Gerald M. Knapp

Determining the most important criteria in maintenance decision making

Many maintenance decisions require the evaluation of alternative solutions in terms of complex maintenance criteria such as cost, repairability, reliability and availability requirements. Such problems can be formulated as multi‐criteria decision‐making problems. The relative importance of maintenance criteria is difficult to assess, and therefore a sensitivity analysis becomes a necessity. The sensitivity analysis approach presented reveals some counter‐intuitive results and can considerably enhance the application of decision analysis in complex maintenance management.

Statistical‐based or condition‐based preventive maintenance?

The focus of preventive maintenance (PM) programmes in industry is shifting from a pure statistical basis to online condition monitoring. Examines the shortcomings of statistical‐based PM which are contributing to this shift, and the potential benefits of and current research issues within condition‐based PM. Notes that statistics and quality control techniques will continue to play a critical role in this evolution.

Some aspects of measuring maintenance performance in the process industry

Performance indicators of operational maintenance can help maintenance staff improve its operations, so that the direct and indirect costs of failure processes can be reduced. Many papers have been written on performance indicators for operational maintenance. However, no consensus on which indicators to use in a particular industry has been reached so far. The authors take an industrial engineering approach to this problem by describing the information system needed to be able to make any inferences on operational maintenance performance in the process industry. The indicators suggested focus on finding the most costly equipment from a maintenance perspective, the cost of the current maintenance concept and the major components of maintenance costs. It is emphasized that standards and procedures need to be developed and that adherence to them has to be ensured.

A fuzzy neural network approach to machine condition monitoring

This paper is focused on the implementation of a predictive neural network for use as an operators aid in the diagnosis of faults with high prediction accuracy in an automated manufacturing environment. In order to evaluate the performance of the model, the network has been tested using both simulated time series and real time machine vibration data collected in lab experiments.

Emotion Recognition in Text for 3-D Facial Expression Rendering

Emotions are a key semantic component of human communication. This study focuses on automatic emotion detection in descriptive sentences and how this can be used to tune facial expression parameters for 3-D character generation. A comparison of manual and automatic word feature selection approaches is performed to determine the influence of word features on classification accuracy using support vector machines (SVM). The automatic emotion feature selection algorithm presented here builds on the framework used by mutual information for feature selection. Results of the study indicate that the set of automatically selected features was as good as the set of manually selected features. The proposed automatic feature selection algorithm implemented in this study helped to detect new words from the training corpus which were relevant to the classification task but were not considered by the researchers. An example of potential outcomes from facial expression tuning is also presented. The analysis includes initial results for dealing with the class imbalance challenge present in the data.

Multiple setup PCB assembly planning using genetic algorithms

A Genetic Algorithm (GA) program is developed for simultaneously optimizing component placement sequence and feeder assignments in the assembly of Printed Circuit Boards (PCBs). The program extends the application of GA to this problem by providing handling of two practical but complicating factors: (1) feeder constraints forcing multiple setups, and (2) new generation assembly machines which can place from both sequenced tape and component feeders within the same setup. This paper details the algorithm developed, demonstrates the approach on several examples, and investigates the performance of the GA and the impact of its various parameters.

Optimization of maintenance organization and manpower in process industries

The objective of maintenance manpower planning is to have the right number of workers with the right capabilities in the right maintenance areas. In this research, a model was developed for optimizing manpower allocation by maintenance area, craft‐type, training level, and in‐house versus sub‐contracted employees, as well as selecting between a centralized versus decentralized organizational structure.

Equipment assignment to multiple overhaul blocks in series systems

This paper covers the development of a heuristic method for grouping equipment into blocks for application of preventive maintenance overhauls within a series system, so as to minimize total maintenance‐related costs for the system. Previously, group technology (GT) concepts have not been applied to this problem, and this research investigated the applicability of such concepts to the block overhaul problem (specifically, the SLCA method was applied). Performance of the heuristic is analyzed with respect to runtime and solution quality.

A fuzzy scheme for failure mode screening

A fuzzy reasoning algorithm was developed and implemented via an expert system to evaluate and assess the likelihood of equipment failure mode precipitation and aggravation. The scheme is based upon the fuzzification of the effects of precipitating factors provoking the failure. It consists of a fuzzy mathematical formulation which linearly relates the presence of factors catalogued as critical, important or related to the incidence of machine failure modes. This fuzzy algorithm was created to enable the inference mechanism of a constructed knowledge-based system to screen industrial equipment failures according to their likelihood of occurrence.

An ARTMAP neural network‐based machine condition monitoring system

Presents a real‐time neural network‐based condition monitoring system for rotating mechanical equipment. At its core is an ARTMAP neural network, which continually monitors machine vibration data, as it becomes available, in an effort to pinpoint new information about the machine condition. As new faults are encountered, the network weights can be automatically and incrementally adapted to incorporate information necessary to identify the fault in the future. Describes the design, operation, and performance of the diagnostic system. The system was able to identify the presence of fault conditions with 100 percent accuracy on both lab and industrial data after minimal training; the accuracy of the fault classification (when trained to recognize multiple faults) was greater than 90 percent.