Traditional Simulation Methods Research Articles

A New Method for the Simulation of High Reynolds Number Effects on Cooling Tower Models in Wind Tunnels

To protect large cooling towers against strong winds, wind tunnel model tests have traditionally been undertaken. As typical circular cylindrical structures, simulations of high Reynolds number (Re) effects are indispensable on scale models of large cooling towers to validate the truthfulness of test results. However, two issues currently exist that hinder the effectiveness of high Re effects simulation: (1) the mean wind pressure distribution at high Re is usually utilized as the sole simulation target, and the simulation results are not typically checked for other dynamic wind effects that are equally important; and (2) it is not clear whether the wind effects obtained and generalized by other researchers for small-sized cooling towers can be used as simulation targets for large cooling towers. In light of these issues, a new method for the simulation of high Re effects that takes into account real simulated dynamic wind effects is proposed in this study and compared with the traditional simulation method, which only considers the simulation of the mean wind pressure distribution. By comparing the wind effects measured on a 167 m-high large cooling tower with those obtained on small-sized cooling towers, the irrationality of treating the wind effects measured on small-sized cooling towers as high Re effects simulation targets for large cooling towers is demonstrated. With these findings, future model tests are expected to reproduce more realistic wind effects on large cooling towers.

Refined Simulation Method for Computer-Aided Process Planning Based on Digital Twin Technology.

Simulation technology is widely used in computer-aided process planning (CAPP). The part machining process is simulated in the virtual world, which can predict manufacturing errors and optimize the process plan. Simulation accuracy is the guarantee of process decision-making and optimization. This article focuses on the use of digital twin technology to build a high-fidelity process model, taking the advantage of the integration of multiple systems, in order to achieve the dynamic association of real-time manufacturing data and process models. Making use of the CAPP/MES systems, the surface inspection data of the part is fed back to the CAPP system and associated with the digital twin process model. The wavelet transform method is used to reduce the noise of the high-frequency signal of the detection data, and the signal-to-noise ratio (SNR) is calculated to verify the noise reduction effect. The surface topography, after noise reduction, was reconstructed in Matlab. On this basis, the Poisson reconstruction algorithm is used to reconstruct the high-fidelity process model for the refined simulation of the subsequent processes. Finally, by comparing the two sets of simulation experiments with the real machining results, we found that the simulation results, based on the digital twin model, are more accurate than the traditional simulation method by 58%.

Generalized finite difference method (GFDM) based analysis for subsurface flow problems in anisotropic formation

This paper applies the meshless generalized finite difference method (GFDM) to analyze the subsurface flow problem in anisotropic formations for the first time, and develops the treatment methods of anisotropic permeability tensor in continuous function and discrete distribution expressions respectively. In particular, in case of the common discrete distribution expression of anisotropic permeability tensor in reservoir simulation, this paper suggests combining general difference operators and harmonic average scheme to simply and efficiently obtain the meshless discrete scheme of the anisotropic flow equation. Three numerical examples of steady and transient flow in the rectangular computational domain, transient flow with complex boundary, and steady flow with strongly discontinuous anisotropic permeability are used to verify the high accuracy and good convergence of GFDM to handle anisotropic flow problems. Compared with the traditional mesh-based reservoir numerical simulation method that replies on laborious gridding to discretize the reservoir domain and complicated multi-point flux approximation (MPFA) to handle discrete distribution anisotropic permeability tensor, the meshless GFDM can be more practically applied to the anisotropic flow problems in the reservoir domain with complex geometry and complex boundary conditions. Besides, it is found that the radius of the influence domain has little impact on the accuracy of steady flow. For the transient flow, it generally holds that the larger the radius of the influence domain, the lower the calculation accuracy in case of Cartesian collocation. In sum, this work provides an efficient and simple meshless solver to handle anisotropic flow problems in porous media under GFDM framework, which reveals the great application potential of GFDM in reservoir numerical simulation.

