Computational Inverse Methods Used In Accuracy Of The Artillery Firing

As a main battle weapon of the conventional weapon system, artillery plays an important role in the national defense and armed forces system. Improving the precision of the artillery firing is very important, thus it is more and more necessary to make the artillery of high hit precision efficiently with low cost. In order to reach the hit accuracy, it usually repeats optimizing the design parameters to achieve the requirements, but this process causes many problems such as large amount of calculation, long manufacture cycle, high manufacture cost and low efficiency. Thus,if we can adopt the practical computational inverse method to identify the state parameter error of the artillery muzzle, it will avoid repeating optimizing process to save the workload and manufacturing cost. Thus this provides an effective way to obtain the accurate state parameter error.This paper carries out a series of research and application to deal with the problems in the state parameter error identification process of muzzle. The main work of this thesis is as follows:(1) An intensity model for the state parameter error identification of the artillery is established. Firstly, it introduces the common coordinate systems and their coordinate transformation of the outside ballistic trajectory. Then it establishes the6-dimensional trajectory equation, and according to the classification of the intensity and the corresponding expressions, it finally establishes an intensity model for the state parameter error identification of the artillery.(2) Sobol’ global sensitivity analysis method based on the quadratic polynomial response surface model is used to analyze the sensitivities of each state parameter of the artillery muzzle. This method adopts Latin sampling samples to fits the quadratic polynomial response surface by least squares, then uses the quadratic polynomial response surface to replace the original model for the calculation of Sobol’ method.This method obtains the global sensitivity results correctly and efficiently of each the state parameter, and filters some state parameters with weak sensitivity, thereby it reduces the computational calculation and improves the computational efficiency.(3) Uncertain computational inverse method based on the dimension reduction integral is used to identify the state parameter error of the artillery muzzle. The state parameter error identification problem is a black-box problem and the calculation willinvolve in the high-dimensional integral process which is difficult or even unable to deal with. The error can be expressed by the moment. The moment of the original system model can be converted the moment of every single-input state parameter submodel by this method, then the calculation moment can be obtained with applying the reduction dimension decomposition to calculate the moment of each sub-model.The objective function can be constructed by making the error between the moment of measuring output model and the calculation moment. However, the unknown input variables mean and standard deviation can be identified by sequential quadratic programming SQP optimization method. Therefore, it ensures the hit accuracy of the artillery, improves the design efficiency and saves the manufacturing cost.