workflow as well as to identify interactions between tasks, actors, and system elements, serving as basis for the printing information modeling.The analysis is carried out in the context of the AEC industry using BIM concepts to identify information exchange requirements and to preserve semantic information, improving the interoperability along the digital workflow.The concrete printing system and process analysis follows the methodology developed for BIM information delivery manuals [32], in which processes are discretized into tasks and the information exchanged between the tasks (i.e.inputs and outputs) are identified as information exchange requirements.In addition, a survey is conducted among concrete printing practitioners and researchers, whose results are integrated into the system and process analysis.The survey aims to address the following research points.
 Common types of printing systems and software applications currently used for concrete printing.
 Work areas, actors, and tasks along the digital workflow of concrete printing.
 Information exchanged among the actors (e.g.inputs and outputs) and material-process interactions observed or relevant for each work area.
The survey is designed as a questionnaire and shared online among 20 practitioners and researchers that have experience and publications in the domains of concrete printing, including design/architecture, civil engineering, mechanical engineering, material science, and robotics.The questions are formulated as a set of multiple choice and open-ended questions regarding (i) printing systems and software applications employed for concrete printing, (ii) overview of digital workflows, roles, and areas relevant to the work of the practitioners and researchers, as well as (iii) information required and generated in their work areas.The open-ended questions are intended for the practitioners and researchers to elaborate further the multiple-choice answers.The results of the survey have been collected and processed to complement the system and process analysis with practical knowledge.Overall, the system and process analysis provides insight into the information exchanged along the digital workflow of concrete printing.Hence, the information exchange requirements provide the basis to translate concrete printing parameters and concepts into a general valid semantic model.In the following subsections, the concrete printing system and the concrete printing process are described and analyzed.Then, the information exchange requirements are identified.

Concrete printing system
In extrusion-based concrete printing, concrete is mixed and transported to a printhead, which is attached to a printing system (e.g.gantry and robot).Concrete is extruded through a nozzle, located at the tip of the printhead, and it is deposited in place to build a component from a digital 3D model.The elements in concrete printing systems are identified, using a robot concrete printer as an illustrative example.As shown in Figure 1, the robot concrete printer is comprised of (i) a system command and a controller (control unit), (ii) a robotic arm and a printhead (printing system), and (iii) a mixer and pump system (material transportation system).

Figure 1. System elements of a robot concrete printer
 Control unit: In the control unit, the system command prepares the data necessary for printing, such as user-defined inputs and generating machine instructions.The system command communicates with the controller via human-machine interfaces.The controller handles the kinematics and the electromechanical aspects of the robotic arm and printhead.The controller also processes the machine instructions that define toolpaths and process parameters, controls the multiple axes executing motions of the robotic arm and the printhead, and monitors the performance of the concrete printer using sensing technologies.Depending on the complexity of the printing system, the controller may also handle the material transportation system.
 Printing system: In the printing system, the robotic arm facilitates the deposition of concrete at desired locations with desired speeds under desired angles.The printhead, as the end effector of the robotic arm, is an element used to extrude concrete, and it consists of an extruding mechanism and a nozzle.The extruding mechanism is a series of parts of the printhead that pushes the concrete through the nozzle.The nozzle, the end part of the printhead, is a hollow element that gives shape to the concrete layer as it is deposited in place to build up a component.

 Material transportation system:
In the material transportation system, the mixer mixes the raw materials to obtain concrete and the pump transports the concrete from the mixing unit to the printhead, avoiding segregation and bleeding.
The system elements have a direct effect on the manufacturing process, the printing material, and the quality of the printed component.Concrete should be extrudable and buildable, where each concrete layer, once in place, is capable of retaining shape and of adhering to and carrying the load of subsequent layers.Key rheological properties provide the characteristics necessary for concrete to be printable, defining the evolution of viscosity and yield stress of the material over time.A detailed review of the manufacturing process in extrusion-based concrete printing is presented in [19], including the effects of material parameters (e.g.yield stress) and machine settings (e.g.nozzle height and material flow) in pumping, extrusion, and deposition of concrete layers.In the following subsection, the concrete printing process is described and analyzed.

