Design Of Complex Mechanical Systems Research Articles

REVIEW OF MODERN USE OF GENETIC AND EVOLUTIONARY ALGORITHMS. STRATEGIES, POSSIBILITIES (REVIEW ARTICLE)

Modern trends in the optimal and rational design of technical objects cross a large number of directions of their implementation. One of the interesting and promising directions is genetic and evolutionary algorithms (GА and EA). Authors promote the use of GА and EA as a tool for solving problems of optimal and rational design of complex mechanical systems. The relevance of highlighting modern methods, approaches and strategies for the implementation of GА and EA is described, as well as consideration of their applied implementation, which makes it possible to identify interesting directions of research that, with further adaptation or modifications, can be used to solve the problems of optimal and rational design of gearboxes, boxes gears and transmissions. The main general directions of the literature on GА and EA are highlighted, as well as the practical use of GА and EA in: technical and technological activities, physics, construction, water systems, nanotechnologies, analytical and simulation modeling, electrical and electronic systems, modeling of artificial intelligence and neural networks, information technologies, economic theory, administration and management, marketing, sociology, biology and medicine. This made it possible to understand the course of scientific thought on this issue, to determine the advantages and disadvantages of existing directions and approaches, and helped to choose the vector of further scientific thought, to decide on interesting approaches, strategies and methods. Considering certain features of EA, the authors prefer them. And in terms of strategies, hybridization with other methods, maximum saturation of all stages with "randomness" and the possibility of learning (memory organization) of the algorithm similar to neural networks are promising. Keywords: optimal design, research directions, genetic algorithms, evolutionary algorithms

Improving the Visual Interactive Analysis Method for Automation and Control of the Decision-Making Process in Multi-Criteria Design of Complex Mechanical Systems

For multi-criteria design problems of complex mechanical systems with a large number of control parameters, technical constraints, and quality criteria, the search for Pareto solution domain takes quite a lot of time varying from hours to days. In fact, the decision-maker (DM) desires to examine a small number of reasonable Pareto optimal solutions in order to understand the problem itself and control the decisionmaking in a simple manner. This paper presents the improvement of a visual interaction analysis method or VIAM developed by the authors with the aim of providing a tool for DM to define the optimal and mutually-agreed solutions in the multi-criteria decision making (MCDM). Indeed, VIAM allows for evaluating the distribution domain of the Pareto optimal solutions defined by the genetic algorithm, which supports the DM to set additional thresholds for the objectives to filter the desired solutions and suggest to shrink or expand the threshold to control the search. In case of mutually-agreed solution non-existence, VIAM allows for providing instruction to reestablish the multi-objective problem that new Pareto solution domains can be found as desired by the DM. Based on VIAM, a visual interaction analysis tool or VIAT was developed by means of Matlab. VIAT was then used for the multi-criteria design of slider-crank mechanism for an innovative fruit vegetable washer with three objectives. Comparative study on the obtained results from VIAT with the existing design option and the obtained solution from the traditional method "concession by priority" has shown the effectiveness of the method proposed in this paper. VIAT is actually a very user-friendly tool that makes the multi-criteria design more practical especially for the mechanical system.

A structured method for the design-for-frequency of a brace-soundboard system using a scalloped brace

Design-for-frequency of mechanical systems has long been a practice of iterative procedures in order to construct systems having desired natural frequencies. Especially problematic is achieving acoustic consistency in systems using natural materials such as wood. Inverse eigenvalue problem theory seeks to rectify these shortfalls by creating system matrices of the mechanical systems directly from the desired natural frequencies. In this paper, the Cayley–Hamilton and determinant methods for solving inverse eigenvalue problems are applied to the problem of the scalloped braced plate. Both methods are shown to be effective tools in calculating the dimensions of the brace necessary for achieving a desired fundamental natural frequency and one of its higher partials. These methods use the physical parameters and mechanical properties of the material in order to frame the discrete problem in contrast to standard approaches that specify the structure of the matrix itself. They also demonstrate the ability to find multiple solutions to the same problem. The determinant method is found to be computationally more efficient for partially described inverse problems due to the reduced number of equations and parameters that need to be solved. The two methods show great promise for techniques which could lead to the design of complex mechanical systems, including musical instrument soundboards, directly from knowledge of the desired natural frequencies.

