Probabilistic Boolean network modeling and model checking as an approach for DFMEA for manufacturing systems

Modeling manufacturing processes assists the design of new systems, allowing predictions of future behaviors, identifying improvement areas and evaluating changes to existing systems. Probabilistic Boolean networks (PBN) have been used to study biological systems, since they combine uncertainty and rule-based representation. A novel approach is proposed to model the design of an automated manufacturing assembly processes using PBNs to generate quantitative data for occurrence assessment in design failure mode and effects analysis. FMEA is a widely used tool in risk assessment (RA) to ensure design outputs consistently deliver the intended level of performance. Effectiveness of RA depends upon the robustness of the data used. Temporal logic is applied to analyze state successions in a transition system, while interactions and dynamics are captured over a set of Boolean variables using PBNs. Designs are therefore enhanced through assessment of risks, using proposed tools in the early phases of design of manufacturing systems. A two-sample T test demonstrates the proposed model provides values closer to expected values; consequently modeling observable phenomena (\(p\,\hbox {value} > 0.05\)). Simulations are used to generate data required to conduct inferential statistical tests to determine the level of correspondence between model prediction and real machine data.     

Probabilistic verification of Boolean functions

We present a novel method for verifying the equivalence of two Boolean functions. Each function is hashed to an integer code by assigning random integer values to the input variables and evaluating an integer-valued transformation of the original function. The hash codes for two equivalent functions are always equal. Thus the equivalence of two functions can be verified with a very low probability of error, which arises from unlikely collisions between inequivalent functions. An upper bound, , on the probability of error is known a priori. The bound can be decreased exponentially by making multiple runs. Results indicate significant time and space advantages for this method over techniques that represent each function as a single OBDD. Some functions known to require space (and time) exponential in the number of input variables for these techniques require only polynomial resources using our method. Experimental results indicate that probabilistic verification can provide an attractive alternative for verifying functions too large to be handled using these OBDD-based techniques.     

Advanced data modeling for industrial drying machine energy optimization

The Journal of Supercomputing - In this article, a digital twin approach is proposed for modeling a pharmaceutical drying process using machine learning techniques, driven by data from different...     

Analysis of synchronous machine modeling for simulation and industrial applications

This paper analyses the synchronous machine modeling by taken into account the machine parameters usually used in industry and those used in researcher’s domains. Two models are presented. The first one is developed in the (d, q) natural reference frames and the other one is referred to the (d, q) stator reference frame. To do this, two methods are proposed to compute the reduction factor of the field winding without any input from design information. It is shown that the reduction factors of the (d, q) damper windings do not influence on the terminal behavior of the machine. This means that it is possible to know the terminal behavior of the machine without knowing the real inductances and resistances of the damper windings. The accuracy of these models is validated by experimental tests.     

Microcomputer control for an industrial sewing machine

The application of the microcomputer to various industrial machinery or equipment is expanding. The sewing industry is no exception to this trend, positively employing microcomputers for more advanced automation, labour saving and further industrialization. To satisfy recent needs, a sewing machine motor equipped with eddy current variable-speed clutch, LIMI-STOP Z, was developed. The LIMI-STOP Z features, besides a friction-free low inertial clutch used instead of the conventional friction type clutch, the use of a 4-bit microcomputer in the control system. This paper describes the principle, composition and functions of the LIMI-STOP Z.     

使用模型检测解决概率布尔网络优化控制

系统生物学期望对复杂生物系统建立一个真实的、可计算的模型,以便于以系统的角度去理解生物系统的演变过程。在系统生物学中,一个重要的主题是通过外部的干预控制发展关于基因调控网络的控制理论,以作为未来基因治疗技术。目前,布尔网络及其扩展的概率布尔网络已经被广泛用于对基因调控网络进行建模。在控制问题的研究中,概率布尔控制网络的状态迁移本质上构成一条有限状态空间的离散时间马尔科夫决策过程。依据马尔科夫决策过程的理论,通过概率模型检测方法解决网络中有限范围优化控制问题和无限范围优化控制问题。针对带有随机干扰且上下文相关的概率布尔控制网络,使用概率模型检测器PRISM对其进行形式化建模,然后将两类优化控制问题描述为相应的时序逻辑公式,最后通过模型检测寻找出最优解。实验结果表明,提出的方法可以有效地用于生物网络的分析和优化控制。     

Friction modeling and compensation for an industrial robot

This article describes the detailed friction modeling of the General Electric GP132 industrial robot. Friction is an area whose importance is often discounted in the development of control systems because it is thought to be insignificant or unmodelable. This work demonstrates that friction does have a predictable structure, and that significant performance improvement can be realized through its proper compensation. Experiments performed to determine the static, Coulomb, and viscous friction of the GP132 are presented. In addition to these components, the robot is shown to have significant gravity load-dependent and position-dependent friction. The accuracy of the friction models are verified through several experiments and are shown to be considerably better than previously formulated models.     