Concrete printing process
To describe and to analyze the concrete printing process, a process map is developed, where tasks, actors, and information exchange requirements are identified along the digital workflow for concrete printing.Synergies with the standardized requirements and process chain described in DIN SPEC 17071:2019-12 [33] and with data models developed for metal-based additive manufacturing [25] as well as the results of the survey are used as basis to develop the process map.Due to the material-process intractions within concrete printing, material-related tasks are considered as part of the concrete printing process.
Following the business process modeling notation [34], the process map is defined focusing on designing and planning of concrete printing projects, as shown in Figure 2 in terms of an activity diagram.The actors in the activity diagram, defined as designer, engineer, material scientist and machine operator, develop specific tasks or subprocesses and exchange information, following a sequence that translates digital models into printed components.The concrete printing process starts with design concepts and design specifications to generate geometric models and to specify the printing systems (i.e.machines) required to execute the concrete printing project.On the one hand, machine settings are initially defined and used to design the printing material (i.e.concrete) in an iterative process, which may include material testing, until the design specifications are satisfied and material specifications are derived.On the other hand, with respect to designing the geometry, the geometric models are sliced into layers considering machine specifications (e.g.nozzle sizes).For each layer, toolpaths are planned and process parameters are assigned to the toolpath profiles according to the process data and the material specifications.Within toolpath planning, simulations are carried out to evaluate the manufacturing process, including the material behavior, for ensuring build success.As an output of the toolpath planning, AM models are created and, if accepted, used as basis to generate machine instructions (e.g.

CNC code).
Figure 2. Extract of the process map describing the concrete printing process [31]

Information exchange requirements
The information exchange requirements are described according to information units and attributes, which are collected from literature -the interested reader is referred to [31] -and from the survey, i.e. from experienced practitioners and researchers in the domains of design/architecture, civil engineering, mechanical engineering, material science, and robotics.Considerations regarding the material-process interactions between concrete and the manufacturing process are included in the information exchange requirements.Furthermore, fabrication information modeling (FIM) concepts introduced in [35] and BIM concepts explored in [36], [37] are considered when defining the information exchange requirements for concrete printing.Previous experiences in integrating FIM-based frameworks into BIM, presented in [26], provide insights into the interactions within concrete printing when translating digital models into physical components.
The attributes are analyzed according to completeness (i.e. if an attribute is required or optional) and with respect to interoperability, where completeness ensures the inclusion of all attributes necessary to manufacture components, and interoperability ensures a common understanding between tasks and actors.The survey results provide insight into the completeness and interoperability of the attributes.
For illustration purposes, the information exchange requirements for AM models are presented in Table 1, where prerequisites of the information exchange requirements are highlighted in gray.AM models, representing the output of the toolpath planning, encompasses all information necessary to generate machine instructions, including process data and material specifications.Toolpaths are usually described by sets of points defining printing and axes paths.Material-process interactions are represented using toolpath profiles, which are evaluated via material modeling and manufacturing process simulations.
Further information on the information exchange requirements for concrete printing may be found in a previous study of the authors presented in [31].Based on the information exchange requirements, the classes as well as the interactions between the classes necessary for AM data modeling are identified.In the following section, the printing information model for additive manufacturing of concrete structures is presented.

Printing information model for additive manufacturing of concrete structures
In this section, the printing information model is developed, aiming to formalize the information necessary for designing and planning concrete printing projects through a single semantic model based on object-oriented modeling concepts, materialized in the form of a semantic model and an ontology.
Since the terminology with respect to "semantic models" and "models" differ, depending on the field of research, it must be noted that in this study a "semantic model" is considered a metamodel to be instantiated into specific "models", in compliance with [38].In other words, the models, representing extracts of the real world, are instances of the metamodel, which formally defines structure, semantics, and constraints of the models.Metamodels are used as schemas to develop software applications as well as databases and may be extended to support future needs of the extracts of the real world.
The model developed for printing information modeling, hereinafter referred to as PIM model, focuses on extrusion-based AM of concrete structures and describes the semantics of material-process interactions.The printing information model is developed in four steps, (i) conceptual modeling, (ii) formal modeling, (iii) verification, and (iv) validation.
In the first step, aspects and interactions within concrete printing are conceptualized in the form of a knowledge map and structured as the semantic model.The knowledge map aids to categorize input parameters in AM data modeling into process-related, material-related, and geometry-related information.Then, the information exchange requirements defined in Section 2, encompassing the input parameters and interactions within concrete printing, are mapped into the semantic model in terms of classes and interactions between the classes.The semantic model is described using Unified Modeling Language (UML) because of its comprehensibility to engineers and rich semantics [39].In the second step, the semantic model is translated into an ontology that allows knowledge-based reasoning to be used for refining the semantic model.In the third step, the PIM model is verified for correctness, performing logic inference and SPARQL queries on the ontology.SPARQL is the standard protocol and resource description framework (RDF) query language and is used to query information from data sources that can be described based on the RDF standard [40].In the fourth step, the PIM model is validated though a software tool using an example case.In the following subsections, the steps to develop the PIM model are presented.

