Zhonghua Ni

Application of ant colony optimization algorithm in process planning optimization

One objective of process planning optimization is to cut down the total cost for machining process, and the ant colony optimization (ACO) algorithm is used for the optimization in this paper. Firstly, the process planning problem, considering the selection of machining resources, operations sequence optimization and the manufacturing constraints, is mapped to a weighted graph and is converted to a constraint-based traveling salesman problem. The operation sets for each manufacturing features are mapped to city groups, the costs for machining processes (including machine cost and tool cost) are converted to the weights of the cities; the costs for preparing processes (including machine changing, tool changing and set-up changing) are converted to the ‘distance’ between cities. Then, the mathematical model for process planning problem is constructed by considering the machining constraints and goal of optimization. The ACO algorithm has been employed to solve the proposed mathematical model. In order to ensure the feasibility of the process plans, the Constraint Matrix and State Matrix are used in this algorithm to show the state of the operations and the searching range of the candidate operations. Two prismatic parts are used to compare the ACO algorithm with tabu search, simulated annealing and genetic algorithm. The computing results show that the ACO algorithm performs well in process planning optimization than other three algorithms.

An approach to mapping machining feature to manufacturing feature volume based on geometric reasoning for process planning

Efficient construction in three-dimensional working procedure model has an important effect on the efficiency and quality of machining process planning. This article proposes an algorithm to mapping machining features to manufacturing feature volumes from B-rep of mechanical part, and then, the working procedure models are generated. The mapping strategy is performed in three steps. In the first step, three types of machining features, such as depression feature, protrusion feature and transition feature, are introduced. In the second step, the edges of the chosen seed face and its neighboring faces are searched in order to generate the machining feature faces. The last step applies a closure to the machining feature faces to generate a compact solid by applying additional neighboring faces and their extensions. Then, the working procedure models are formed by combining the mapped manufacturing feature volumes with the final parts model. Compared with the existing working procedure model’s generation methods, this approach can avoid the unnecessary conversions from the engineering drawing to three-dimensional process models and from the machining knowledge to the modeling knowledge, which can greatly reduce the planning time on modeling. To validate the feasibility and validity of this approach, two machined parts with complex machining features are tested in the developed prototype computer-aided process planning system.

A new method of reusing the manufacturing information for the slightly changed 3D CAD model

Nowadays, growing quantities of process models in the process planning system are generated in enterprises based on the 3D design model. Usually, these process models contain a great deal of process information. So, how to reuse the embedded manufacturing information plays an important role in the competitive global market for the enterprises increasingly, because it can save enormous time and cost. In this paper, a new method is presented to reuse the existing manufacturing information for the slightly changed 3D design model. First, the hierarchical data structure of process route is introduced to represent the dynamic evolution process of mechanical part model, which takes working procedure model as the carrier of manufacturing information. Second, a multilevel machining feature descriptor which captures different levels of information for each process is designed to search the query machining feature. Moreover, the proposed algorithm of obtaining the main machining feature face for accelerating feature match is given to filter out unmatched feature. Then, the strategies of reusing the manufacturing information are proposed: the matched machining features can automatically obtain the manufacturing information and the unmatched machining features are dealt with by interactive mode. Finally, some slightly changed aircraft structural parts are utilized to verify our proposed method for reusing the manufacturing information. A prototype system has been developed to verify the effectiveness of the proposed method.

An intelligent approach for dimensioning completeness inspection in 3D based on transient geometric elements

Abstract Complete dimensioning plays an important role in digital product design. This paper proposes an intelligent reasoning approach to inspect the dimensioning completeness for 3D mechanical parts. Firstly, the problem of inspecting the dimension completeness is transformed into the problem of determining the geometric element (GE) constrained states. Then, 8 types of general geometric elements are decomposed into basic geometric elements (BGE) and basic dimensional elements (BDE), and an intermediate model named transient geometric element (TGE) is proposed to reflect the influences of geometric constraints (GC) on the BGEs. Thereafter, 3 types of complex constraints are decomposed into 6 types of simple constraints, which are then utilized to form the generation rules of the TGEs. Then, an upgrading geometric reasoning method of creating new TGEs is developed to get the BGE’s constrained state and the dimension usage status. Finally, the completeness states of dimensions are acquired according to the dimension usage status and the BGE’s constrained states. The presented approach is tested by a pre-dimensioned mechanical part composed of many typical geometric elements and constraints, and the result demonstrates that the dimensioning completeness states can be successfully inspected.

