[IFC Tunnel][Geotechnics] Modeling of IfcGeoScienceObservation

[larswik]

Related issue is #408.

The current list of proposed predefined types for IfcGeoScienceObservation is a bit bloated and also capture different aspects. This structure needs to be revisited. List of currently proposed predefined types:

• INSITUTEST : A test carried out on site directly in place, e.g. in a borehole, a gallery or from the surface. Distinguished from geophysical surveys which are more complex, see separate description.
• LABTEST : A test on a soli (geologic) or fluid specimen carried out in a laboratory
• BOREHOLEDATAREPRESENTATION : Any kind of information related to intervals or points on the borehole axis
• GEOLOGINTERVAL : Typical geological information contained an a logging sheet like lithology, weathering, RQD, core recovery,...
• BOREHOLECONSTRUCTION : Installed casing, material, filter segments, different casing and drilling diameters,... (This does not sound like anything close to an observation)
• GEOPHYSICALINTERVAL : Typical information collected in continuous geophysical logs, e.g. Pwave velocity, Resistivity, Gamma ray, Density, Caliper, Acoustic and optic scan,…Distinguished from geophysical surveys that measure the conditions in greater distance around the borehole
• MWDINTERVAL : "Measure While Drilling" data are collected during destructive drilling in short intervals and conatin information like water flow, advance rates, drilling resistence,...For exchanged in IFC, interpreted MWD-data or key-attributes are expected
• BOREHOLETESTREPRESENTATION : Interval or point representing the geometry /location of the test results
• GEOPHYSICALSURVEY : No definition (Remove and only keep the specific types below)
• SURFACEGEOPHYSICALSURVEY : Geophysical surveys conducted at or near the ground surface
• BOREHOLEGEOPHYSICALSURVEY : Larger scale geophysical test performed from inside a borehole, distinguished from detailed geophysical log because the ground in greater distance around the boreholes is beeing investigated. Distinguished from logging of the borehole wall with geophysical methods, e.g. sonic or gamma log.
• SAMPLE : OGC GeoSciML.Specimen : A Specimen is a physical sample, obtained for observation(s) carried out ex situ, sometimes in a laboratory.
• GEOPOINTOBSERVATION : No definition. Proposal: Observation at a point regarding geological or geotechnical phenomena such as discontinuities, groundwater, weathering, rockbursts etc.
• MAPPEDUNIT : Part of a MappedZone which was assigned to distinct category of "Book B" models, e.g. a GeologicalUnit, GeotechnicalUnit or fault zone, including detailed observations like weathering, lithologies, discontinuity spacing etc.
• LINEATION : Linear elements / measurements of e.g. striae on a fault surface, mineral lineations,...
• FOLDAXIS : Linear structure that defines the geometry of a fold structure
• DISCONTINUITYSURFACE : shape of a structure like a contact or tectonic fault mapped as a planar feature or trace line

[larswik]

Revised proposal (keeping mainly the base classes): INSITUTEST : A test carried out on site directly in place, e.g. in a borehole, a gallery or from the surface. Distinguished from geophysical surveys which are more complex. LABTEST : A test on a soli (geologic) or fluid specimen carried out in a laboratory BOREHOLEOBSERVATION : Any kind of observation or measurement result related to intervals or points on the borehole axis GEOPHYSICALSURVEY : Geophysical surveys conducted either at or near the ground surface or inside boreholes SPECIMEN (why not use the GeoSciML term?): A physical sample, obtained for observation(s) carried out ex situ, sometimes in a laboratory. GEOSCIENCEOBSERVATION : Distinctly mapped structures that have been observed on MappedZones such as lineation, fold axis, discontinuities etc. (to be consistent, I removed the subclass types, e.g. LINEATION & FOLDAXIS and only kept the "mother").

[mbeaufils]

SAMPLE / SPECIMEN (why not use the GeoSciML term?): A physical sample, obtained for observation(s) carried out ex situ, sometimes in a laboratory.