Conceptual modeling of the printing information model
To conceptualize the information necessary for designing and planning concrete printing projects, aspects and interactions within concrete printing are summarized in the knowledge map and then structured in the semantic model.The knowledge map, shown in Figure 3, abstracts printed components according to process-related, material-related, and geometry-related aspects.Printed components are designed and planned within the context of a building or construction project, with design specifications that consider specific AM methods and printing systems.Printed components, represented by the Component entity, can be described as outputs of manufacturing processes (Process entity), are composed by materials (Material entity), are characterized by specific geometries (Geometry entity), and are situated within specific environments (Environment entity).The manufacturing processes receive printable materials and digital models of the geometry of the components as inputs and are executed according to process parameters.The manufacturing process, to ensure manufacturability, may modify the geometry features of the components and the material properties, due to the effects of the underlying physics of the manufacturing process.Material properties may constrain process parameters, e.g.viscosity limiting the material flow rate, and may modify geometry features, such as overhang angles and deposition deformation.Structural properties of the printed components are related to material properties and to geometry features of the components.For example, the stiffness of printed components is related to the elastic modulus of the material and the inertia of the components.
Furthermore, environmental conditions may modify material properties, e.g.yield strength during curing, as well as structural properties, such as durability.where the AM methods and printing systems are selected before defining geometry features and materials of the components.Three types of interactions are identified, (i) process-geometry interactions, (ii) material-process interactions, and (iii) material-geometry interactions.
 Process-geometry interactions include features constrained due to manufacturability (e.g.print resolution, infill pattern, and supports) [41] and reinforcement solutions [17].Process-geometry interactions may be complemented with FIM concepts, as discussed in [35], to bridge the gap between virtual design tools and digital fabrication tools.
 Material-process interactions include process parameters that are constrained by material parameters and material parameters that are modified by the manufacturing processes.
Following the insight into material-process interactions presented in [12][18], [19], on the one hand, the rheological behavior of concrete constrains process parameters during pumping and extrusion.On the other hand, process parameters can modify material properties, for example by incorporating admixtures or local vibration during extrusion and deposition.Material-process interactions are usually evaluated by modeling the material and simulating the manufacturing process during design and planning concrete printing projects, as has been shown in [16].
 Material-geometry interactions include geometric features that are limited by material properties, such as overhangs and buildability (e.g.stacking height and slenderness), or by changes due to material properties, such as deformation during deposition due to early-age creep and shrinkage.Material-geometry interactions have been briefly discussed in [18].
Building upon the knowledge map as an outcome of the system and process analysis, which aids to categorize the input parameters and interactions in AM data modeling, the UML representation of the PIM model is developed.As mentioned above, the main parameters involved in concrete printing are coherently categorized into classes, and the interactions between the classes are described with semantic relationships.Moreover, the PIM model is refined in an iterative process based on knowledge-based reasoning.As a result, the PIM model is an understandable and instantiable metamodel, which can be instantiated for BIM-based concrete printing in compliance with the Industry Foundation Classes (IFC) standard.
The PIM model is based on the system and process analysis presented in the previous section.When designing the PIM model, three simplifications are assumed to ensure hardware and software independence.First, the geometry information is assumed to be describable using concepts of standardized data models for geometry representations, such as the IFC standard.Second, nesting features in the PIM model can aid the description of support structures necessary for overhangs assuming that the same process method and printing system are employed for the components and the support structures.Third, similar to the support structures, aggregation features in the PIM model can aid the description of reinforcement solutions by aggregating components.Material-process interactions are described in the UML class diagram as semantic relationships, as highlighted in blue in Figure 4.While planning concrete printing projects, material properties must be used as inputs for toolpath planning, and in particular those pertinent to the rheological behavior of concrete that changes over time.Profiles of the material properties may be generated during simulations that allow to evaluate the success of the build.Process boundary conditions depend on the material boundary conditions, that impact both, process constraints and machine constraints.In particular, process constraints may modify the hardened properties of the materials due to the material anisotropy resulting from the layered build up and the effect of loads acting upon the layers.For example, dry environments may have detrimental effects on the material properties, diminishing the yield strength develop during curing and constraining the range of optimum process parameters, such as maximum build height.To provide knowledge-based reasoning, the PIM model is formalized into the ontology in the following subsection.

