EP3942451A1 - Création automatique d'un plan de construction - Google Patents

Création automatique d'un plan de construction

Info

Publication number
EP3942451A1
EP3942451A1 EP20717174.5A EP20717174A EP3942451A1 EP 3942451 A1 EP3942451 A1 EP 3942451A1 EP 20717174 A EP20717174 A EP 20717174A EP 3942451 A1 EP3942451 A1 EP 3942451A1
Authority
EP
European Patent Office
Prior art keywords
dimensional
dimensional model
building
model
system arrangement
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
EP20717174.5A
Other languages
German (de)
English (en)
Inventor
Fabian Schmidt
Rolf KRETSCHMER
Frank PELZER
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Raumdichter GmbH
Original Assignee
Raumdichter GmbH
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Raumdichter GmbH filed Critical Raumdichter GmbH
Publication of EP3942451A1 publication Critical patent/EP3942451A1/fr
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2210/00—Indexing scheme for image generation or computer graphics
    • G06T2210/04—Architectural design, interior design

Definitions

  • the present invention is directed to a system arrangement for the automated creation of an error-free, deterministic and three-dimensional building plan.
  • the present invention is also directed to a correspondingly set up method and to a computer program product with control commands which implement the method or operate the system arrangement. According to the invention, it is possible to create three-dimensional building plans on the basis of two-dimensional building plans and to take a complex set of rules into account.
  • DE 10 2016 105 480 A1 shows a method for installing, managing and / or maintaining components in a building complex. In particular, it is provided that installation data or maintenance data are visualized.
  • DE 10 2016 116 572 A1 shows an alignment of point clouds for modeling interior spaces. For this purpose, partial models of the interior of a building structure are created.
  • EP 2 913 796 A1 shows the creation of panorama views in a mobile system.
  • a system arrangement for the automated creation of an error-free, deterministic and three-dimensional building plan is proposed, with an interface unit set up to provide a plurality of two-dimensional building plans of a building, a learning unit set up for iterative and adaptive provision of consistency rules, a transformation unit set up for the empirical transfer of the plurality of two-dimensional building plans into the three-dimensional model of the building, a logic unit set up to apply the consistency rules to the transferred three-dimensional model, an identification unit set up to identify control deviations in the three-dimensional model and an adaptation unit set up to adapt the three-dimensional model in such a way that depending on a empirically determined transformation model, a rule-compliant three-dimensional model results.
  • the proposed system arrangement enables automated creation, since no human intervention is necessary.
  • Two-dimensional building plans can be entered and a three-dimensional building plan results. This is done error-free and deterministically, since a set of rules is provided which enables errors to be corrected or underspecifications to be identified and the system to propose solutions for replenishing.
  • a three-dimensional building plan is generally a virtual one, i.e. digital, building plan.
  • Analogue, two-dimensional building plans are processed digitized. This can imply preparatory procedural steps, for example the digitization of analog construction plans.
  • a plurality of two-dimensional building plans is provided by means of an interface unit. This can imply a component set up in terms of network technology, which obtains corresponding building plans from a server, for example.
  • a plurality of consistency rules are then provided by means of a learning unit. This rule set is provided iteratively and adaptively, which means that rules can be entered repeatedly and existing rules can be adapted. As a result, a learning model is created and the consistency rules are refined in each iteration. Consequently, the system arrangement can be operated iteratively and each processing is followed by a feedback such that the set of consistency rules is adapted.
  • the consistency rules describe a consistent three-dimensional model of a building in abstract form.
  • rules are created that describe when a three-dimensional model has reached a desired state.
  • Examples of such a consistent model are that the outer walls typically have to create an enclosed interior that is only opened via doors and windows. If, for example, a two-dimensional construction plan shows that the walls are not in a form-fitting manner, this is typically a fault. Such a building can generally be constructed, but according to life experience this is not desired and the consistency rules describe that a first wall typically stands orthogonally on a second wall.
  • Another criterion for a consistent model is that every room is accessed by at least one door. If a room does not have a door, it is typically a planning error.
  • Such errors can be identified by a consistency rule and a model can then be proposed which corresponds to the consistency rules.
  • An under-specification can also occur in such a way that a certain position within the plan cannot be seen. Such an under-specification should be avoided and consequently the consistency rule can be used to describe that each component must have a predefined position. If such a sub-specification is identified, an action can be proposed as to how the model is to be adapted in such a way that a deterministic model is created.
  • the majority of two-dimensional building plans are transferred into the three-dimensional model, whereby existing algorithms can be used.
  • the two-dimensional building plans can be put together in such a way that the three-dimensional model is created.
  • the present invention is directed to converting two-dimensional building plans into a three-dimensional model.