In OGC Geotech IE we use the term MaterialSample (it seems that some people have strict definitions of samples vs specimen that are not shared by everyone). Also a Sample / specimen is not an Observation and so shall not be a type of IfcGeoscienceObservation.

BOREHOLECONSTRUCTION : Installed casing, material, filter segments, different casing and drilling diameters,... (This does not sound like anything close to an observation). We propose that BOREHOLECONSTRUCTION is just a separate Pset connected to IfcBorehole and not an observation type.

Agree.

INSITUTEST : A test carried out on site directly in place, e.g. in a borehole, a gallery or from the surface. Distinguished from geophysical surveys which are more complex. LABTEST : A test on a soli (geologic) or fluid specimen carried out in a laboratory

This is our classical way of splitting in two the Geotech O&M familly. LabTest shall be linked to a sample / specimen. InSituTest shall be linked to either a borehole, or observedzone / mappedzone.

BOREHOLEOBSERVATION : Any kind of observation or measurement result related to intervals or points on the borehole axis

Borehole observations can be considered as Insitu tests, so the term should be changed. By BoreholeDataRepresentations and BoreholeTestRepresentations we wanted to introduce the capacity to build (hopefully 3D) representations of Borehole Logs. I would suggest the term "BoreholeLogRepresentation". Perhaps if could be defined as an IfcLinearElement?

GEOSCIENCEOBSERVATION : Distinctly mapped structures that have been observed on MappedZones such as lineation, fold axis, discontinuities etc. (to be consistent, I removed the subclass types, e.g. LINEATION & FOLDAXIS and only kept the "mother").

It would be strange to have GEOSCIENCEOBSERVATION as a type of IfcGeoscienceObservation. Also I imagine the objective is to provide an equivalent to what we called "ObservedObjects" in our miro board, isn't it?

GEOPHYSICALSURVEY : Geophysical surveys conducted either at or near the ground surface or inside boreholes (crossholes)

Semantically, geophysical surveys are insitu tests, performed either on a TestZone/ObservedZone or in a borehole. I think we wanted here to propose an equivalent of BoreholeLogRepresentations but for CrossHoles. So I propose to be more explicit here. We may also want to mention that Geophysical surveys in 1 borehole should be dealt via InSituTest and/or BoreholeLogRepresentation.

[j-weil]

We discussed this proposal in the GeoSubgroup based on Mickaels answer above, and share ourt common feedback:

SAMPLE / SPECIMEN (why not use the GeoSciML term?): A physical sample, obtained for observation(s) carried out ex situ, sometimes in a laboratory.

we propose to use the term MaterialSample (it seems that some people have strict definitions of samples vs specimen that are not shared by everyone). Also a Sample / specimen is not an Observation and so shall not be a type of IfcGeoscienceObservation. --> Type of SpatialGeoObsElement

BOREHOLECONSTRUCTION : Installed casing, material, filter segments, different casing and drilling diameters,... (This does not sound like anything close to an observation).

We propose that BOREHOLECONSTRUCTION is just a separate Pset connected to IfcBorehole and not an observation type. Agree.

INSITUTEST : A test carried out on site directly in place, e.g. in a borehole, a gallery or from the surface. (Not part of the definition in IFC, but only explanatory: Distinguished from geophysical surveys which are more complex.) LABTEST : A test on a rock/soil (geologic) or fluid specimen carried out in a laboratory

This is our classical way of splitting in two the Geotech O&M family. LabTest shall be linked to a sample / specimen. InSituTest shall be linked to either a borehole, or observedzone (GeotechIE) / MappedZone or TestedZone (IFC). --> is this linkage going to happen via the existing options for IfcSpatialZones?

BOREHOLEOBSERVATION : Any kind of observation or measurement result related to intervals or points on the borehole axis

We suggest the term "BoreholeLogRepresentation" (or just “BoreholeLog”?) How does this fit to the concept of “IfcLinearElement”?