Deep Learning Accelerated Hydraulic-Mechanical Simulation with Prior Smoothness Constraints for Heterogeneous Petroleum Reservoirs

Summary Coupled hydraulic-mechanical (HM) reservoir simulation requires tremendous efforts of development and is usually time-consuming. Consequently, the accurate yet fast simulation of poroelastic reservoirs is a challenge to traditional reservoir simulation methods. In this work, we aim to resolve this issue by replacing the geomechanical simulation module with a proxy stress predictor. We have developed a deep learning (DL)-based stress inference module to accelerate geomechanical simulation. The DL is based on convolutional neural network. We have constructed a 2D U-Net network, which takes the pressure, rock properties, and initial and boundary conditions as input and predicts the induced stress fields. We use the upper bound of the gradient of the stress field, which is from the a priori analysis of the mechanical governing equation, as a Lipschitz smoothness constraint. The model is trained with 80,000 pressure–stress pairs and demonstrates accuracy that is greater than 99%. We have augmented the trained network to a hydraulic reservoir simulator to conduct coupled HM simulation. Our results show that the proxy network effectively reduces the computational time of the mechanical module by more than 90% while still maintaining the accuracy of the physical simulator. The smoothness-constrained U-Net demonstrates significantly higher convergence rate and generalization capability. The novelty of this work is that it is arguably the first effort to combine a priori analysis of governing partial differential equations (PDE) with convolutional neural networks.

Intelligent Design Model of Urban Landscape Space Based on Optimized BP Neural Network

Urban landscape space planning is an important application field of landscape ecology. With the continuous development of the field of architectural optimization, more and more optimization methods have sprung up, including various intelligent optimization algorithms. Such intelligent optimization algorithms usually rely on traditional building performance simulation methods to obtain building performance indicators for optimization in the optimization process. However, intelligent optimization algorithms generally require large-scale calculations. At the same time, the time required for building performance simulation is often limited by the complexity of the building model and the configuration of the computer, which leads to too long performance optimization time for designers in the project. With efficient and accurate feedback, building performance optimization methods based on intelligent optimization algorithms are mainly used in scientific research and are difficult to invest in actual projects. Because the traditional BP neural network has its own limitations and its insufficient sample size and weak generalization ability in complex prediction problems, this paper uses the learning algorithm of optimizing the BP neural network to propose an urban landscape space intelligent design model. This article introduces the artificial neural network, a new application technology with the assistance of a geographic information system, and establishes a BP neural network model for urban landscape ecological planning. Seven elements of distance and number of residential points are used as input variables, patch density, fractal dimension, the Shannon diversity index, and aggregation degree are selected as output changes, and 20 samples are carefully collected to train the network. The results show that the network convergence effect is ideal and the generalization ability is strong, which provides a new simulation analysis method for landscape ecological planning.

Discover your potential: The influence of kinematics on a muscle's ability to contribute to the sit-to-stand transfer.

Existing methods for estimating how individual muscles contribute to a movement require extensive time and experimental resources. In this study we developed an efficient method for determining how changes to lower extremity joint kinematics affect the potential of individual muscles to contribute to whole-body center-of-mass vertical (support) and anteroposterior (progression) accelerations. A 4-link 2-dimensional model was used to assess the effect of kinematic changes on muscle function. Joint kinematics were systematically varied throughout ranges observed during the momentum transfer phase of the sit-to-stand transfer. Each muscle’s potential to contribute to support and progression was computed and compared to simulated potentials estimated by traditional dynamic simulation methods for young adults and individuals with knee osteoarthritis. The new method required 4-10s to compute muscle potentials per kinematic state and computed potentials were consistent with simulated potentials. The new method identified differences in muscle potentials between groups due to kinematic differences, particularly decreased anterior pelvic tilt in young adults, and revealed kinematic and muscle strengthening modifications to increase support. The methods presented provide an efficient, systematic approach to evaluate how joint kinematic adjustments alter a muscle’s ability to contribute to movement and can identify potential sources of pathologic movement and rehabilitation strategies.

Analysis of Mixing in Corotating Twin Screw Extruders through Numerical Simulation

Abstract Mixing is a major ingredient in many industrial processes to obtain desired and uniform material properties. Nowadays, many materials, pigments, additives, gas or reactants, are mixed to create new products combining the characteristics of different raw materials to obtain specific product properties. The quality of the mixing, i.e. the uniformity of the mixture, is a key issue that will determine the morphology and the properties of the resulting compound [1, 2]. An insight in the physics of mixing is therefore necessary in order to achieve a good quality of mixing or maintain it when scaling equipment, for example. Such information can now be obtained through the numerical simulation of the transient flow in extruder components. To improve greatly the ease of obtaining such information, a new technique introduced in a finite elements software is presented. This technique simplifies the meshing, reduces the meshes needed and eliminates complex remeshing algorithms to simulate flow in screws, pumps and mixing devices. This technique is validated versus traditional simulation methods (i.e. conforming meshes). 3-D transient numerical simulations of two twin screw extruder configurations are then presented. A further quantitative comparison of their mixing behavior is developed as statistical information (of the RTD, deformations, dispersion, etc.) can be obtained to compare both configurations in a synthetic and quantitative way.