Formal modeling of the printing information model
Unified Modeling Language, allowing to visualize semantics in great details and being easily comprehensible for engineers, supports the development of software tools and information systems necessary for designing and planning concrete printing projects.However, UML is bounded to close world assumptions and may contain design errors leading to unsatisfiable concepts [42].To enable open world assumptions for knowledge-based reasoning, to provide vocabulary, and to verify the PIM model satisfiability, a printing information ontology, which formalizes the previous UML class diagram to the Web Ontology Language (OWL), is developed.Based on the ontology, the PIM model is refined in an iterative process.The ISO-standardized Basic Formal Ontology (BFO) is used as upper-level ontology, as it provides support for information exchange [43].Classes referring to objects and materials are aligned with the top-level entity BFO:Material_entity, classes referring to geometry representations and sites are aligned with the top-level entity BFO:Immaterial_entity, classes referring to quality information are aligned with the BFO:Quality entity, and classes referring to processes are aligned with the BFO:Occurrent entity.Attributes of the classes are defined as data properties, while multiplicity conditions are described with cardinality axioms.Furthermore, the semantic relationships between the classes are described using class properties, axioms, and description logic (DL) rules.Figure 5 shows an extract of the alignment of the classes with the BFO hierarchy for the printing information ontology, hereinafter referred to as PIM-O.The PIM-O is built following a bottom-up construction strategy, allowing the alignment to other domain ontologies, such as ifcOWL [44].For example, concepts in ifcOWL have been enhanced using BFO as basis to improve interoperability [45].

Verification of the printing information model
The PIM model is verified for correctness using the PIM-O through model checking and conducting competency questions.On the one hand, model checking aims to identify errors in the design of the ontology, such as classes that cannot be instantiated.On the other hand, competency questions are a set of questions that determine the scope or the intention of an ontology and must be able to be answered correctly based on the ontology.Therefore, the PIM-O is checked on a terminological level (through model checking), which delineates classes and relationships, and on an assertional level (through the competency questions), which relates instances to classes.Furthermore, it should be noted that model checking is carried out using logic inference using a reasoner, while the competency questions are queried using SPARQL queries.In the following paragraphs, an overview of the verification is presented Model checking: The model checking tests conducted in this study comprise (i) ontology consistency checking, (ii) concept satisfiability checking, and (iii) concept subsumption checking, as illustrated in   property_GreenStrength  Based on the positive results of the competency questions on the PIM-O, the correctness of the PIM model is verified.Hence, the PIM model, representing a metamodel, can be instantiated into specific models that are used for designing and planning concrete printing jobs.In the following subsection, the PIM model is validated.