  • the invention is also aimed at converting a three-dimensional model into two-dimensional building plans. So it is not only possible to put the two-dimensional building plans together, but rather it is also possible to visually cut up the three-dimensional model, with corresponding cuts in each case generating a two-dimensional building plan.
  • the two-dimensional building plans are transferred to the three-dimensional model, which is initially done empirically. This is done empirically because it is not yet clear whether the two-dimensional building plan is based on an error or whether there is an under-specification. Consequently, this transfer can take place iteratively, in such a way that all possible three-dimensional models arise. If there is a sub-specification, several three-dimensional models are possible and these can all be generated. Then z. B. by means of an artificial intelligence or a neural network, which three-dimensional model is most suitable and corresponds to the consistency rules. It is thus possible to create several three-dimensional models, some of which are rejected again.
  • the three-dimensional model is checked using the consistency rules and deviations from the rules are determined.
  • a control deviation describes the state in which a three-dimensional model does not conform to the consistency rules.
  • This is generally not desired and the three-dimensional model is adapted in such a way that a rule-compliant three-dimensional model results as a function of an empirically determined transformation model. Consequently, transformation steps are performed using artificial intelligence or a neural network and then the transformation model can be selected, which particularly advantageously adheres to the consistency rules with regard to the three-dimensional model. For this purpose, parameters such as efficiency can be taken into account. If several transformation models have been identified empirically, the transformation model can be selected which produces a consistent three-dimensional model with as few changes as possible.
  • the transformation model can be created using artificial intelligence and / or a neural network.
  • a neural network can be trained and various two-dimensional building plans can be converted into three-dimensional models.
  • a correspondingly defined neural network can then be provided, for example by means of a server, and continuously refined.
  • the consistency rules can be mapped using a neural network, and the transformation model can also be mapped using a neural network.
  • the two-dimensional building plans can be generated from a three-dimensional model by means of graphic section formation.
  • the consistency rules can be changed and / or expanded as a function of the adaptation of the three-dimensional model. This has the advantage that the rule set can be adjusted and refined in each iteration, and rules can be added or deleted.
  • the consistent three-dimensional model corresponds to physical properties and / or specified empirical values. This has the advantage that physics can be mapped and empirical values can be modeled, on the basis of which the three-dimensional model is created. In this way, specifications can be entered manually or specifications can be extracted from projects that have already been carried out and reused in future projects.
  • the consistency rules describe a structural analysis. This has the advantage of being ensured It becomes that the resulting model always describes a safe structure and consequently it can be automatically checked whether statics are guaranteed and then a correspondingly faulty model can be automatically adapted.
  • the consistency rules have a number of rules which are at least partially dependent on one another. This has the advantage that a complex set of rules can be mapped and implemented. Typically, this cannot be done by a human processor or rules can be so complex that results would be highly error-prone.
  • the transformation model describes adaptations that modify individual components of the three-dimensional model in such a way that a rule-compliant three-dimensional model is created.
  • This has the advantage that automated adaptations are possible that convert a faulty model or a non-deterministic model into a consistent model. This can be done empirically or can be done here z.
  • B. an artificial intelligence or a neural network can be used.
  • a three-dimensional model can thus be created by trial and error, which conforms to the rules and then a suitable model can be selected from the several possibilities.
  • a model can be suitable if it requires as few calculation steps as possible.
  • the transformation model transforms physical and / or physical properties of the components of the three-dimensional model. This has the advantage that both the material of the individual components is changed can be as well as the position or dimensions. Two-dimensional building plans often have corresponding specifications. For example, a line thickness or a color can be used to define how the component is to be designed. If an error is identified here that affects the structural engineering, for example, it can be adjusted automatically.
  • a system deviation describes an error and / or an underspecification. This has the advantage that it can be clearly defined when there is a control deviation and consequently it is possible to react accordingly.
  • the transformation unit has an interface to already implemented, remote transformation services. This has the advantage that existing implementations can be reused and, in particular, the transformation unit can access a server in terms of network technology, which provides the corresponding functionalities.
  • physical and / or physical properties can be taken from the two-dimensional building plans. This has the advantage that a holistic adaptation of the three-dimensional model can take place and thus it is also ensured that the three-dimensional model is consistent.
  • the three-dimensional model is stored and can be used as a reference in further iterations of creating a further building plan.
  • This has the advantage that one can learn from the individual projects and then corresponding results can be found in other projects can be reused.
