Optimisation of the design of a component

1. A method of using a data processor for optimizing a design of a component, in which method principal design variants are generated, the principal design variants having design parameters which are common to principal design variants and having design variables which differ between the principal design variants, the method comprising;

a) inputting a first principal design variant to said data processor and generating, by means of a rule-based geometry means, at least one secondary design variant from said first principal design variant,

b) analyzing the design variants to generate analysis results,

c) subjecting the analysis results to an optimization process, and

d) iterating steps a) to c) a desired number of times in said data processor, and in each such iteration, changing a design variable to generate a further principal design variant which replaces said first principal design variant, wherein said optimization process comprises a first stage employing a genetic algorithm, and a second stage employing a gradient search algorithm.

2. A method as claimed in claim 1 , wherein said at least one secondary design variant is derived from the first principal design variant by applying a mathematical operation to at least one of the design parameters or at least one of the design variables of the principal design variants.

3. A method as claimed in claim 1 , wherein said analyzing the design variants comprises a finite element analysis, involving applying input conditions to determine the behavior of the design variants.

4. A method as claimed in claim 1 , wherein the at least one secondary design variant represents a component made to the design at a tolerance limit.

5. A method as claimed in claim 1 , wherein the at least one secondary design variant represents a component failure.

6. A method as claimed in claim 1 , wherein the at least one secondary design variant represents damage to a component made to the design.

7. A method as claimed in claim 1 , wherein the at least one secondary design variant represents a simplified geometry based on feature or dimension reduction.

8. A method as claimed in claim 1 , in which the component is a component of a gas turbine engine.

9. A method as claimed in claim 8 , in which the component is a blade for a gas turbine engine, said blade having a fir tree root.

10. A method of using a data processor for optimizing a design of a component, comprising;

a) inputting in a data processor, to a rule-based geometry means, a design file, which includes a plurality of design variables, and the rule-based geometry means generating from said design file a plurality of geometry design variants,

b) analyzing models based upon said design variants to generate analysis results,

c) subjecting the analysis results to an optimization process, and

d) iterating steps a) to c) a desired number of times in said data processor, and in each such iteration, changing one of said design variable within said design file, wherein said optimization process comprises a first stage employing a genetic algorithm, and a second stage employing a gradient search algorithm.

11. Apparatus for optimizing a design of a component, comprising;

a) database means for storing design parameters which are common to design variants, and for storing design variables which differ between the design variants, and the database means for providing an output comprising a design file which includes a plurality of design variables,

b) a rule-based geometry means for receiving said design file and for generating from said design file a plurality of geometry design variants,

c) analysis means for analyzing models based upon said design variants to generate analysis results,

d) optimization means for subjecting the analysis results to an optimization process, and

e) control loop means for coupling said optimization means to said rule-based geometry means, for modifying a design variable in said design file, and for enabling iteration a desired number of times in said apparatus of items b), c) and d), wherein said optimization means includes genetic algorithm means, and gradient search means.

12. Apparatus according to claim 11 wherein said rule-based geometry means comprise an ICAD system for generating a geometry file containing said geometry design variants.

13. Apparatus according to claim 11 wherein said analysis means comprises finite element analysis means.

14. A method of using a data processor for optimizing a design of a component, in which method principal design variants are generated, the principal design variants having design parameters which are common to principal design variants and having design variables which differ between the principal design variants, the method comprising;

a) inputting a first principal design variant to said data processor and generating, by means of a rule-based geometry means, at least one secondary design variant from said first principal design variant,

b) analyzing the design variants to generate analysis results,

c) subjecting the analysis results to an optimization process, and

d) iterating steps a) to c) a desired number of times in said data processor, and in each such iteration, changing a design variable to generate a further principal design variant which replaces said first principal design variant, wherein the component is a blade for a gas turbine engine, said blade having a fir tree root and said blade attached by said fir tree root to a disk having a last continuous radius, wherein the optimization process minimizes the area outside the last continuous radius of said disk to which said blade is attached by its fir tree root.

15. A method as claimed in claim 14 wherein the optimization process comprises a first stage employing a genetic algorithm and a second stage employing a gradient search algorithm.

16. A method of using a data processor for optimizing a design of a component, in which method principal design variants are generated, the principal design variants having design parameters which are common to principal design variants and having design variables which differ between the principal design variants, the method comprising;

a) inputting a first principal design variant to said data processor and generating, by means of a rule-based geometry means, at least one secondary design variant from said first principal design variant,

b) analyzing the design variants to generate analysis results,

c) subjecting the analysis results to an optimization process, and

d) iterating steps a) to c) a desired number of times in said data processor, and in each such iteration, changing a design variable to generate a further principal design variant which replaces said first principal design variant, wherein the component is a blade for a gas turbine engine, said blade having a fir tree root, an optimum tooth profile and a maximum notch stress, wherein the optimization process determines the optimum tooth profile to minimize the maximum notch stress.

17. A method as claimed in claim 16 wherein the optimization process comprises a first stage employing a genetic algorithm and a second stage employing a gradient search algorithm.

18. A method of using a data processor for optimizing a design of a component, in which method principal design variants are generated, the principal design variants having design parameters which are common to principal design variants and having design variables which differ between the principal design variants, the method comprising;

a) inputting a first principal design variant to said data processor and generating, by means of a rule-based geometry means, at least one secondary design variant from said first principal design variant,

b) analyzing the design variants to generate analysis results,

c) subjecting the analysis results to an optimization process, and

d) iterating steps a) to c) a desired number of times in said data processor, and in each such iteration, changing a design variable to generate a further principal design variant which replaces said first principal design variant, in which the component is a turbine blade having a fir tree root and said blade attached by said fir tree root to a turbine disk having a last continuous radius, wherein the optimization process minimizes the area outside the last continuous radius of the turbine disk.

19. A method as claimed in claim 18 wherein the optimization process comprises a first stage employing a genetic algorithm and a second stage employing a gradient search algorithm.

20. A method of using a data processor for optimizing a design of a component, in which method principal design variants are generated, the principal design variants having design parameters which are common to principal design variants and having design variables which differ between the principal design variants, the method comprising;

a) inputting a first principal design variant to said data processor and generating, by means of a rule-based geometry means, at least one secondary design variant from said first principal design variant,

b) analyzing the design variants to generate analysis results,

c) subjecting the analysis results to an optimization process, and

d) iterating steps a) to c) a desired number of times in said data processor, and in each such iteration, changing a design variable to generate a further principal design variant which replaces said first principal design variant, in which the component is a turbine blade having a fir tree root, an optimum tooth profile and a maximum notch stress, wherein the optimization process determines the optimum tooth profile to minimize the maximum notch stress.

21. A method as claimed in claim 20 the optimization process comprises a first stage employing a genetic algorithm and a second stage employing a gradient search algorithm.