GEOSCIENCEOBSERVATION : Distinctly mapped structures that have been observed on MappedZones such as lineation, fold axis, discontinuities etc. (to be consistent, I removed the subclass types, e.g. LINEATION & FOLDAXIS and only kept the "mother").

It would be strange to have GEOSCIENCEOBSERVATION as a type of IfcGeoscienceObservation. -->Use “IfcGeoscienceObservedFeature”? This is seen as an equivalent to what we called "ObservedObjects" in our miro board, isn't it? We think it is required to have at least two subtypes:

• One that can cover complex mapped structures, such as our “MappedUnit”, DiscontinuitySurface,… --> Proposal “GeoscienceMappedFeature” (would be in line with OGC concept)
• A second one for information that is just attached to a “dummy-object” like a point, e.g. discontinuity measurements (“structural data”) and others
• --> proposal: “GeoscienceLocalInformation”

GEOPHYSICALSURVEY : Geophysical surveys conducted either at or near the ground surface or by using boreholes changed definition, add: "and measuring the whole volume in between (crosshole)"

Semantically, geophysical surveys are insitu tests, performed either on a TestZone/ObservedZone or in a borehole. We adjusted the definition as above, to make clear we want to support other surveys e.g. outside a borehole.

We may also want to mention that Geophysical surveys in 1 borehole should be dealt via InSituTest and/or BoreholeLogRepresentation.

[j-weil]

Here is our fwork file with better formatting 220623_GeotechSubgroup Answer.docx :

[larswik]

MATERIALSAMPLE: There is no IFC entity SpatialGeoObsElement. We have collected everything for SpatialGeoObsElement into some existing entities:

• IfcSpatialZone : Re-used and extended predefined types for mapped zones and tested zones
• IfcBorehole : Re-use with slight changes in semantics and added predefined types according to geo taxonomy
• IfcSensor : Re-use - I am not sure about the status regarding new predefined types.

New entities in the IFC structure are IfcGeoScienceObservation (Book A), IfcGeoScienceModel and IfcGeoScienceFeature (Book B). Besides that, we also re-use IfcLinearElement with new subtypes to cover for e.g. GeotechTypicalSection. This provides a bit of an issue on where to put MaterialSample (if it is not an observation). For GeoScience, we have the structure:

• IfcProduct
  • IfcElement
  • IfcGeoScienceElement
    • IfcGeoScienceModel
    • IfcGeoScienceFeature
    • IfcBorehole
  • IfcObservation
    • IfcGeoScienceObservation
  • IfcLinearElement
  • IfcLinearZone
    • IfcGeotechTypicalSection

If "MaterialSample" can't be a kind of IfcGeoScienceObservation, it still needs to be an IfcGeoScienceElement. Does it deserve its own IFC entity as a subtype of IfcGeoScienceElement?

[mbeaufils]

From my point of view:

New entities in the IFC structure are IfcGeoScienceObservation (Book A), IfcGeoScienceModel and IfcGeoScienceFeature (Book B)

I did not get that an IfcGeoScienceFeature was planned. I am personally happy with it if this is the replacement for IfcGeoscienceInterpretedObject.

If "MaterialSample" can't be a kind of IfcGeoScienceObservation, it still needs to be an IfcGeoScienceElement. Does it deserve its own IFC entity as a subtype of IfcGeoScienceElement?

Yes. MaterialSample as a subtype of IfcGeoscienceElement or even subtype of IfcElement. In a nutshell, it should be at the same level as IfcBorehole. Is IfcBorehole a subtype or IfcGeoscienceElement? Some may argue that Borehole and MaterialSample are important for geoscience (> IfcGeoscienceElement) but others could say they can also be useful for other purposes (> IfcElement), as IfcSensor…

[larswik]

BOREHOLECONSTRUCTION : Ok!

INSITUTEST : Ok!

LABTEST : Ok!

