Residual stresses and dimensional deviation in metal additive manufacturing: prediction and mitigation methods

Abstract

Additive manufacturing (AM) has attracted wide attention over the past years due to its various advantages, such as design freedom, less production cost, and short production cycles. However, residual stresses and dimensional deviation are two of the key challenges in terms of structural integrity and the finish quality of printed components. To predict and minimize the residual stresses and dimensional deviation, time-consuming thermal-mechanical simulations and costly experimental studies are often required. Efficient and generic methods are in urgent need to tackle the great challenges. In this thesis, the following simple and effective methods as well as a novel metal deposition pattern have been developed to predict and minimize the residual stresses and dimensional deviation of AM components.

In order to develop an efficient method to predict residual stresses in AM, a point heat source model is developed to study the relationship between the heat flux and the temperature and residual stresses in a representative manufacturing process. Numerical results show that the residual stresses at any position can be expressed as a function of its relative spatial position to the heating surface and the corresponding peak temperature it has experienced during the thermal cycles. The distribution of residual stresses can be divided into three segments according to the peak temperature. The peak temperature only depends on the heat flux and the distance to the point heat source center. A semi-analytical solution is proposed to predict the peak temperature and residual stresses; given the heat flux is known. The proposed solution is further validated by a full-scale numerical case study, and a good agreement has been achieved.

Deposition patterns can significantly affect the residual stress distribution in AM. However, none of the existing patterns can perform well in all aspects: part strength, printing efficiency, applicability, dimensional accuracy, and especially reducing residual stresses. To address this shortcoming of the current practices, the effect of pattern feature on the product performance is analyzed and a novel pattern integrating various advantageous features, called the S-pattern, is proposed for metal AM processes. The finite element (FE) simulation is used to study the temperature and stress field of a cuboid structure under the S-pattern and five other representative patterns: zig-zag, raster, alternate-line, in-out spiral, and out-in spiral. The results show that the S-pattern achieves the lowest magnitude of both equivalent residual stress and maximum principal residual stress. The warpage of the S-pattern is close to that of counterparts.

In addition, the S-pattern possesses multiple advantages over other existing patterns and therefore could be considered as the optimum one among the six deposition patterns considered.

A rapid evaluation of the consequence of the chosen scanning or deposition strategy is of outstanding importance in AM. In order to realize real-time evaluation of deposition patterns in terms of residual stresses, two methods, one based on measured temperatures and one based directly on the information from AM G-code, are proposed. In the first method, through the thermo-mechanical simulations of the temperature and stress fields of a square part with six typical deposition patterns, it is found that the residual stress distribution is determined by the uniformity of temperature distribution which is correlated with the peak temperatures measured at corners. The equivalent residual stress and the maximum principal residual stress are inversely correlated with the average peak temperature and the minimum peak temperature at the corners, respectively. The correlations between temperature and residual stresses provide an efficient strategy to evaluate patterns for different structures, materials, and AM processes, and can be easily implemented in practice.

One step further, the second proposed pattern evaluation method is based on bead sequence number which can be directly obtained from the G-Code design. In this method, starting from the discretization of the deposition pattern by a series of sequence numbers, we introduce two interconnected concepts for assessing the resulting residual stresses: “equivalent bead sequence number” and “bead sequence number dispersion index” which can be physically interpreted as a representation of the localized heat accumulation and the global heat accumulation gradient, respectively. The temperature fields and residual stresses of a square part with six deposition patterns predicted by thermo-mechanical FE simulations are used to develop and verify the proposed criterion. It is found that the “equivalent bead sequence number” of a given pattern is closely correlated to the distribution of the associated temperature and residual stresses. More interestingly, both the highest equivalent and highest maximum principal residual stresses of a pattern linearly increase with its corresponding value of “bead sequence number dispersion index”. Guided by this relation, two new patterns with lower residual stresses are developed and evaluated. The proposed sequence-driven approach allows not only real-time evaluation but also the optimization of deposition patterns.

Dimensional deviation is another great challenge for AM, particularly for thin channels, which was the last focus of the thesis. In order to improve the dimensional accuracy of AM thin channels, we propose a concept called "melting cell" to describe and quantify the overmelting area. Based on the geometrical relationship between the melting cell and target channel, a method for predicting and optimizing the final geometry of thin channels is outlined. To verify the method, geometries of thin horizontal circular channels of various sizes are studied as examples. The predicted results by the proposed method show a remarkable agreement with the available experimental results. Moreover, a mathematical description of the compensated design of a thin channel is also presented. The proposed method is simple yet effective. It can be easily extended to the manufacturing of thin channels with various shapes, materials, and different powder bed fusion processes.