Validation of the printing information model
The validation aims to test if the PIM model is suitable to describe process, geometry, and material input parameters as well as interactions between the parameters in AM data modeling of concrete structures.
Building upon previous work of the authors [30], a software tool has been developed using the PIM model as the backbone, and it integrates BIM concepts into the concrete printing processes.As an illustrative example case, a cylindric tank is planned for concrete printing using the software tool to generate toolpaths and to gather inputs for simulating the build-up process.
The software tool is written as a plug-in for BIM platforms for designing and planning concrete printing projects, and it generates specific models (i.e.instances) from the PIM model.The software tool, referred to as "PIM tool", incorporates algorithms for (i) slicing, (ii) toolpath planning, and (iii) CNC code generation.The inputs of the PIM tool are user-defined parameters for process-related information (e.g. machine data, slicing data, and toolpath data) and material-related information (e.g.material specifications and material properties), which extend the BIM models.The PIM tool updates the BIM models as the slicing and toolpath planning are executed, and outputs CNC code as machine instructions for concrete printers.
The cylindric tank is prepared for simulating the concrete printing process, based on a BIM model, as shown in Figure 7.The cylindric tank is designed with an inner radius of 250 mm, a thickness of 60 mm, and a height of 600 mm.The printing system is a robot concrete printer with a printing area of 2,500 mm × 3,000 mm × 4,000 mm (L × W × H), a nozzle size of 25.4 mm, and a printing speed range between 3,000 mm/min and 48,000 mm/min.Based on the nozzle size of the printing system, a layer height of 15 mm and an extrusion width of 30 mm are defined as slicing parameters.Toolpath parameters include a spiral layer-by-layer strategy, with a boundary thickness of two adjacent filaments.A printing speed of 5,000 mm/min is chosen as a machine setting with a layer interval time of 0.31 min.The material-related information described based on the material model presented in [48].Early-age concrete is modeled as a cohesive material with a Mohr-Coulomb failure criterion.The temporal behavior of early-age concrete regarding yield stress (i.e., green strength) is modeled using time-dependent parameters, such as cohesion and Young's modulus, as well as constant parameters, such as internal friction angle, Poisson's ratio, and dilatancy angle.The concrete printing process is simulated by gradually adding layers and updating the yield stress over time.Border conditions include free radial deformation, a fixed support at the bottom layer, and self-weight loading conditions.The input parameters are fed into a finite element model for further analyses.The finite element model is used to perform a deformation analysis to determine the maximum build height that can be achieved with the defined printing strategy and machine settings.
Figure 7. Simulation of the concrete printing process of a cylindrical tank The validation example case demonstrates that the PIM model is suitable for designing and planning concrete printing projects using BIM models, representing components to be printed, as inputs.With the PIM tool, information from BIM models can be processed, enabling model updates.User-defined parameters are used as a basis to execute algorithms and to generate information necessary for simulating the build-up process.Hence, BIM models are extended to include material-related information, such as temporal behavior, and process-related information, such as profiles for each layer.
The PIM model has therefore shown to be suitable to describe parameters and material-process interactions in concrete printing for simple components.In the following section, the results of this study are discussed.

Results and discussion
The PIM model has been conceptualized, formalized, verified, and validated as a metamodel for AM of concrete structures.The PIM model, being a metamodel, has been instantiated into the PIM tool, which is capable of processing and updating BIM models.With the PIM tool, algorithms are implemented to facilitate the design and planning of concrete printing projects based on a single metamodel, improving geometry conformity, manufacturability, and performance.The PIM model has thus shown to be suitable to describe parameters and material-process interactions in concrete printing for simple components and has the potential to describe complex components.
As has been corroborated in this study, efficient solutions to enrich BIM models are software tools implemented as add-ons for BIM software that can interpret, process, and update the information contained in the BIM models.Here, the PIM model, as a metamodel for AM of concrete structures, has easily been instantiated into a software tool written for designing and planning concrete printing projects.The potential of coupling AM data modeling and BIM concepts has already been extensively discussed in literature [7], as BIM models may serve as basis for complete digital models for concrete printing, including processes-related, material-related, and geometry-related information.By enhancing the capabilities of BIM concepts, BIM models may shift from a functional focus to a manufacturing focus [26], following the fabrication information modeling concept discussed above, to match the demands for geometry, material, and process representations in concrete printing.Therefore, it can be concluded that the PIM model has potential to achieve BIM-based concrete printing, serving as a first step towards improving current data modeling concepts currently deployed for concrete printing.