BOREHOLELOG : Ok! (it is not considered a linear element since that is more "longitudinally/along alignment" oriented "stuff")

GEOSCIENCEOBSERVATION/IfcGeoScienceObservation/IfcGeoScienceObservedFeature : Since predefined types does not allow subtyping, my proposal is to keep only the specific types in this case, i.e. MAPPEDFEATURE and LOCALINFORMATION ("GeoScience" is already in the name of the entity itself). Rename IfcGeoScienceObservation to IfcGeoScienceObservedFeature if that feels as a better term.

GEOPHYSICALSURVEY : Ok with modified definition: "Geophysical surveys conducted either at or near the ground surface or by using boreholes and measuring the whole volume in between (crosshole)"

[larswik]

Is IfcBorehole a subtype or IfcGeoscienceElement?

Yes it is. I forgot in my comment above. Updated now. However, IfcSensor is an IfcDistributionControlElement that was introduced in IFC 4.0 and not GeoScience-specific. Not moving that one.

[larswik]

MaterialSample: Could we use IfcMaterialDefinition (e.g. IfcMaterialConstituentSet) and associate to an IfcGeoScienceObservation? Change definition for IfcMaterial? Current definition: "IfcMaterial is a homogeneous or inhomogeneous substance that can be used to form elements (physical products or their components)" Changed definition: "IfcMaterial is a homogeneous or inhomogeneous substance that can be used to form elements (physical products or their components) and describe natural samples (or their composites)."

[larswik]

Summary: BoreholeConstruction : Pset_BoreholeConstruction

IfcGeoScienceObservation : New Entity (subtype of IfcObservation) PDTs: INSITUTESTRESULT: Result from a test carried out on site directly in place, e.g. in a borehole, a gallery or from the surface. LABTESTRESULT: Result from test on a rock/soil (geologic) or fluid specimen carried out in a laboratory BOREHOLELOG: Any kind of observation or measurement result related to intervals or points on the borehole axis MAPPEDFEATURE: Distinctly mapped structures that have been observed on MappedZones such as lineation, fold axis, discontinuity surfaces etc. LOCALINFORMATION: Other observations made locally (e.g. at a point) such as discontinuities, water inflow, weathering, rockburst etc.
GEOPHYSICALSURVEYRESULT : A systematic collection of geophysical data that was gathered either at or near the ground surface or by using boreholes and measuring the whole volume in between (crosshole). (Changed name and definition so it fits more with the result of the survey rather than the process of surveying) - Please check!

MaterialSample (proposal): Attach an IfcMaterialDefinition (e.g. an IfcMaterialConstituentSet) to e.g. IfcGeoScienceObservation/INSITUTEST or IfcGeoScienceObservation/LABTEST.

[mbeaufils]

MaterialSample: Could we use IfcMaterialDefinition (e.g. IfcMaterialConstituentSet) and associate to an IfcGeoScienceObservation? Change definition for IfcMaterial? Current definition: "IfcMaterial is a homogeneous or inhomogeneous substance that can be used to form elements (physical products or their components)" Changed definition: "IfcMaterial is a homogeneous or inhomogeneous substance that can be used to form elements (physical products or their components) and describe natural samples (or their composites)."

Be careful to material meanings in MaterialSample vs MaterialConstituent. In MaterialSample, "material" is used to oppose to "conceptual". It means "physical" sample. Whereas in MaterialConstituent this is what the "thing" is made of.

I still support the idea to have MaterialSample as an IfcGeoscienceElement.

[j-weil]

I agree to Mickael, supporting to have MaterialSample as a Gwoscience element. From my experience IFC Material is used in a different context and usually refers to clearly defined materials with often standardized properties. We are dealing with unknown material in the first place, with the option to attach a wide ränge of loobservations or measurements to it

[SergejMuhic]

@j-weil and @mbeaufils could you explain the exchange scenario where you would use a material sample? By this I mean, what sort of a data model would this be? An instance sketch of this exchange model would help a lot.

