1 Foreword In the study of the overall performance of the engine, the accuracy of the engine component characteristic diagram has a significant influence on the calculation results of the overall performance. Experience has shown that if the component characteristics are not selected properly, the matching operating point in the engine is often not obtained in some working conditions; accurate component characteristics are also of great significance for engine performance monitoring and fault diagnosis during the whole machine test. When performing performance research, analysis, and monitoring of an engine that has been put into operation, due to the theoretical and random errors in the mathematical model of component characteristics or the measurement device itself, as well as changes in factors such as the installation impact of each component and the working environment, The actual characteristics of the engine components are not completely consistent with the calculated or tested results. Therefore, it is necessary to infer the characteristics of components close to actual work from the limited known performance data of the engine. The coupling optimization calculation of engine performance is used to obtain the correction coefficients of the relevant component characteristics, and the correction coefficients are used to correct the original characteristics.
The coupling optimization calculation method is relatively mature, but when it is used to modify the component characteristics, a simple approximation is generally adopted. In this paper, the coupling optimization calculation of the steady-state performance of the engine is used to obtain the coupling coefficients of the component characteristics at each operating point, and the non-linear * small square fitting based on the similarity of the characteristic diagrams is used to modify the characteristic surface of the engine to obtain Component characteristics closer to actual operating conditions.
Since the sample points can be arbitrarily distributed on the characteristic diagram, this paper uses the form of curved surface to describe the component characteristics. If there are a large number of sample points distributed over the entire characteristic range, these discrete data points can be directly used to estimate the first and second derivatives of the surface and fit a smooth characteristic surface. However, it is usually difficult to get enough Data points, therefore, must be corrected with reference to the changing law of the original characteristic diagram. Based on the similarity of the change rule of the characteristic diagram of the same type of parts, in the correction calculation, the node value of the characteristic surface mesh is used as the independent variable when the surface is fitted; the first and second derivatives and mixed partial derivatives of the nodes are used as the optimization target Part of it is to basically maintain the changing law of the original characteristic surface. Curved surface fitting adopts non-linear *little squares method, so that the characteristic value of each sample point is sufficiently close to the value calculated by coupling optimization in the sense of *small squares (equal to the product of coupling coefficient and reference characteristic value) . In this way, the corrected characteristics can better conform to the actual characteristics, and the change law of the characteristic curved surface can also be better maintained.
Since local bicubic interpolation is used at the sample points, a sample point value only affects the correction of 16 adjacent nodes. In order to increase the calculation speed, the characteristic area is divided into the part with sample points and the part without sample points, and the nonlinearity * small square fitting correction of the characteristic diagram of the parts respectively is in these two domains, and iteratively loops in sequence to perform nonlinearity. *Small squares fitting until the accuracy requirements are met. The above method can make the total error level of each sample point tend to be the smallest in the sense of the smallest square. When verifying the accuracy of the corrected component characteristics, the coupled optimization algorithm for engine performance is also used. It can be seen from the definition of the coupling coefficient that if the modified characteristic calculation is used, the closer the obtained coupling coefficient is to 1, the better the consistency with the real characteristic.
4 The calculation results are explained by taking the efficiency characteristics of the combustion chamber as an example.
In order to facilitate the observation of changes, the coupling coefficient of each sample point before characteristic correction is drawn in ascending order, and the correction coefficient of each point after characteristic correction is drawn in the order of these points. It can be seen from Figure 1 that after characteristic correction, the correction coefficients of each sample point are concentrated around 1.0, especially the original large deviation points have been well improved. In the characteristic correction, it is considered that the overall error level of all samples tends to be the smallest. Therefore, some correction coefficients close to 1.0 will have a tendency to deviate not too much. This aspect is due to the influence of other points, and is also related to the accuracy of the steady-state mathematical model.
5 Conclusion Using the non-linear*little squares fitting algorithm based on the similarity of the characteristics of similar components, the characteristic graphs that are more in line with the actual characteristics of the components can be obtained from the characteristic correction coefficients of the discrete samples. The algorithm is feasible and effective. The revised component characteristic diagram can more accurately reflect its actual working characteristics, and the accuracy of the result is close to the steady-state model of the engine and the actual working conditions (such as the accuracy of the algorithm, the accuracy of the actual working conditions and process simulation, etc.) ), the amount of raw data and its accuracy are closely related. In addition, detailed mathematical descriptions should be made for the characteristics of each component, such as its change law and range, etc., and these characteristics should be fully considered when implementing the nonlinear * small squares fitting algorithm, so that the revised characteristic map can be better Reflect the true law of component characteristics. For automotive parts and parts machining, PTJ Shop offers the highest degree of OEM service with a basis of 10+ years experience serving the automotive industry. Our automotive precision shop and experts deliver confidence. We have perfected the art of producing large component volumes with complete JIT reliability, backed by the quality and long-term reliability our customers expect.