Turbomachine Blade Distribution Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for distributing blades around a turbomachine disk face challenges in achieving balance with high computational loads, leading to vibrations and acoustic noise due to the large number of possible blade distributions that need to be calculated, which can converge slowly to an optimal solution.
Innovation Solution
A method using different similarity criteria to generate neighboring blade distributions, allowing for quicker exploration of the distribution space through a Tabou-type algorithm, which reduces the number of distributions tested and arbitrates between strategies to converge more rapidly to an optimal solution, while thresholding unbalances ensures focus on meeting specific criteria.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a Tabu-type algorithm switches between distributions by swapping two blades to explore all possible distributions, then the exploration of distribution space is achieved, but the convergence to an optimal solution is relatively slow when the cost function depends on several blade balancing parameters
Solution Approach 1:
The invention segments the exploration of distribution space by introducing multiple similarity criteria (first, second, and third criteria) that define different neighborhoods. Each criterion represents a different strategy for generating neighboring distributions, allowing the algorithm to explore the solution space through multiple pathways simultaneously rather than a single slow path
Solution Approach 2:
The invention dynamically adapts the exploration strategy by selecting different similarity criteria based on the current state of the search. The algorithm can switch between different neighborhood definitions (different criteria) to accelerate convergence, making the exploration process flexible and responsive to the optimization progress rather than following a fixed slow path
2Productivity
If multiple similarity criteria are used to generate neighboring distributions with different neighborhoods, then different exploration strategies are defined, but the computational load increases
Solution Approach 1:
The invention applies partial action by not requiring all possible distributions to be examined exhaustively. Instead, it generates a limited set of neighboring distributions according to each similarity criterion and selects the best one. This partial exploration through multiple criteria provides faster convergence without the prohibitive computational cost of complete enumeration
Solution Approach 2:
The invention changes the parameters of the search process by introducing multiple similarity criteria with different neighborhood definitions. This allows the algorithm to adjust its exploration behavior dynamically, switching between different search strategies (parameter settings) to balance computational effort and convergence speed based on the optimization needs
Data Source
Figure 1
Figure 2a
Figure 2b~2c
AI summary
The invention proposes a method for simulating the distribution of blades on a turbomachine disc, the method comprising the steps of:⋅providing a plurality of blade configurations, each blade configuration being associated with a blade and comprising at least one measurement of a balancing parameter measured on the associated blade, ⋅ search for and selection of a bladed-disc distribution combining the configurations of blades supplied with positions on the disc, the bladed disc distribution encouraging the attainment of at least one criterion defined according to a predetermined cost function dependent on the balancing parameter measurements, the search and selection being performed by successive iterations, the method being characterized in that a current iteration in the search and selection involves the steps of: ○ generating a plurality of distributions from a reference distribution, said plurality comprising at least one first nearby distribution generated according to a strategy encouraging the attainment of a first predetermined set of criteria of resemblance to the reference distribution and at least one second distribution generated according to another strategy encouraging the attainment of a second predetermined set of criteria of resemblance to the reference distribution, ○ selecting, from the nearby distributions, a nearby distribution that encourages optimization of the cost function as the new reference distribution.