AI-ML powered BIM framework for smarter 3D models

An AI-powered BIM framework can automate the modeling process and handle large amount of data, enabling better and faster 3D model creation.

Current methods for creating BIM models require extensive manual work, which are tedious, time-consuming, costly, and prone to errors. An AI-ML powered approach can address these challenges by automating the BIM process, consuming huge amount of data at once, avoiding errors, reducing time and costs, thereby enabling better and faster 3D model creation.

The BIM process consists of four phases: scanning of the facility, 3D modeling, system configuration, and solution adoption (Figure 1).

Figure 1: AI-powered BIM modeling framework

• Scanning: Scan and capture the entire facility (360-degree view) using highly advanced devices such as LiDAR and 360° cameras that can collect accurate and minute data at high speed and with precision. For example, lasers scan millions of points in minutes, and create a ‘point cloud’ data set. The images captured (or point cloud datasets) are then cleansed and stitched together and stored on cloud. You need to decide the format of the processed files (collected across various phases) during the scanning phase itself to enable quick retrieval and usage.

• 3D modeling: The images stored in the cloud environment are processed using AI-ML models to generate robust, highly accurate 3D spatial data. The AI integrated into the image processing platform helps to recognize objects present in the facility, identify and blur people captured, and stitch millions of images together to create a detailed digital 3D model of the building. The 3D model contains attributes such as elevations, measurements, and coordinates, and are stored in an output file. Using this 3D model, one can easily navigate through the entire facility, record accurate measurements, and maneuver the digital asset based on the requirement—for expansion, redesigning, or remodeling of any part of the facility.

• System configuration: Using the output file (data) generated in the modeling process, supply chain systems such as warehouse management system (WMS) and labor management system (LMS) can be configured and deployed in a production environment. The output file needs to be maintained and updated regularly to ensure data is accurate and relevant for controlling and managing day-to-day operations.

• Solution adoption: In this phase, business process validation is executed to ensure that performance and efficiency of the systems are not affected by the changes implemented. The spatial attributes can be tested and validated by running algorithms or through a manual checking process to ensure data accuracy.

The above framework can be customized as per business requirements. For example, a leading automotive aftermarket parts retailer improved the operational efficiency of their distribution center by accurately capturing information about their storage facilities using BIM. The cost of BIM depends on the purpose, size, and complexity of the mapping facility, devices used for capturing spatial data, software used for modeling and configuration, and other operational costs involved. High-level collaboration of different stakeholders is vital to the success of the modeling process. SMEs of various operational areas can enable the modeling team to extract data with high accuracy and in configuring relevant systems. Any discrepancies can be identified by physically inspecting the facility at random and cross-checking with the virtual model of the warehouse. The different functional or operational KPIs should be continually monitored and performance ascertained before scaling the modeling process.

Business benefits are significant

An AI-ML-powered BIM strengthens business operations across the value chain.

The application of BIM ranges from planning and designing to entire facility lifecycle, which includes:

• Space utilization: BIM models increase space visibility by providing accurate and complete dimensions of the facility and its volumetric capacities. This will be of immense help in situations where space is constrained.

• Operational efficiency: Digital layouts can help plan DC operations like receiving, putaway, and picking, in both human and co-botic environment.

• Dynamic inventory slotting: The 3D models can help operations managers stock and prioritize inventory based on customer demand. This helps reduce fulfillment time and optimize labor operations.

• Digital twin: Using digital twin, operations managers can digitally plan, optimize, and make any facility changes or scenarios virtually. For example, any scenario changes related to the facility can be made as per their requirements and the layout of the infrastructure can be redesigned to reduce risk and understand the repercussions in advance. This can increase DC efficiency by streamlining complex operations and ensuring managers have a full overview of their facilities.

• Labor productivity: BIM can help improve labor productivity by optimizing layout and reducing the distance that workers need to travel to pick or store products. For example, a 3D model of the warehouse can be used to optimize the pick path mapping and identify the optimal location for each product, reducing the time required to locate and pick products. This information can be used to optimize the picking process, reducing the time required to complete each task and improving labor productivity. Virtual interactive 3D model of the facility can be used to train employees or to familiarize them with the space after remodeling the facility.

• Sustainability: BIM can help reduce environmental impact by optimizing energy usage (for example, designing energy-efficient machine handling equipment), reducing inventory wastage, and improving the overall sustainability of DC operations.