Application of Big Data Analysis in Intelligent Industrial Design Using Scalable Computational Model

Zhe Zhang, H. ZhangScalable 2025 · DOI

TL;DR

This paper introduces the Integrated Concentric Framework for Intelligent Industrial Design (ICF-IID), a scalable computational model designed to harness big data for smarter manufacturing. By combining distributed computing with real-time predictive analytics, the framework achieves significant improvements in design accuracy (98.32%) and process optimization (96.21%) compared to traditional methods.

What Problem Does This Paper Address?

In the era of Industry 4.0, the industrial sector is drowning in data. Modern manufacturing generates massive, high-dimensional datasets from sensors, logs, and production metrics. The challenge isn't just storing this data, but processing it fast enough to make "live" design decisions.

Traditional design workflows often operate in silos, struggling with data integrity, real-time processing demands, and the sheer computational weight of advanced analytics. The authors aim to bridge the gap between "big data" and "intelligent design" by creating a framework that can scale resources dynamically to meet these heavy computational needs.

Methods

The core of the research is the ICF-IID framework. It isn't just one algorithm but a multi-layered architecture:

  • Scalable Distributed Computing: Built on a distributed architecture (utilizing tools like the Hadoop Distributed File System), the framework allows for parallel processing of massive datasets across clusters.
  • Parallelized K-Means Clustering: For real-time data classification, the model uses an adaptive k-means algorithm that constantly recalculates cluster centroids as new data arrives, allowing the system to "learn" and adapt to shifting production patterns.
  • Digital Twins: The framework incorporates digital representations of physical objects to run simulations and generate "fresh ideas" or optimizations before they are implemented on the factory floor.
  • Concentric Integration: The "concentric" aspect refers to how the framework layers data collection, management, and distributed computing around a core of predictive analytics and machine learning tools (like CNNs and ANNs).

Datasets

The paper evaluates the framework's efficacy through simulation analysis. Interestingly, for the specific evaluation of the efficiency and accuracy ratios, the authors discuss a dataset related to employee attrition (predicting when staff might leave) as a case study for the model's predictive capabilities. While the primary focus is industrial design, this dataset serves to demonstrate how the framework handles complex, multi-variable data to provide actionable insights.

Key Insights and Ideas

  • Dynamic Resource Allocation: One of the "aha" moments is the system's ability to scale processing power based on demand. This ensures that real-time perception doesn't lag when data volume spikes.
  • From Static to Living Models: By integrating Digital Twins with big data, industrial design moves from a static, one-time process to a continuous loop of feedback and refinement.
  • The "Concentric" Advantage: By placing data management and distributed architecture at the center, the framework ensures that the "fancy" ML algorithms at the outer layers are always fed high-quality, well-organized data.
  • Standardization is Key: The authors emphasize that the first step to intelligent design isn't the AI itself, but the creation of an exhaustive, standardized database of "alternatives" (Bill of Materials, demand lists, etc.) from which the AI can choose.

Key Results

The ICF-IID framework was compared against existing methods like standard Deep Learning (DLT), Artificial Neural Networks (ANN), and Fuzzy Clustering (FCA). The quantitative gains were substantial:

  • Accuracy Ratio: 98.32%
  • Product Innovation Ratio: 97.65%
  • Scalability Ratio: 97.41%
  • Optimized Design Process Ratio: 96.21%
  • Efficiency Improvement: A 9.21% increase in the efficiency ratio compared to baseline methods.
  • Throughput: The proposed method increased throughput by 99.21% compared to other existing simulation studies.

Limitations and Future Work

While the results are promising, the authors acknowledge several hurdles:

  • Resource Intensity: The system requires significant computational infrastructure, which might be a barrier for smaller organizations.
  • Data Dependency: The model's performance is heavily dependent on the quality and completeness of the input data; "garbage in, garbage out" remains a risk.
  • Integration Complexity: Merging this framework into existing, legacy industrial workflows can be difficult and resource-heavy.

Future directions include refining data gathering methods, solving the "data island" issue (where data is trapped in disconnected systems), and developing even more specialized algorithm models for specific niche industrial tests.

Summary updated

Abstract

Smart industrial design’s incorporation of big data analytics is reshaping the manufacturing industry by boosting product innovation, optimizing design processes, and increasing overall efficiency. Massive amounts of data may be processed using scalable computing, leading to crucial insights that propel more informed, data-driven design choices. Integrating advanced analytics into current design workflows, dealing with diverse and large-volume data, and guaranteeing data quality and integrity are all obstacles to implementing extensive data analysis in industrial design. Many challenges must be solved, including keeping data secure and meeting the computational needs of real-time processing data. Intelligent industrial design benefits significantly from scalable computing’s extensive data analysis capabilities, which allow systems to analyze huge quantities of data in real time. Its dynamic resource allocation achieves efficient resource utilization and optimum performance, guaranteeing that processing power scales with the demand. This research suggests an Integrated Concentric Framework for Intelligent Industrial Design (ICF-IID) that applies big data analysis using scalable computational resources. The framework analyses big datasets from different parts of the design and production process using powerful visualization tools, machine learning algorithms, and predictive analytics. Adaptive algorithms developed for unique demands in industrial design, strong data management protocols, and a distributed computing architecture for efficient data processing are essential components. The framework is useful for predictive maintenance, product lifecycle management, and design parameter optimization in industrial design. The framework may find design defects, predict equipment failures, and suggest improvements by analyzing historical and real-time data. The efficacy and scalability of the suggested framework are assessed through simulation analysis. These results show that it can efficiently and accurately process industrial data on a wide scale. Based on the results, the framework seems useful for making decisions in complicated design contexts and providing practical insights. The proposed method increases the efficiency ratio of 9.21%. accuracy ratio of 98.32%, product innovation ratio of 97.65%, scalability ratio of 97.41%, and optimized design process ratio of 96.21% compared to existing methods.
BibTeX
@inproceedings{zhang2025application,
  author = {Zhe Zhang and H. Zhang},
  title = {{Application of Big Data Analysis in Intelligent Industrial Design Using Scalable Computational Model}},
  booktitle = {Scalable},
  year = {2025},
  doi = {10.12694/scpe.v26i3.4381},
}
Figure