一个工业控制传感器网络系统的实现

由于传统大型工业企业生产的有线通讯方式存在着对传播媒介要求很高、节点的删除和添加十分复杂等问题,无线传感器网络被应用于工业环境的数据采集和传输。但同时,如何构建节能的无线传感器网络并使它在工业生产中满足实时控制的需要、在恶劣物理环境中能拥有足够的抗干扰能力也成为目前传感器网络的研究重点。广工计算机学院研发中心自主研发的工业控制传感器网络系统G2NET,能够自适应的形成拓扑结构,进行节点的添加和删除,并针对无线通讯协议实时性和可靠性这两个问题进行了研究,提出了解决方案。     

概率级联布尔网络的集镇定及其应用

随着系统生物学和医学的迅速发展,基因调控网络已经成为一个热点研究领域.布尔网络作为研究生物系统和基因调控网络的一种重要模型,近年来引起了包括生物学家和系统科学家在内的很多学者的广泛关注.本文利用代数状态空间方法,研究了概率级联布尔网络的集镇定问题.首先给出概率级联布尔网络集镇定的定义,并利用矩阵的半张量积给出了概率级联布尔网络的代数表示.其次基于该代数表示,定义了一组合适的概率能达集,并给出了概率级联布尔网络集镇定问题可解的充要条件.最后将所得的理论结果应用于概率级联布尔网络的同步分析及n人随机级联演化布尔博弈的策略一致演化行为分析.     

Simulation of a production line system with machine breakdowns using network modeling

In a manufacturing system, several factors can affect the reliability of the system in producing expected output levels, including resource input rates, labor rate variability, product quality, and machine failures. Although all of these factors will be reflected in the modeling process presented in this paper, the primary focus will be on the effect of machine breakdowns on system output. Network modeling and simulation with Q-GERT is the vehicle of analysis that is employed. This technique will be demonstrated via a realistic case example encompassing a complex production system consisting of several assembly lines, each containing several machines. The statistical results of the simulation of the example system are presented and discussed. In addition, examples of how the simulation model can be used to test changes in machine repair times and breakdown rates will be presented.     

Dynamic modeling and optimization of an industrial fluid catalytic cracker

Fluid Catalytic Cracking (FCC) is an important process which is used to convert heavy petroleum fractions into more valuable lighter products. In this work, the FCC process consists of the reactor, the regenerator and the fractionation units. Modeling is challenging due to the complex reaction chemistry and the interactions among the different process units. The reaction medium is modeled by the method of discrete lumping that uses narrow fractions. As a result, the number of discrete lumps (or pseudo-components) to model the process increases and this enables better prediction of fractionation products. For the reactor, we present a new kinetic model that includes a yield function for the cracking products. Kinetic constants and heat of cracking are correlated with the average boiling point of the pseudo-components. These correlations are next used in the development of first-principles models for the riser and the regenerator units. In addition, an empirical model is constructed for the purpose of predicting the individual amounts of the fractionation products from the reactor's effluent. Using parameter estimation, model parameters are estimated from actual industrial data. Model predictions match the plant measurements closely. Simulation and optimization results show that the developed model offers significant potential for use in real-time optimization and control.     

基于改进单类支持向量机的工业控制网络入侵检测方法

针对单类支持向量机（OCSVM）入侵检测方法无法检测内部异常点和离群点导致决策函数偏离训练样本的问题,提出了一种结合具有噪声的密度聚类（DBSCAN）方法和K-means方法的OCSVM异常入侵检测算法。首先通过DBSCAN算法,剔除训练数据中的离群点,消除离群点的影响;然后利用K-means划分数据类簇的方法筛选出内部异常点;最后利用OCSVM算法为每一个类簇建立单分类器用于检测异常数据。工控网络数据集上的实验结果表明,该组合分类器能够利用无异常数据样本检测出工控网络入侵,并且提高了OCSVM方法的检测效果。在气体管道网络数据集入侵检测实验中,所提方法的总体检测率为91.81%;而原始OCSVM算法则为80.77%。     