3D reconstruction of porous media by combining scaling transformation and multi-scale discrimination using generative adversarial networks

The modeling and characterization of porous media is quite significant to explore and develop oil and natural gas resources . Traditional numerical simulation methods obtain reconstruction results through the statistical probability in training images (TIs), while the reconstruction process is lengthy and the probability information cannot be reused. The development of deep learning has provided reliable support for the reconstruction of porous media due to its strong ability of learning and extracting characteristics from TIs and the reuse of network parameters. As a common deep learning method, generative adversarial networks (GANs) can obtain images highly similar to the structural characteristics of the TI through adversarial training between the generator and the discriminator . Based on GAN, a three-dimensional multi-scale pattern generative adversarial network (3D-MSPGAN) that can achieve 3D super-resolution reconstruction of porous media is proposed in this paper. The use of scaling transformation allows 3D-MSPGAN to reconstruct super-resolution images of porous media that are much larger than the TI, while the multi-scale discrimination in the discriminator provides a guarantee for the generation of high-quality super-resolution images, by which the discriminator can retain the global structure and local characteristics of the TI simultaneously during the reconstruction process. In addition, due to the multi-scale patch information used in the discriminator, only a single TI is needed to complete the reconstruction of porous media. Experimental comparison with some typical methods proves that 3D-MSPGAN can achieve the reconstruction of 3D porous media with faster speed and higher quality. • Our method can achieve super-resolution reconstruction of porous media. • Our method only needs a single training image for the reconstruction. • Our method can complete large-quantity reconstructions quickly.

A Novel Trilinear Flow Model for Capturing Dynamic Behavior of Water Flooding in Core Plug with Different Fracture Inclinations in Carbonate Reservoirs

Abstract The purpose of this study is to introduce a new three-linear flow model for capturing the dynamic behavior of water flooding with different fracture occurrences in carbonate reservoirs. Low-angle and high-angle fractures with different occurrences are usually developed in carbonate reservoirs. It is difficult to simulate the water injection development process and the law of water flooding is unclear, due to the large variation of the fracture dip. Based on the characteristics of water flooding displacement streamlines in fractured cores with different occurrences, the matrix is discretized into a number of one-dimensional linear subregions, and the channeling effect between each subregion is considered in this paper. The fractures are divided into the same number of fracture cells along with the matrix subregion, and the conduction effect between the fracture cells is considered. The fractured core injection-production system is divided into three areas of linear flow: The injected fluid flows horizontally and linearly from the matrix area at the inlet end of the core to the fracture and then linearly diverts from the fracture area. Finally, the matrix area at the outlet end of the core also presents a horizontal linear flow pattern. Thus, a trilinear flow model for water flooding oil in fractured cores with different occurrences is established. The modified BL equation is used to construct the matrix water-flooding analytical solution, and the fracture system establishes a finite-volume numerical solution, forming a high-efficiency semianalytical solution method for water-flooding BL-CVF. Compared with traditional numerical simulation methods, the accuracy is over 86%, the model is easy to construct, and the calculation efficiency is high. In addition, it can flexibly portray cracks at any dip angle, calculate various indicators of water flooding, and simulate the pressure field and saturation field, with great application effect. The research results show that the greater the fracture dip angle, the higher the oil displacement efficiency. When the fracture dip angle is above 45°, the fracture occurrence has almost no effect on the oil displacement efficiency. The water breakthrough time of through fractures is earlier than that of nonthrough fractures, and the oil displacement efficiency and injection pressure are more significantly affected by the fracture permeability. With the increase of fracture permeability, the oil displacement efficiency and the injection pressure of perforated fractured cores dropped drastically. The findings of this study can help for better understanding of the water drive law and optimizing its parameters in cores with different fracture occurrences. The three-linear flow model has strong adaptability and can accurately solve low-permeability reservoirs and high-angle fractures, but there are some errors for high-permeability reservoirs with long fractures.