Multidisciplinary Design and Collaborative Optimization for Excavator Backhoe Device

The excavator working device is a typical mechanical system of electromechanical liquid that is complex. Traditional optimization design methods are difficult to get global optimized results of excavator backhoe device through the serial mode of “mechanism-load-structure”. Thus, the theory of parallel collaborative optimization (CO) is applied. To establish a sophisticated CO model of the backhoe device, a certain excavator is investigated as a sample multidisciplinary CO (MDCO) design. To generate the CO model, an improved optimization algorithm called the particle swarm-genetic algorithm （PS-GA）is proposed. To verify the MDCO design of the excavator backhoe device, a parameterized virtual prototype (VP) of the backhoe device is established in ADAMS. This VP is optimized by applying the MDCO design results to the parameterized VP. The VP of the backhoe device is also optimized by a single discipline when the optimization results from a single discipline are inputted into the parameterized VP. Both optimized VPs are simulated under similar conditions, and results show that in the MDCO design, the arm crowd force of the backhoe device is 8.1% stronger than that in the design optimized by a single discipline under constant power and oil pressure conditions. Similarly, breakout force increased by approximately 8.3%. The quality (volume) of the entire backhoe device decreased by 9.5%; however, the maximum stress of each characteristic partition changed only slightly. Therefore, the MDCO design effectively and practically addresses problems regarding the optimization of the design of complex mechanical systems.

Analytical mechanics approaches in the dynamic modelling of Delta mechanism

SUMMARYThe increasing importance of computational models for the design of complex mechanical systems raises a discussion on defining some criteria for the selection of adequate modelling methods. This paper aims to contribute to such discussion from an educational point of view. By choosing the Delta parallel mechanism as a typical representative of multi-body mechanical systems, four approaches – one based on the Principle of Virtual Work, two based on Lagrange's formalism, and one based on Kane's formalism – are analysed from the perspective of modelling procedures. Finally, inverse dynamic simulations are carried out along with qualitative comparisons of the considered approaches.

Discrete Time Transfer Matrix Method for Launch Dynamics Modeling and Cosimulation of Self-Propelled Artillery System

In many industrial applications, complex mechanical systems can often be described by multibody systems (MBS) that interact with electrical, flowing, elastic structures, and other subsystems. Efficient, precise dynamic analysis for such coupled mechanical systems has become a research focus in the field of MBS dynamics. In this paper, a coupled self-propelled artillery system (SPAS) is examined as an example, and the discrete time transfer matrix method of MBS and multirate time integration algorithm are used to study the dynamics and cosimulation of coupled mechanical systems. The global error and computational stability of the proposed method are discussed. Finally, the dynamic simulation of a SPAS is given to validate the method. This method does not need the global dynamic equations and has a low-order system matrix, and, therefore, exhibits high computational efficiency. The proposed method has advantages for dynamic design of complex mechanical systems and can be extended to other coupled systems in a straightforward manner.

A procedure for the design of complex mechanical systems by decomposition

The design of many practical mechanical systems usually requires dealing with large-scale problems incorporating different physical phenomena with diverse behavioural constraints that are generally non-linear. This paper presents a generalised procedure to aid the designer in making such problems tractable through appropriate segmentation. Accordingly, the interactions between the different components and parameters become clearer, allowing interactive modelling improvement and design decisions to be made serially with immediate feed back on how they influence the design objective. The procedure also aids the designer in identifying and easily correcting any constraint inconsistency or programming errors. Although the methodology is developed for large-scale systems, simple examples are considered for illustration and good feasible solutions can be obtained without extensive computing effort. There is no guarantee, however, that a global optimum design can be achieved by this approach, although in all the tried cases, the solution was optimum.

A form verification system for the conceptual design of complex mechanical systems

This paper presents a summary of research into the development and implementation of a domain-independent, computer-based model for the conceptual design of complex mechanical systems [1]. The creation of such a design model includes the integration of four major concepts: (1) the use of a graphical display for visualizing the conceptual design attributes: (2) the proper representation of the complex data and diverse knowledge reguired to design the system; (3) the integration of quality design methods into the conceptual design: (4) the modeling of the conceptual design process as a mapping between functions and forms. Using the design of an automobile as a case study, a design environment was created which consisted of a distributed problem-solving paradigm and a parametric graphical display. The requirements of the design problem with respect to data representation and design processing were evaluated and a process model was specified. The resulting vehicle design system consists of a tight integration between a blackboard system and a parametric design system. The completed system allows a designer to view graphical representations of the candidate conceptual designs that the blackhoard system generates.

Effective component design synthesis using structural and shape optimization methods

An efficient optimization-based component design synthesis methodology is developed to expedite the mechanical design of complex mechanical systems that contain several components. This method offers a modular design approach wherein the system design is obtained by representing all the components using lumped properties first and then synthesizing each component shape later to achieve those pre-optimized lumped properties. Methodology consists of four steps: (1) develop a lumped parameter analysis model of the component, (2) determine those lumped parameters such as stiffness and mass using structural optimization methods, (3) parameterize the component shape using pre-selected topology, and (4) optimize the component shape as a separate problem to achieve predetermined stiffness/mass characteristics. A new shape optimization software Optshape was developed using MSC/ Nastran's grid sensitivity capability to facilitate component shape optimization. Design of a spring joint for an automotive exhaust system is presented to demonstrate the effectiveness of the method and the Optshape software.