BN-LSTM-based energy consumption modeling approach for an industrial robot manipulator

Industrial robots (IRs) are widely used to increase productivity and efficiency in manufacturing industries. Therefore, it is critical to reduce the energy consumption of IRs to maximize their use in polishing, assembly, welding, and handling tasks. This study adopted a data-driven modeling approach using a batch-normalized long short-term memory (BN-LSTM) network to construct a robust energy-consumption prediction model for IRs. The adopted method applies batch normalization (BN) to the input-to-hidden transition to allow faster convergence of the model. We compared the prediction accuracy with that of the 1D-ResNet14 model in a UR (UR3e and UR10e) public database. The adopted model achieved a root mean square (RMS) error of 2.82 W compared with the error of 6.52 W achieved by 1D-ResNet14 model prediction, indicating a performance improvement of 56.74%. We also compared the prediction accuracy over the UR3e dataset using machine learning and deep learning models, such as regression trees, linear regression, ensemble trees, support vector regression, multilayer perceptron, and convolutional neural network-gated recurrent unit. Furthermore, the layers of the well-trained UR3e power model were transferred to the UR10e cobot to construct a rapid power model with 80% reduced UR10e datasets. This transfer learning approach showed an RMS error of 3.67 W, outperforming the 1D-ResNet14 model (RMS error: 4.78 W). Finally, the BN-LSTM model was validated using unseen test datasets from the Yaskawa polishing motion task, with an average prediction accuracy of 99%.     

A machine vision system for enhancing the teleoperation of an industrial robot

This paper describes a real-time vision system that enhances the teleoperation of a servicing tool used in the heat exchangers of nuclear power plants. The vision system is used to track the position of the tool as it moves over a sheet of tube ends. A map-based strategy is adopted for the estimation of the position. The system incorporates a novel method for a foreshortening correction that is applied prior to map referencing. A hypothesize and verify scheme locates two image features that correspond to two map features. An efficient scheme for extracting image features is developed to locate these two features (tube-end centers) in the image. Two different types of heat-exchanger tube sheets are accounted for. They are those with tube ends placed in a square grid and those with tube ends placed in a triangular grid. The map-based strategy minimizes the cumulative errors in the estimate of the tool head position. The resulting low-cost system has been tested on synthetic and real data. Performance results are given.     

Reasoning Boolean operation based modeling for heterogeneous objects

A reasoning Boolean operation based CAD modeling approach applied to construct heterogeneous material objects is presented. This reasoning Boolean operation consists of a merging operation and an extracting operation, and it is executed according to the material-dominant information defined in the designed heterogeneous object database. In addition to the material-dominant Boolean union, subtraction, and intersection, a new Boolean complex_union operation is defined and introduced to the set of the reasoning Boolean operation. The Boolean complex_union ‘assembles’ the results of the material-dominant Boolean intersection and subtraction to form a new modeling assembly for the heterogeneous object. Due to the CAD-based nature, the thus formed heterogeneous model can be readily implemented with advanced CAD/CAE/CAM software for integrated design, analysis, and simulation. An example of such an application, its hierarchy of the model database, and the major steps of the model construction are described.     

Neural network modeling of torque estimation and d–q transformation for induction machine

This paper presents a neural network approach in modeling of torque estimation and Parks d–q transformation for an open-loop induction machine. The nonlinear approximation capability of neural networks makes it possible to map the Parks d–q transformation and torque estimation in an induction motor, which would otherwise require extensive complex calculations. The neural network simulation results will be compared to those of directly DSP calculated transformation and estimation. The results show improved performance with the neural network approach. We conclude that machine systems transformations and estimations can take advantage of the neural network technology for improved performance and cost reduction in the long run.     

A sparse matrix algorithm on the Boolean vector machine

The Boolean Vector Machine (BVM) is a large network of extremely small processors with very small memories operating in SIMD mode using bit serial arithmetic. Individual processors communicate via a hardware implementation of the Cube Connected Cycles (CCC) network. A prototype BVM with 2048 processing elements, each with 200 binary bits of memory, is currently being built using VLSI technology.

The BVM's bit-serial arithmetic and the small memories of individual processors are apparently a drawback to its effectiveness when applied to large numerical problems. In this paper we analyze an implementation of a basic matrix-vector iteration algorithm for sparse matrices on the BVM. We show that a 2²⁰ Pe BVM can deliver over 1 billion (10⁹) useful floating-point operations per second for this problem. The algorithm is expressed in a new language (BVL) which has been defined for programming the BVM.