One of the most important issues in the execution of engineering projects in BIM is the quality control of the models. This control is essential for the final models to satisfy the benefits promoted and established with the use of the BIM methodology. It is necessary to design an appropriate strategy to guarantee the quality of the models. It is also important to monitor the execution of this strategy. The results must also be interpreted appropriately and documented in reports. In a BIM model quality control process, one of the most relevant and present elements are the conflicts between geometric components of models from the design disciplines.
In order to execute this BIM Use, it is essential that we use resources to standardize the analyses and identify the hierarchy of flexibility between disciplines. After each analysis performed, it is important to carry out evaluations in relation to the previous analysis. It is also important to present an overall evaluation in relation to the entire process developed. This analysis may consume significant time and not have the necessary assertiveness.
It is also important to produce documents that materialize the adopted strategy and the results presented. In this context, is it possible to use artificial intelligence to enhance data analysis and assist in the preparation of quality reports for models?
In this article, I will demonstrate how artificial intelligence can help us achieve good results in these fields.
This post was written by BIM Corner’s Guest Author, Guilherme Guignone.
Table of contents
Chapter 1 - Brief description of the use of the Clash Matrix
The Clash Matrix is a tool that allows the categorization and prioritization of interferences detected between components of different disciplines in a BIM model. Essentially, it is a table that cross-references the disciplines involved in the project, such as: Architecture, Structures, Hydraulics, Electrical, and Mechanical Systems, and structures the record of conflicts between them. Each cell in the matrix represents a possible point of interference between the components of two different disciplines. The matrix makes it easier to see where the most frequent conflicts are and which disciplines are most involved, helping in decision-making and prioritizing resolutions.
How to Run a Clash Matrix?
Running a Clash Matrix generally follows the steps below:
- BIM Preparation and Modeling: Before starting the analysis, all disciplines must develop their models.
- Importing Models: All models from the different disciplines are federated. This is done in BIM model checking software, such as Autodesk Navisworks or Solibri Model Checker, which allows the combination and analysis of models from multiple disciplines.
- Building the Clash Matrix: The planning for the execution of clashes, in terms of selecting pairs of disciplines and the sequencing of the clash executions, is defined.
- Configuring Clash Criteria: The software is configured to detect clashes based on specific criteria. The clash criteria may include aspects such as:
- Geometric tolerances (defining the level of proximity required for something to be considered a clash);
- Types of components that may or may not overlap;
- Priority level (determining the importance of clashes based on factors such as impact on schedule or cost); among others.
- Clash Detection: The BIM model verification software performs an automatic check and generates a list of clashes.
- Clash Classification and Prioritization: After creating the Matrix, clashes are classified by priority. This can be done based on criteria such as severity level (whether the clash impacts the structure or safety, for example), ease of resolution, and project phase. Critical clashes, which may result in risks to structural integrity or major schedule adjustments, are prioritized for resolution.
- Communication and Resolution of Clashes: The Clash Matrix facilitates communication between teams from different disciplines, allowing them to work together to resolve identified clashes. The disciplines involved can adjust their components in the model to eliminate clashes, ensuring that the final project is viable and efficient.
Importance of the Clash Matrix in the BIM Process
The Clash Matrix is a fundamental tool in the BIM process for several reasons:
- Reduction of Rework: By identifying and resolving clashes in the virtual environment, it is possible to avoid problems during physical construction, which reduces rework and the costs associated with corrections on the construction site.
- Increased Efficiency: The matrix allows the team to quickly identify which disciplines and components are most involved in clashes, allowing the resolution effort to be focused where it is most needed.
- Planning and Coordination: The Matrix facilitates collaborative planning between disciplines, allowing a clear view of the points of contact and possible conflicts between systems and components. This allows for more accurate coordination and more efficient resource allocation.
- Prioritization of Interventions: Classifying clashes into priority levels allows teams to focus their efforts on the most critical problems first, ensuring that safety or structural issues are addressed before minor clashes.
- Improved Project Quality: Resolving clashes before physical construction results in a more integrated and higher quality project, as each discipline is adjusted and coordinated to avoid unnecessary interference.
In short, the Clash Matrix is essential to ensure that the project developed in the BIM environment is prepared for physical execution without significant conflicts between the systems and components of the engineering disciplines.
Chapter 2 - What is Artificial Intelligence?
What is Artificial Intelligence?