Analyzing vulnerability of optical fiber network considering recoverability

With the development of optical transmission technology, optical fiber networks have become critical infrastructures in supporting information transmission on the Internet. However, the fiber cable is very vulnerable to large-scale damage such as earthquakes or pulse bombs. What is more serious is that it will take a long time to locate and repair the damages on fiber links. The long-term repair process will cause continuous network performance degradation and severe economic loss. The fact is that these dangerous areas may be ignored by traditional vulnerability analysis models. To solve this problem, this paper proposes a method to analyze the vulnerability of fiber networks based on network recoverability. We first improve the traditional fiber network simulation methods and damage simulation methods to provide a compatible foundation for the network recoverability simulation. Then, we present the network vulnerability analysis model: the Damage Measurement and Location Model (DMLM). The model employs the heuristic traversal algorithm based on random points to locate the candidate attack positions. We design three vulnerability metrics: two metrics are related to network recoverability and one metric is used for comparison. We also build their corresponding theoretical frameworks to determine the appropriate model parameters to satisfy the specified estimation error requirements. Numerical results prove the proposed model’s effectiveness and excellent sensitivity for essential parameters. The visual results of the vulnerable zones prove the necessity of considering network recoverability in vulnerability analysis of optical fiber networks.

A Physics-Informed Deep Learning Paradigm for Transient Power Angle Stability Assessment

Transient stability assessment (TSA) plays an essential role in the safe and stable operation of power system. The traditional time-domain simulation method and direct method have great limitation in modeling and computing, which can not meet the need of fast and accurate assessment of power system stability. Therefore, combined physics model with data-driven algorithm, this paper proposes a physics-informed deep learning paradigm to solve this problem. In the data-driven algorithm, recurrent neural networks are trained to predict the rotor speed and power angle with the guiding of the power system’s physical information. Finally, the simulation on IEEE14 system is carried out by using PSS/E software, which proved that the proposed method can predict the power angle trajectory quickly and accurately.

Research on numerical simulation of wear temperature field of vehicle brake disc based on SIFI feature

In order to overcome the problems of low simulation accuracy and long time in traditional numerical simulation method of vehicle brake disc wear temperature field, a numerical simulation method of vehicle brake disc wear temperature field based on SIFI characteristics is proposed. The temperature field simulation model of vehicle brake disc wear is built, and the temperature field estimation results are obtained. On this basis, after the Gaussian scale space is generated by using SIFI features, the key point descriptor is generated by determining the key extreme points and the direction of the key points, so as to extract the characteristics of the vehicle brake disc wear temperature field. According to the feature extraction results, the numerical simulation of the vehicle brake disc wear temperature field is carried out. The experimental results show that the simulation accuracy of this method is more than 92%, the average simulation time is 33.9 ms, and the comprehensive performance is better.

An integrated simulation method for PVSS parametric design using multi-objective optimization

An adequate strategy for achieving energy efficiency when designing a photovoltaic shading system (PVSS) shall find an equilibrium between sunlight heat gain and daylight transmittances through effective analysis tools in a building's early design phases. However, traditional simulation methods are either time-consuming or lacking architectonical thinking. This paper proposes a new method for architects to integrate thermal and daylighting performance by using parametric script modelling and optimize their balance with multi-objective optimization (MOO) algorithm in PVSS design. A case study was conducted to demonstrate the workflow of proposed integrated simulation method in PVSS design, and further compared the results with that of three single-objective optimizations under the same design requirement. The findings show that the integrated framework is a feasible method for PVSS design and can be extended into the design of other advance shading system or building integrated photovoltaic.

Study on Production Optimization Method of Fractured Reservoir Based on Connectivity Model

Due to the extensive development of fractures and serious heterogeneity in fractured reservoirs, it is difficult for the traditional numerical simulation method to invert its geology, which greatly limits the efficiency and accuracy of simulation and cannot realize the real-time optimization of production scheme. The connectivity model can only consider the two characteristic parameters of conductivity and connectivity volume, which does not involve complex and rigorous geological modeling. It can quickly and accurately reflect the state of the real reservoir, greatly reducing the simulation time, and is suitable for real-time production performance prediction of the reservoir. Due to the large difference in conductivity of fractured reservoirs, the difficulty of fitting increases. In this paper, the connectivity model is first applied to fractured reservoirs to realize the production dynamic simulation of fractured reservoirs. The optimization principle is used to optimize the injection-production scheme with the economic net present value as the objective function. In order to verify the method, the connectivity model is applied to Mu 30 of Changqing Oilfield in this paper. The results show that this method can effectively reflect the real production situation of the oilfield and the connectivity of the reservoir, and the simulation time is relatively fast. After optimization, the cumulative oil production of the reservoir increases by 8.1%, the cumulative water injection decreases by 2.3%, and the rising rate of water cut decreases by 58.8%, indicating that the connectivity model can realize the real-time production optimization of the reservoir.