Summary and conclusions
A printing information modeling approach has been proposed, serving as a basis for data modeling for additive manufacturing of concrete structures.A need to formalize the parameters and material-process interactions in concrete printing has been identified and a generic printing information model has been developed.The printing information model, a semantic (or meta) model for AM of concrete structures, describes process, geometry, and material input parameters as well as interactions between the parameters, providing instantiable models for data modeling.To develop the printing information model, a requirements analysis, based on a system and process analysis, has been carried out to identify exchange requirements, bringing light to the input parameters and material-process interactions that must be considered to design and plan concrete printing jobs.The PIM model has been formalized as an ontology, verified for correctness, and validated as an instantiable model.Furthermore, a software tool has been developed using the PIM model as the backbone, where BIM concepts are integrated into the concrete printing processes, showing the potential of the PIM model for BIM-based concrete printing.
In conclusion, with an example case of a printing simulation of a printed concrete cylindrical tank, the value of the printing information model for BIM-based concrete printing could be demonstrated, where the printing information model is used to adequately define the input parameters and material-process interactions for designing and planning concrete printing projects.Still, limitations exist, as simple algorithms for process planning have illustratively been executed in this study, which may require extensions when printing complex components.Future work may be conducted towards coupling digital twin frameworks to extend the PIM model for supporting process monitoring and process control, advancing the adoption of BIM-complainant data models for additive manufacturing.

Figure 3 .
Figure 3. Knowledge map abstracting printed components according to process-related, materialrelated, and geometry-related aspects

Figure 4 Figure 4 ,Figure 4 .
Figure 4 presents an extract of the PIM model in the form of a UML class diagram depicting the structure of a printed component, i.e. classes and semantic relationships.On the one hand, as can be seen from Figure 4, printed components (AMComponent class) are defined as part of buildings (Building class), providing context to concrete printing projects.On the other hand, printed components are described as outputs of AM processes (AMProcess class), have constituent materials (Material class), and have geometry representations (Geometry class).The AMProcess class includes process-related information necessary for slicing, toolpath planning, and machine control in extrusion-based concrete printing.TheMaterial class refers to material-related information necessary for concrete printing that is generalized for materials, such as concrete (main material) and plaster (support material).The Geometry class describes geometry-related information that may be inherited from 3D digital models and that may be generated during planning.A more detailed UML class diagram for the PIM model is presented in

Figure 4 .
Figure 4. Extract of the printing information model (PIM model) in UML representation, with material-process interactions highlighted in blue

Figure 5 .
Figure 5. Extract of the alignment of classes according to the BFO hierarchy for the printing information ontology (PIM-O)

Figure 6 Figure 6 .Listing 1 .
Figure 6.Extract of the printing information ontology, including (a) component view, (b) process view, (c) material view, and (d) geometry view

Formal
expression: PIM-O ⊭ T ⊑ Ʇ  Concept satisfiability checking A concept expression C is satisfiable with respect to the PIM-O, answering the question "is it possible to instantiate a concept C?" Formal expression: PIM-O ⊨ C ⊑ T  Concept subsumption checking A concept expression C is a subsumption of a concept expression D with respect to PIM-O.Formal expression: PIM-O ⊨ C ⊑ D  Competency questions: Applying competency questions is a well-established method for ontology checking [47].Accordingly, competency questions are developed based on the information exchange requirements to check the PIM-O on the assertional level.SPARQL queries are used to answer the competency questions based on the structure and axioms defined in the PIM-O.For illustration purposes, querying of a competency question (CQ) is shown by the example of competency question CQ1 and CQ2.

Table 1 .
Information exchange requirements for AM models, with prerequisites highlighted in gray Process dataThe process data will have been specified prior to developing the additive manufacturing model.The process data includes printing strategy (e.g.layer-by-layer strategy, infill pattern, and nozzle height), boundary conditions (e.g.process constraints and machine constraints), and machine parameters (e.g. printing speed, acceleration, and pump pressure).aggregate size, slump, and open time), support material (e.g.material type and design strength), and reinforcement material (e.g.material type and design strength).

Table 2 ,
[46]g the Pallet reasoner[46].The Pallet reasoner is a tool used to check whether an ontology O satisfies an axiom  (formally written as O ⊨ ).The results of the model checking tests are shown in Table2.As can be observed, the PIM-O positively satisfies the model check tests on the terminological level.

Table 2 .
Model checking results for PIM-O

Table 3 .
The results obtained in the query Q1 show that subclasses of the class AMComponent for walls, beams, columns, support structures, and reinforcement structures.Results of the competency question query results for PIM-O