Artificial Intelligence (AI) is an area of computer science that seeks to develop systems capable of performing tasks that normally require human intelligence. These tasks can include pattern recognition, decision-making, language translation, natural language processing, among others. AI allows machines to process large volumes of data, identify trends and make decisions based on complex and dynamic information, imitating human learning and problem-solving processes. There are different approaches and techniques to AI, including supervised learning, unsupervised learning, reinforcement learning, and deep neural networks. These techniques allow AI systems to learn from historical data and make predictions or recommendations based on new data. AI is applied in numerous areas, such as healthcare, finance, transportation, engineering, and communications, and is becoming increasingly present in everyday activities.
How does ChatGPT work?
ChatGPT, developed by OpenAI, is a language model based on a deep neural network architecture called Transformers. It is trained to process and generate natural language, interpreting questions and providing coherent and contextually appropriate responses. ChatGPT is a Natural Language Processing (NLP) model, which means it is designed to understand, interpret, and generate text in human language.
In the context of the project analyzed, ChatGPT provided a detailed and understandable view of interferences between disciplines, assisted in the generation of reports and graphs, and provided practical guidance to improve project coordination and quality. The use of AI, such as ChatGPT, allowed for a significant gain in efficiency and accuracy, allowing teams to focus on resolving conflicts and improving collaboration, which is essential for project success.
Chapter 3 - Use of artificial intelligence in practice for interpreting interference analysis data and preparing reports
The following artificial intelligence tool was used: ChatGPT-4.
ChatGPT-4 is the fourth version of the language model developed by OpenAI, evolving significantly compared to previous versions in terms of natural language processing capacity, context understanding, and generation of coherent and detailed responses. This model is designed to interpret and answer natural language questions with an advanced level of depth and accuracy, making it useful in a variety of applications, from virtual assistants and customer support to assisting with complex data analysis and interpretation tasks in technical and engineering projects.
Applications of ChatGPT-4 in Engineering and Clash Analysis ChatGPT-4 is especially useful in complex technical contexts, such as engineering and clash analysis in BIM projects.
Its advanced capabilities allow it to act as an assistant for data analysis, interpretation of results and production of technical reports.
Below are some of the ways in which ChatGPT-4 can be applied in this context:
- Data Interpretation and Clash Analysis: In engineering projects that use BIM models, ChatGPT-4 can assist in the interpretation of clash data between disciplines. It can explain the implications of the data, suggest solutions and provide a contextual analysis of how certain conflicts (or “clashes”) can affect the project, facilitating understanding for professionals from different areas.
- Generation of Reports and Technical Documentation: With its ability to produce detailed and coherent texts, ChatGPT-4 is a powerful tool for generating reports and technical documentation. It can synthesize analysis results, describe methods (such as the clash matrix), and generate recommendations based on engineering best practices, which speeds up the documentation creation process.
- Automation in Creating Visualizations and Scripts: ChatGPT-4 can assist in creating scripts to generate graphs and data visualizations, providing guidance for visualization tools such as Python (with Matplotlib and Pandas) or even VBA for automation in Excel and PowerPoint. This support helps engineers and project managers transform data into intuitive graphs that highlight patterns and facilitate decision-making.
- Assistance in Troubleshooting and Suggesting Best Practices: ChatGPT-4 can act as an assistant in solving specific problems, such as conflicts between engineering disciplines in a BIM model. It can suggest best practices for mitigating interference, recommending, for example, periodic model reviews, best practices for coordination between teams, and the use of clash detection tools to detect conflicts in real time.
- Team Training and Training Support: ChatGPT-4 can be used to create educational and informative content for engineering teams, especially regarding the use of BIM, clash matrix, and artificial intelligence. It can answer questions in a personalized manner and provide detailed explanations, serving as a continuous learning resource for technical professionals.
In summary, ChatGPT-4 represents an important advance for the engineering sector,allowing professionals to have access to a tool that interprets data, provides insights and supports decision-making in an efficient and accurate manner. Its ability to process complex information and generate technical documentation streamlines and improves project management, increasing quality and coordination in highly complex environments, such as BIM.
Use of artificial intelligence in practice, aiming at better interpretation of data from interference analyses and increasing the potential for preparing reports (a project in which I worked as BIM Coordinator was adopted as a reference)
Below is the step-by-step guide for using ChatGPT-4 for the adopted Case Study:
Step 1
All the results of the interference analyses performed for the given project were brought in. Next, ChatGPT was asked to organize the information in table format.
Step 2
I then asked for the percentage variations in relation to each analysis to be presented.