A Novel Data-Driven Model for Dynamic Prediction and Optimization of Profile Control in Multilayer Reservoirs

Establishing reservoir numerical simulation and profile control optimization methods considering the mechanism of profile control has always been a difficult research problem at home and abroad. In this paper, firstly, a physics-based data-driven model was established on daily production data of injection and production wells following the principle of material balance. Key parameters including transmissibility, control pore volume, water injection allocation factors, and injection efficiency are derived directly from history matched model, and the dominated flow channels could be quantitatively identified. Then, combined with the evaluation results of the plugging ability of the plugging agent, imaginary well nodes are added to the existing interwell relationship to characterize the heterogeneity of interwell-specific parameters. This process performs flow processing along the interwell control units, forming a new and rapid method for simulation and prediction. Lastly, based on the calculated interwell transmissibility, water injection efficiency, and allocation factors, injection wells with low water injection efficiency can be preferentially selected as profile control wells. In addition, taking the production rates, injection rates, and the amount of plugging agent as optimization variables, we established an optimal control mathematical model and realized the parameter optimization method of the profile control. We demonstrated the results of one conceptual model and two indoor experiments to verify the feasibility of the proposed method and completed two actual field applications. Model validation and actual field application show that the proposed method successfully eliminates the complicated geological modeling procedure and the tedious calculation process associated with the profile control treatment in traditional numerical simulation methods. The calculation speed improves tens or hundreds of times, and water channeling paths are accurately identified. Most importantly, this method realizes the overall decision-making of profile control well selection, dynamic production prediction, and parameter optimization of profile control measures quickly and accurately by mainly using the daily production data of wells. The findings of this study can help for better understanding of the optimization design and application of on-site profile control schemes in large-scale oilfields.

The development of fuel management code TRACS for HFETR based on Unstructured-mesh variational nodal method

High Flux Engineering Test Reactor (HFETR) focuses on the irradiation test of reactor fuel elements, thus the accurate determination of power or flux distribution becomes important. Since HFETR contains complex geometry assembly and different types of irradiation channels, the traditional simulation methods such as Monte Carlo method and deterministic method based on rectangle and hexagonal mesh are not applicable in taking into account calculation accuracy and efficiency. In this study, based on “two-step” calculation scheme, a fuel management code named TRACS was developed. Homogenization of few-group cross-sections was performed by KYLIN-II lattice code, and reactor core calculation was performed by unstructured-mesh variational nodal method. Criticality position of control rod searching adopts difference method, and micro-depletion calculation is performed by solving microscopic burnup equation using an open source matrix exponential solver EXPOKIT of Chebyshev rational approximation method. Finally, two benchmarks of HFETR sample model with different irradiation channels were used to verify TRACS code, and the results by Monte Carlo code OpenMC serve as reference. The comparison results show that the keff error is less than 300 pcm for both two cases, and the relative power error of nearly all the fuel assembly and regions in irradiation assembly at both BOL and EOL stay below 5%, and the critical positions of control rods between TRACS and OpenMC during reactor operation are nearly the same. Overall, TRACS agrees well with OpenMC, and it can be used for fuel management calculation of HFETR with complex geometrical irradiation channel.

Retraction Note: Influence of soil rock mixture in mountain area based on machine learning and psychological intervention of left-behind children

With the rapid popularization of the Internet plus China and China’s rapid economic development, machine learning is playing an increasingly important role in model research. This paper presents a cold start method based on a hybrid Dirichlet process model and isolated forest. As we all know, the soil rock mixed slope is between the slope of soil and rock, most of its structure is loose and has large porosity, high permeability, and poor slope stability. Compared with soil slope and rock slope, the research on soil rock mixed slope is not comprehensive enough. Based on the study of geological environment conditions of highway slope, the mechanical properties of soil rock mixed slope are tested by traditional boundary equilibrium method and numerical simulation method. The stability of the three-dimensional slope is systematically studied in this paper. Nowadays, the mental health of left-behind children is still the focus of social attention. In this paper, a special school for left-behind children in a city is selected as the research object. The school is located between urban and rural areas. Since the children who stay in the school are from urban and rural areas, they will be selected as the research objects. Compared with ordinary children, left-behind children are more likely to have concerns, and their psychological changes are difficult to predict. What is more worrying is the long-term mental health problems. This paper aims to find the correlation between sports games and left-behind children’s mental health by introducing sports games in order to provide some methods and theoretical reference materials for solving the mental health problems of left-behind children.