Some additional questions:

1. How does a sample fit into the built environment?
2. Who would produce these IFC files?
3. Which stakeholders would receive such IFC files?
4. What are the conceptual requirements of MaterialSample? i.e. does it require geometry and any other representation concepts such as colour or texture ..., identity, to which other objects is it related to, where is it supposed to be allocated in the project structure Site - Facility - FacilityPart, version history, classification, material and typing
5. In which phases (design, build, operations) of a tunnel project is this data supposed to be exchanged?

It is really not clear how this concept fits into BIM exchanges/use cases, who is supposed to produce the data, how is it supposed to be received and incorporated into the facility model, when is it supposed to be produced and when exchanged. We would need a lot more input.

[j-weil]

1. How does a sample fit into the built environment?

Not sure what “built environment” means, but the “sample object” carries the information on physical or chemical parameters that have been measured on a laboratory. It represents the “piece of ground” that was taken to the lab.

2. Who would produce these IFC files?

Either the laboratory that id the measurement and investigation, the contreactor for site investigation works or the responsible geologists that digitizes the results

3. Which stakeholders would receive such IFC files?

4. The ingeneering geologist that produces the ground model and does the interpretation

5. A checking engineer or infrastructure owner

6. What are the conceptual requirements of MaterialSample? i.e. does it require geometry and any other representation concepts such as colour or texture ..., identity, to which other objects is it related to, where is it supposed to be allocated in the project structure Site - Facility - FacilityPart, version history, classification, material and typing

Geometry – it requires a location, can be a point, or a dummy-geometry Color: can be used to represent a certain measured value (= key property) lie e.g. compressive strength Identity: ? Relation: place where it was collected, SpatialGeoObsElement, and Psets with results and meta data depending on the tests that have been performed on it

6. In which phases (design, build, operations) of a tunnel project is this data supposed to be exchanged?

design: yes, important part of site investigation build: yes, to check measured properties against prediction (prognosis model) operation: the same concept could be used for samples of asphalt, concrete,… that are taken in road and tunnel maintenance

[SergejMuhic]

From your answers, I have to stick to the proposal above. You are talking about Materials in IFC.

Expecting that laboratories would produce IFC files is in my view a bit far fetched. Maybe a bit more than a bit. 😄 There are other materials of which the characteristics come from laboratories but none of them incorporate the samples that were used for testing.

In a nutshell, the requirements listed imply that you need the material characteristics which are located somewhere. That is a typical IfcProduct (something spatial or physical - in our case physical) with an assigned material. I do not see other requirements.

[mbeaufils]

Completing what @j-weil wrote, but maybe focusing more on the perspective for BIM / IFC.

First : A MaterialSample aka Sample/Specimen is something collected from the field (piece of soil, rock, fluid) that is then studied in a laboratory. Several tests will be performed on them to identify their characteristics. In a MaterialSample, several materials (here with the same meaning as IfcMaterial) can be found. They can then be studied separately.

For us, IfcMaterial would be a property of an object, a kind of geoscience observation. Whereas with the expected IfcMaterialSample we would like to have a concept that represent the physical object that has been collected on the field and to which we expect to associate several lab tests.

The MaterialSample would be positionned where it has been collected. It will be a point. We also expect to be able to describe sub sampling: MaterialSamples built from MaterialSamples.

Expected benefits of that approaches:

• Highlights dependencies between observations (if they are made from the same MaterialSample),
• Reduction of the amount of info attached to the geoscience observation (provision of the Id of the MaterialSample, instead of repeating the geoscience observation location),
• Reduction of the amount of the objects at the same place: as a lot of geoscience observations are made from the same borehole or materialSamples. Having a geometry for each geoscience observation would lead to a lot of objects at the same location.

Also I am not sure that laboratories will use IFC for sharing their data, yet in terms of exchange scenario when Book A data are exchanged (in order to be inputs for Book B) the description of the MaterialSamples should be included. Geomodellers would be interested in seing the Boreholes, MaterialSamples sampling locations, ObservedZones and then discover the associated geoscience observations.