The stem-based vector space model for automatic resource allocation in workflow

In the context of workflow, resource allocation is often performed manually and empirically. In this paper, we present a novel approach based on stem-based vector space model (VSM) to automate the resource allocation works. The VSM similarity is utilized to compute the cohesion measures between the resources and the tasks, and the cohesion measures are then used to rank the resource candidates, and the resource with the highest score will be allocated to the specific workflow task. The initial evaluation on our experiments shows that our method can efficiently predicate the performers for the workflow tasks with good precision.

A Closed-loop Workflow Management Technique Based on Process Mining

A closed-loop workflow management system (C-WfMS) with process mining and result feedback technique was proposed to improve the efficiency of integrated manufacturing information system. The structure of organizational mining module (OME), the framework, and working method of C-WfMS were presented. The performer-to-performer relation (PPR) matrix and the activity-by-performer relation (APR) matrix were extracted from the workflow log. The PPR matrix is turned into social network matrix and put into SNA (Social Network Analysis) tools for further analysis; an algorithm for mining the APR matrix was presented, with which was the extraction of the weight matrix of APR. The implement of weight matrix of APR was also proposed so that the mining results could feedback to the WfMS organization dataset to make decision support for the assignments of tasks. Finally, a numerical example was presented.

The Parametric Casting Process Modeling Method Based on the Topological Entities Naming

In order to maintain the topological relations for the elements of casting process model and rebuild the casting process correctly when process parameters is changed, the topological entities naming method is used in this paper. Firstly, topological entities naming method, including naming rule for entity ID, Geometry_Name, and Process_Name, is proposed, and the mapping strategy between topological entity name and process information is established. Then, the geometry modeling method for casting process is studied in three steps: (1) process parameter setting method for gating and riser system; (2) parametric modeling procedure for gating and riser system; (3) the 3D process dimension design procedure for gating and riser system. Thirdly, the reconstruction method for parametric casting process model based on topological entity identification is given. A prototype of 3D casting process planning system is developed, and the reconstruction method is tested by an example part.

The backward growing method for constructing 3D process models in the machining process planning

A 3D process model shows the manufacturing information intuitively in different machining stages. Efficiently and easily constructing 3D process model is one of the most important key technologies in 3D computer aided process planning system and it meets the requirements of the current 3D manufacturing environment. Thus, this paper focuses on creating the 3D process models (3D-PMs) based on the backward growing methodology. Firstly, the 3D-PM is analyzed and decomposed into two parts: the working procedure model and the process attribute information. Then, the WPM construction method is proposed by using the geometrical reasoning knowledge and it is composed by the removal material volume and the after processing model. Then the process attribute information is appended on the working procedure models based on the proposed rules. Finally, the results of taking a machined part as the study case verify the validity of the developed method.

Notice of Retraction Partner selection of virtual enterprise using two-layer ant colony optimization

To improve the partner selection quality of virtual enterprise (VE), an improved ant colony optimization (ACO) named two-layer ACO (TL-ACO) was proposed in this paper. According to the features of partner selection, TL-ACO divided each selection step into two stages and each stage was given independent pheromone and visibility. As a result of that, the objective function could be much simpler than if original ACO was used, and the local optimum solution can be easily avoid by using the local obstruction (LO) method which was developed based on bionics. The simulation experiment manifested that the TL-ACO was effective and high-efficiency.

Development of CAPP based on 2.5D machining feature recognition

This paper proposes an architecture of CAPP based on machining feature recognition. The key techniques such as machining feature recognition, process planning optimization are discussed. In order to recognize the machining feature, the edges are divided into two types: convex Edge and concave Edge. The machining features are divided into four types: AOMF, HOMF, BCMF, and TCMF. The recognition process of the basic machining feature and the interactive process are given. Process planning optimization is to optimize the operation order and the machining resource. A mathematical model for process planning is given. A prototype is developed, and a example part is used for test.