Injection-production optimization of fault-karst reservoir\u2014considering high-speed non-Darcy effect

Fault-karst reservoirs are featured by complex geological characteristics, and accurate and fast simulation of such kind of reservoirs using traditional simulator and simulation methods is pretty hard. Herein, we tried to discrete the complex fault-karst structures into one-dimensional connected units connecting the well, fracture and cave based on reservoir static physical parameters and injection-production dynamics. Two characteristic parameters, conductivity and connected volume, are proposed to characterize the inter-well connectivity and material basis. Meanwhile, the high-speed non-Darcy seepage term is introduced into the material balance equations for well-fracture-cave connected units to describe the actual seepage characteristics within the fault-karst reservoirs, and to better simulate the oil/water production dynamics. Based on this method, a fracture reservoir model of 1 injection-3 production was established. The change of oil–water action law in different injection and extraction systems under two production regimes of fixed production rate and fixed pressure is analyzed. A case study was conducted on S fault zone, where the flow of oil and gas did not follow the Darcy seepage rule and with a β value of 103–104, the single well flow pressure and oil production were perfectly matched with the real data. In addition, connected units with more prominent high-speed non-Darcy features were found to have better connectivity, which might shed light on the more accurate prediction of inter-well connectivity. Moreover, an improved injection-production well pattern and was proposed based on connectivity prediction model and reservoir engineering method to solve the problems of insufficient natural energy supply and overhigh oil production rate in Block S. Furthermore, the injection/production rate as well as the timing and cycle of water injection was predicted and optimized so as to better guide to site operations.

Mesh-Free Surrogate Models for Structural Mechanic FEM Simulation: A Comparative Study of Approaches

The technical world of today fundamentally relies on structural analysis in the form of design and structural mechanic simulations. A traditional and robust simulation method is the physics-based finite element method (FEM) simulation. FEM simulations in structural mechanics are known to be very accurate; however, the higher the desired resolution, the more computational effort is required. Surrogate modeling provides a robust approach to address this drawback. Nonetheless, finding the right surrogate model and its hyperparameters for a specific use case is not a straightforward process. In this paper, we discuss and compare several classes of mesh-free surrogate models based on traditional and thriving machine learning (ML) and deep learning (DL) methods. We show that relatively simple algorithms (such as k-nearest neighbor regression) can be competitive in applications with low geometrical complexity and extrapolation requirements. With respect to tasks exhibiting higher geometric complexity, our results show that recent DL methods at the forefront of literature (such as physics-informed neural networks) are complicated to train and to parameterize and thus, require further research before they can be put to practical use. In contrast, we show that already well-researched DL methods, such as the multi-layer perceptron, are superior with respect to interpolation use cases and can be easily trained with available tools. With our work, we thus present a basis for the selection and practical implementation of surrogate models.

Deterministic modeling of solidification microstructure formation in directed energy deposition fabricated Ti6Al4V

Metal additive manufacturing (MAM) technology is now widely applied in the manufacturing and remanufacturing industry, among which microstructure simulation gradually shows its importance. Traditional solidification microstructure simulation methods all have their merits as well as drawbacks when applied in MAM cases. In this work, a deterministic solidification microstructure model, named “invasion model”, is established to avoid the essential defects of traditional methods. This model focuses on the interaction between the neighboring bi-crystals instead of simulating the growth kinetics of each columnar grain or deriving the field form of variables. Within a bi-crystal system, the grain boundary tilt from thermal gradient direction is understood as a transient invasion behavior of one grain to another, and the competitive grain growth behavior along the buildup process of MAM is a summary of all the invasions in bi-crystal systems. To fill the gaps in the rapid solidification theory, a database recording the anisotropic growth effect under a rapid directional solidification condition was established with the help of an artificial neural network (ANN). Wire feed directed energy deposition (DED) fabricated Ti6Al4V thin-wall samples with full dendritic columnar grains (prior beta) are used as a benchmark to test the validation of the novel simulation model. The grain geometrical structure of reconstructed prior β grains along the build-up direction has a good agreement with the simulation result. The model can also be applied to other cases of MAM or combine with various models to achieve real-time as-solidified crystallographic feature prediction when it meets the scope of application. • Traditional solidification microstructure simulation methods all have their merits and drawbacks when applied in MAM cases. • Invasion model is established to investigate the transient competitive grain growth behavior within bi-crystal systems. • Wire DED Ti6Al4V thin-wall samples are used as a benchmark to test the validation of the novel simulation model.