  • exemplary three-dimensional models are stored which are already advantageous and these can then serve as templates for further iterations. This means that it is always clear which transformation rules are to be used, since a template already exists which describes how a consistent three-dimensional model should look.
  • the object is also achieved by a method for the automated creation of an error-free, deterministic and three-dimensional building plan, with the provision of a plurality of two-dimensional building plans of a building, an iterative and adaptive provision of consistency rules which describe a consistent three-dimensional model of the building, an empirical transfer the plurality of two-dimensional building plans in the three-dimensional model of the building, applying the consistency rules to the transferred three-dimensional model, identifying system deviations in the three-dimensional model and adapting the three-dimensional model in such a way that, depending on an empirically determined transformation model, a rule-compliant three-dimensional model results.
  • the object is also achieved by a computer program product with control commands which implement the proposed method or operate the proposed system arrangement.
  • the method has method steps which can be functionally simulated by the structural features of the system arrangement.
  • the system arrangement has structural features which provide functionalities that can be mapped by the method steps. So the method is used to operate the system arrangement and the system arrangement is set up to carry out the proposed method.
  • Figure 1 a schematic flow diagram of the proposed
  • a first system which converts a three-dimensional model into a two-dimensional model or two-dimensional construction plans.
  • a second system can be proposed which converts the two-dimensional construction plans into a three-dimensional model. This can be done, for example, by means of a neural network or artificial intelligence. These two systems can be part of the proposed system arrangement.
  • a 3D building model is automatically generated from 2D plans.
  • One system 1, supported by AI, has a digital twin of the building.
  • the building model is created as a graph database, for example, all components and their logical relationship to one another are represented by functions that are inherently logical. Thus, in accordance with one aspect of the present invention, only parameters are changed.
  • a 3D BIM building model can be created from these graph databases, e.g. in RDF format.
  • schedules which are necessary for the entire process for the physical creation of the component, are also stored with regard to planning, scheduling, administrative, logistical and legal aspects.
  • system 1 generates floor plans, sections and views from the 3D model, e.g. as 2D DXF or 2D DWG files and schedules. It sends these building plans and schedules to system 2, supported by AI.
  • System 2 now has the task of generating the appropriate 3D BIM model as a DXF or DWG file and schedules from the transmitted floor plans, sections and views, taking into account the correct administrative, logistical and legal aspects.
  • System 2 transmits the result of this generation back to system 1.
  • System 1 checks the result proposed by system 2 for correctness. If the result is incorrect, system 1 transmits this to system 2 with the correct solution. System 2 thus learns from its mistakes.
  • System 1 now generates, according to one aspect of the present invention, a new variant of the building in which there is e.g. Moving a component like a wall, changing the material of a component or the deadline specifications for the construction of the building, without violating the overall logic of the building. Or it changes the location and orientation of a building and based on the resulting logic (e.g. solar radiation, climate zones and country-specific regulations) the components are adapted accordingly.
  • System 1 generates floor plans, sections, views and schedules again from this and transmits these to System 2, so the process begins again.
  • System 2 learns to create 3D models and correct schedules from corresponding 2D files.
  • the new variant of the building generated by system 1 can be done automatically using appropriate algorithms as well as by manual input.
  • the aim is also to create 3D building models of existing buildings, of which only 2D files exist, and also to check their accuracy with regard to data consistency, see also 2D-Checker. This would significantly reduce the cost of creating 3D BIM models or RDF files for existing buildings. If these buildings have not yet been created, but only a 2D plan exists, this could also make it possible, with much less effort, to e.g. 3D BIM building model or RDF files can be created with corresponding schedules.
  • FIG. 1 shows a method for the automated creation of an error-free, deterministic and three-dimensional building plan, with a provision 100 of a plurality of two-dimensional building plans of a building, an iterative and adaptive provision 101 of consistency rules which describe a consistent three-dimensional model of the building, an empirical transfer 102 of the plurality of two-dimensional building plans into the three-dimensional model of the building, an application 103 of the consistency rules to the transferred 102 three-dimensional model, an identification 104 of control deviations in the three-dimensional model and an adaptation 105 of the three-dimensional model in such a way that depending on a empirically determined transformation model, a rule-compliant three-dimensional model results.
  • the person skilled in the art recognizes that the method steps can be carried out iteratively and / or in a different order. In addition, individual process steps can have sub-steps.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Geometry (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Computer Graphics (AREA)
  • Architecture (AREA)
  • Civil Engineering (AREA)
  • Structural Engineering (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Processing Or Creating Images (AREA)