Step 3
Next, I requested that a graph be generated in Excel that would best demonstrate the results. To this end, I requested that a VBA script be generated so that I could insert the graph into Excel.
This made it possible to easily generate graphs.
Step 4
I requested a more suitable graph to demonstrate the positive and negative percentage variations.
A stacked bar chart was suggested.
Step 5
I requested the generation of the graph.
Step 6
I executed the following prompt:
“I need a bar chart showing the percentage variations between pairs of subjects in each analysis. I need the chart to be generated in image format and represented in this chatbot.”
Note: In ChatGPT, a prompt is a message or phrase that is used to start a conversation with the neural language model. It can be a question, a statement, a keyword, or any other message that helps the model understand the context or topic of the conversation that will be generated.
Step 7
I executed the following prompt:
“Ok. Now I need to represent the pairs of subjects that presented the greatest negative percentage variations by means of a graph. I need the graphic representation to be organized so that the subjects with the greatest percentage variation are represented first.”
Step 8
I executed the following prompt:
“Generate a text document for me, in report format, containing the images produced by this chatbot, for the client of this engineering project. I need to demonstrate the analyses performed in relation to the interferences between the constructive components of the disciplines throughout the project. The results should be represented in relation to the isolated disciplines and in relation to pairs of disciplines. The disciplines that presented the greatest positive and negative percentage variations should be represented. The disciplines that need to have special attention in relation to the greatest occurrences of positive variations and which ones achieved the best results in relation to the negative variations should be described. I also need to present suggestions for strategies to improve the process of reducing interferences throughout the development of the project. I need the document to be structured in cover, summary and chapters. The initial chapter is the objective, the second is the methodology adopted, the second is the results, the third is the conclusions and the last is the references. Include all the analyses performed by this chatbot. In the methodology chapter, explain what a clash matrix is and how it is adopted to analyze geometric components of engineering disciplines using the BIM (Building Information Modeling) methodology. Use the article available at the following address as a reference: https://bimcorner.com/rules-for-classifying-interferences-with-the-use-of-clash-matrix/ The report must be written in formal language, using technical jargon, and must be at least 12 pages long.”
It generated a simplistic result:
Step 9
I asked for the following prompt:
“Rewrite chapter [2. Methodology] to provide more details on what a clash matrix is, how to execute it, and why it is important.”
I did this individually for each chapter:
He produced detailed and rigorous content. I copied each chapter into the report generated in .doc format, so it was possible to create a complete document. Below is the final report:
The entire process that I performed manually could be enhanced and, in addition, it was possible to add complementary analyses. Below, the manual process that I performed.
Conclusion
The use of ChatGPT proved to be an important ally in the process of interpreting results and preparing technical documentation. It was possible to speed up the process of developing documents and analyses. These are important assistants for the development of human work. The human will validate or not the product generated by the assistants, review it and complement it. The result of the content produced by ChatGPT is closely linked to the input information executed by humans. The prompt engineering built will be decisive for good responses from ChatGPT and distancing from possible hallucinations of AI tools.
This post was written by BIM Corner’s Guest Author, Guilherme Guignone.
References
Autodesk. Navisworks Manage: Powerful Tools for Detecting Clashes in BIM Models. Disponível em: https://www.autodesk.com/products/navisworks/overview.
Azhar, S. (2011). Building Information Modeling (BIM): Trends, Benefits, Risks, and Challenges for the AEC Industry. Leadership and Management in Engineering, 11(3), 241-252.
BIM Corner. Rules for Classifying Interferences with the Use of Clash Matrix. Disponível em: https://bimcorner.com/rules-for-classifying-interferences-with-the-use-of-clash-matrix/.
Eastman, C., Teicholz, P., Sacks, R., & Liston, K. (2011). BIM Handbook: A Guide to Building Information Modeling for Owners, Managers, Designers, Engineers, and Contractors. 2nd ed. Hoboken, NJ: John Wiley & Sons.
Hardin, B., & McCool, D. (2015). BIM and Construction Management: Proven Tools, Methods, and Workflows. 2nd ed. Hoboken, NJ: John Wiley & Sons.
Volk, R., Stengel, J., & Schultmann, F. (2014). Building Information Modeling (BIM) for Existing Buildings – Literature Review and Future Needs. Automation in Construction, 38, 109-127.
Khemlani, L. (2012). Top Criteria for Selecting a Clash Detection Solution. AECbytes. Disponível em: http://www.aecbytes.com/feature/2012/ClashDetectionSolutions.html.
