Artificial neural network approach for estimating weld bead width and depth of penetration from infrared thermal image of weld pool

Abstract

In this article an artificial neural network based system to predict weld bead geometry using features derived from the infrared thermal video of a welding process is proposed. The multilayer perceptron and radial basis function networks are used in the prediction model and an online feature selection technique prioritises the features used in the prediction model. The efficacy of the system is demonstrated with a number of welding experiments and using the leave one out cross-validation experiments.     

Online quality inspection of resistance spot welded joint based on electrode indentation using servo gun

Abstract

Resistance spot welding is one of the major joining techniques widely used for car body assembly. Weld quality may significantly influence the durability and reliability of the automobile body. Automotive manufacturers often rely on destructive testing and monitoring variables which indirectly reflected weld quality to assess the weld quality and control the welding process. However, these approaches have inherent limitations and are difficult to be implemented in plant environments. Therefore, it is imperative to develop an online inspection method to evaluate weld quality. In the present study, a method of producing a series of substantially uniform spot welds between two metal parts using a servomotor driven movable electrode and an axially opposing fixed electrode is proposed. The indentation in the workpiece surface is suitably measured by the displacement of the movable electrode as it applies an electrode force and welding current is passed through the weld site of the workpiece. The optimal indentation range is determined by peel test and metallographic examinations with respect to various sheet gages and grades. Consequently, online weld quality inspection results are achieved based on developed optimal indentation range.     

Elastomeric and electrically conductive materials on basis of thermoplastic elastomers and their controlled manufacturing

Abstract

The present work presents a possibility to produce a rubber elastic and electrically conductive polymer material on the basis of dynamic vulcanisates. Thanks to the specific morphology of dynamic vulcanisates and the non-uniform carbon black distribution, carbon black filled dynamic vulcanisates can exhibit a very low percolation threshold of ～4 wt-%. Keeping the carbon black content low, a broad spectrum of resistivity properties can be achieved by variation of material factors like type and content of rubber phase and filler, concentration of cross-linking agent and compatibiliser and technological factors like mixing time respectively. In comparison with thermoplastic elastomers on the basis of block copolymers dynamic vulcanisates show a distinct lower percolation threshold. Up to a carbon black content of ～10 wt-% the mechanical properties of carbon black filled dynamic vulcanisates are not negative influenced essentially. To characterise the development of the carbon black dispersion and distribution processes and the conductivity properties in an internal mixer, the method of online measured electrical conductivity is suited very well for carbon black containing rubber mixtures. It could be shown in pre-investigations that this method promises to be a very useful tool for monitoring the mixing processes of carbon black filled dynamic vulcanisates in continuous mixing processes by means of extruders too.     

Precision online model for prediction of strip temperature in hot strip rolling

Abstract

Investigated via a series of finite element (FE) process simulation is the effect of diverse process variables on some selected non-dimensional parameters characterising the thermomechanical behaviour of the roll and strip in hot strip rolling. Then, on the basis of these parameters, online models are derived for precise prediction of the temperature changes occurring in the bite zone as well as in the interstand zone in a finishing mill. The prediction accuracy of the proposed models is examined through comparison with predictions from a FE process model.     

Modelling weld bead geometry using neural networks for GTAW of austenitic stainless steel

Abstract

The authors analyse the importance of different weld control parameters on the weld pool geometry of gas tungsten arc welding using an online feature selection technique that suggests weld voltage and vertex–angle pair as more important than the weld voltage and torch speed pair. Using the selected features multi layer perception and radial basis function networks are developed for prediction of bead width, penetration depth, and bead area. With cross-validation the authors have extensively studied the performance of composite models (one model for all outputs) and individual models (one model for each output). The individual models are found to work better than composite models. Usually, radial basis function networks are found to work better than the multi layer perception networks. To assess the influence of weld control parameters the authors have studied the performance of both networks using different combination of inputs. Overall, the performance of the proposed models is found to be quite satisfactory.     

几种合成绝缘子带电检测方法的比较

分析了观察法、电场法、紫外成像法和红外成像法这 4种合成绝缘子带电检测方法的实验室和现场运行情况 ,总结出它们不同的主要适用范围 :观察法、电场法分别适于外绝缘、导通性内绝缘缺陷 ;紫外成像法、红外成像法分别适于能产生局部放电、局部发热的缺陷。比较而言电场法成本低廉 ,直接检测 ,准确可靠 ,操作简单 ,更适于国内推广使用     

TRNSYS The most complete solar energy system modeling and simulation software

The five computer programs TRNSYS, PRESIM, TRNSED, ONLINE and PREBID have been put together into a program package which is the most complete solar energy system modeling and simulation software that is available today.     

Mixed-Load Transportation Scheduling of Multiple Agents in a Warehouse Environment

In a floor warehouse environment, where many agents transport numbers of products iteratively, efficient product transportation is required for high productivity. In general, the following facts are recognized. (i) Three kinds of jobs exist: picking, replenishment and arrangement. (ii) Mixed-load transportation can decrease the completion time of vehicles. In many former methods, however, the above facts are not considered. We propose a design methodology using a dispatching rule and a local search method based on a simulated annealing algorithm. The rule is applied to obtain the initial solution and the local search method is adopted to obtain a semi-optimal solution in consideration of mixed-load transportation. To evaluate the applicability of the proposed algorithm, we have implemented the algorithm in several product environments. The simulation results verify the effectiveness of the proposed method.     

ROBOT＇S MOTION ERROR AND ONLINE COMPENSATION BASED ON FORCE SENSOR

Robot's dynamic motion error and on-line compensation based on multi-axis force sensor are dealt with.It is revealed that the reasons of the error are formed and the relations of the error are delivered.A motion equation of robot's termination with the error is established,and then,an error matrix and an error compensation matrix of the motion equation are also defined.An on-line error's compensation method is put forward to decrease the displacement error,which is a degree of millimeter,shown by the result of simulation of PUMA562 robot.     

Macro- and microdispersion of carbon black in liquid silicone rubbers

Abstract

A correlation between the online during mixing measured conductance of carbon black (CB) filled silicone rubber and macrodispersion of CB determined by optical microscopy and microdispersion analysed by AFM has been established. The online conductance is a suitable tool to characterise the effect of technological parameters (e.g. rotor type, speed, mixing temperature) and material parameters (matrix viscosity, CB type and content) on the macro- and microdispersion directly during the mixing process. Thus, a modified method of the online conductance measurement can be used as the in-process method for monitoring the production process of electrical conductive polymer materials with low viscosity, e.g. liquid silicone rubbers.