Abstract

La présente invention concerne un système permettant de produire de manière automatisée un plan de construction en trois dimensions, déterministe et sans erreur. La présente invention concerne en outre un procédé correspondant et un produit programme informatique avec des instructions de commande qui mettent en œuvre le procédé ou font fonctionner le système. Selon l'invention, il est possible de créer des plans de construction tridimensionnels à partir de plans de construction bidimensionnels, en tenant compte d'un ensemble complexe de règles.
EP20717174.5A 2019-04-04 2020-04-03 Création automatique d'un plan de construction Pending EP3942451A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102019108812.7A DE102019108812A1 (de) 2019-04-04 2019-04-04 Automatisches Erstellen eines Gebäudeplans
PCT/EP2020/059680 WO2020201563A1 (fr) 2019-04-04 2020-04-03 Création automatique d'un plan de construction

Publications (1)

Publication Number Publication Date
EP3942451A1 true EP3942451A1 (fr) 2022-01-26

Family

ID=70189966

Family Applications (1)

Application Number Title Priority Date Filing Date
EP20717174.5A Pending EP3942451A1 (fr) 2019-04-04 2020-04-03 Création automatique d'un plan de construction

Country Status (3)

Country Link
EP (1) EP3942451A1 (fr)
DE (1) DE102019108812A1 (fr)
WO (1) WO2020201563A1 (fr)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112528368B (zh) * 2020-12-09 2022-11-18 四川蓉信开工程设计有限公司 基于bim的水处理工程构筑物的设计方法
CN112733246B (zh) * 2021-01-22 2023-04-07 上海建工四建集团有限公司 建筑物的自动设计方法、装置、终端、存储介质及处理器
CN117171853B (zh) * 2023-09-15 2024-04-26 中建七局建筑装饰工程有限公司 一种基于bim+c#的下挂板正向设计与施工方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3958214A1 (fr) 2014-02-26 2022-02-23 NavVis GmbH Procédé de génération de vues panoramiques sur un système mobile de cartographie
DE102016105480A1 (de) 2016-03-23 2017-09-28 Jürgen Gehrmann Verfahren und System zum Installieren, Verwalten und/oder Warten von Komponenten in einem Gebäudekomplex
DE102016116572A1 (de) 2016-09-05 2018-03-08 Navvis Gmbh Ausrichtung von Punktwolken zur Modellierung von Innenräumen

Also Published As

Publication number Publication date
DE102019108812A1 (de) 2020-10-08
WO2020201563A1 (fr) 2020-10-08

Similar Documents

Publication Publication Date Title
WO2020201563A1 (fr) Création automatique d'un plan de construction
EP1061422B1 (fr) Système d'ordinateur pour la définition, l'optimisation et la régulation des processus
DE102019131385A1 (de) Sicherheits- und leistungsstabilität von automatisierung durch unsicherheitsgetriebenes lernen und steuern
EP2330469B1 (fr) Procédé et environnement de développement destinés à produire un programme de commande complet exécutable
DE112018003983T5 (de) Türantrieb
EP2671122B1 (fr) Conception automatisée d'une technique d'éclairage d'une installation technique
DE102015111875A1 (de) Prozesssteuerungssystem unter Verwendung typischer bzw. Adapterkomponenten
EP2795414B1 (fr) Planification automatisée d'une technique de gestion d'une installation technique
DE10327614A1 (de) Vorrichtung und Verfahren zur Programmierung und/oder Ausführung von Programmen für industrielle Automatisierungssysteme
WO2020201559A1 (fr) Rapport numérique de l'état de bâtiments
EP1092210B1 (fr) Dispositif et procede pour la realisation d'un modele d'installation virtuel
DE10232659A1 (de) Verfahren und Konfigurator zur Erstellung eines Anlagenkonzepts aus einer Anzahl von Anlagenkomponenten
DE102010021382A1 (de) Verfahren und System zur Erzeugung eines Integrationsmodells
DE102018202626A1 (de) Verfahren zur rechnergestützten Parametrierung eines technischen Systems
DE102019208918A1 (de) Verfahren und Vorrichtung zum Abgleich eines Bauwerksdatenmodells mit der betreffenden Gebäudeausrüstung
EP3909011A1 (fr) Création efficace d'une configuration de bâtiment
EP3944120B1 (fr) Procédé de détermination des paramètres de coupe pour panneaux et procédé de fabrication de panneaux pour un assemblage d'une surface
DE102022131925A1 (de) Selbstlernendes verfahren für einen fertigungs- und/oder montageprozess
WO2021170303A1 (fr) Transmission de bord à nuage à réduction de données basée sur des modèles de prédiction
WO2019211188A1 (fr) Installation d'automatisation et procédé de modélisation
DE102020127060A1 (de) Verfahren zum Modellieren eines Aufzugschienensystems
EP4423659A1 (fr) Procédé mis en oeuvre par ordinateur pour disposer des éléments dans un espace d'un bâtiment
EP3637204A1 (fr) Dispositif d'automatisation de bâtiment et procédé
DE102021116161A1 (de) Methode zur Bearbeitung eines Bauwerksplans
EP3955184A1 (fr) Système de planification automatisée et/ou d'autorisation d'un bâtiment ainsi que procédé de fonctionnement dudit système

Legal Events

Date Code Title Description
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: UNKNOWN

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20211021

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

DAV Request for validation of the european patent (deleted)
DAX Request for extension of the european patent (deleted)
P01 Opt-out of the competence of the unified patent court (upc) registered

Effective date: 20230403