Precision Investment Casting Process Optimization via Integrated Computational Platform

By /

Precision investment casting remains a critical manufacturing method for producing high-performance components in aerospace and energy industries. Traditional optimization approaches relying on empirical trial-and-error face challenges including prolonged iteration cycles, high labor costs, and inefficiency in multi-parameter coordination. To address these limitations, we developed an integrated computational platform combining finite element simulation, design-of-experiments (DOE) algorithms, and multi-objective optimization to achieve rapid process-defect correlation analysis.

Platform Architecture and Functional Modules

The platform integrates ProCAST-based finite element simulations with Python-driven automation tools, enabling high-throughput process optimization. Its three-layer architecture includes:

  1. Data Layer: Manages casting process parameters, simulation results, and surrogate models.
  2. Business Layer: Executes DOE sampling, parallel job scheduling, and optimization algorithms.
  3. Presentation Layer: Provides interactive visualization of simulation progress and Pareto-optimal solutions.
Precision Investment Casting Process

Key Technological Innovations

The platform incorporates three breakthrough features for precision investment casting optimization:

Table 1: Functional Modules of the Integrated Platform
Module Components Optimization Capability
Process Parameterization Latin Hypercube, Box-Behnken, CCD 25-50 variables
Defect Prediction ProCAST, MAGMASOFT interfaces Shrinkage, hot tear, misrun
Multi-Objective Optimization NSGA-II, MOPSO, NLPQLP 3-5 conflicting objectives

The shrinkage porosity prediction model for turbine components follows a quadratic response surface formulation:

$$ Y = 20.02 – 0.4879x_1 + 2.07x_2 + 0.7023x_3 + 0.1656x_1x_2 – 1.92x_1x_3 + 0.5461x_2x_3 – 4.44x_1^2 + 0.476x_2^2 + 10.2x_3^2 $$

where \( Y \) represents shrinkage volume (cm³), \( x_1 \) pouring temperature (°C), \( x_2 \) ambient temperature (°C), and \( x_3 \) thermal emissivity.

Application Case: Turbine Guide Vane Optimization

A K418B superalloy turbine guide vane with 58 thin-wall blades (minimum thickness 0.35mm) was optimized through the platform:

Table 2: Process Parameter Optimization Range
Parameter Lower Bound Upper Bound
Pouring Temperature (°C) 1,450 1,550
Ambient Temperature (°C) 50 100
Thermal Emissivity 0.5 0.7

Using Latin Hypercube sampling with 24 design points, the platform completed simulations 91.66% faster than manual operations. Multi-island genetic algorithm identified the optimal solution:

  • Pouring temperature: 1,549.5°C
  • Ambient temperature: 50.45°C
  • Thermal emissivity: 0.686
  • Predicted shrinkage: 12.721 cm³

Efficiency Analysis

The platform demonstrates remarkable efficiency improvements for precision investment casting development:

Table 3: Computational Efficiency Comparison
Method 24 Simulations Optimization Total Time
Traditional 24 hr Manual 26 hr
Integrated Platform 2 hr Auto 2.5 hr

Industrial Implementation

Validated through X-ray inspection per HB20040-2011 standards, the optimized precision investment casting process achieved 100% qualification rate for turbine guide vanes in serial production. The platform’s adaptability enables rapid parameter adjustment for different superalloys (IN718, Mar-M247) and component geometries.

Future developments will incorporate machine learning-based defect prediction and real-time process monitoring data fusion, further enhancing the platform’s capability for next-generation precision investment casting applications.

Scroll to Top