Inspection Dataset Research Articles

Index-based triangulation method for efficient generation of large three-dimensional ultrasonic C-scans

The demand for high-speed ultrasonic scanning of large and complex components is driven by a desire to reduce production bottlenecks during the non-destructive evaluation (NDE) of critical parts. Emerging systems (including robotic inspection) allow for the collection of large volumes of data in short time spans, compared to existing inspection systems. To maximise throughput, it is crucial that the reconstructed inspection datasets are generated and evaluated rapidly without loss of detail. This requires new data visualisation and analysis tools capable of mapping complex geometries while guaranteeing full coverage. This paper presents an entirely new approach for the visualisation of three-dimensional (3D) ultrasonic C-scans, suitable for application to high data throughput ultrasonic phased array inspection of large and complex parts. Existing reconstruction approaches are discussed and compared with the new index-based triangulation (IBT) method presented. The IBT method produces 3D C-scan representation, presented as coloured tessellated surfaces, and the approach is shown to work efficiently, even on challenging geometries. An additional differentiating characteristic of the IBT method is that it allows for easy detection of lack of coverage (an essential feature for ensuring that inspection coverage can be guaranteed on critical components). The results demonstrate that the IBT C-scan generation approach runs over 60 times faster than a C-scan display based on Delaunay triangulation and over 500 times faster than surface reconstruction C-scans. In summary, the main benefits of the new IBT technique include: high-speed generation of C-scans on large ultrasonic datasets (orders of magnitude improvement compared to surface reconstruction C-scans); the ability to operate efficiently on 3D mapped datasets (allowing 3D interpretation of C-scans on complex geometry components); and intrinsic indication of lack of inspection coverage.

Кластерный анализ изменений пространственного положения трубных секций магистрального нефтепровода по данным внутритрубных обследований

Кластерный анализ изменений пространственного положения трубных секций магистрального нефтепровода по данным внутритрубных обследований

Extreme value analysis (EVA) of inspection data and its uncertainties

Extreme value analysis (EVA) is a statistical tool to estimate the likelihood of the occurrence of extreme values based on a few basic assumptions and observed/measured data. While output of this type of analysis cannot ever rival a full inspection, it can be a useful tool for partial coverage inspection (PCI), where access, cost or other limitations result in an incomplete dataset. In PCI, EVA can be used to estimate the largest defect that can be expected. Commonly the return level method is used to do this. However, the uncertainties associated with the return level are less commonly reported on. This paper presents an overview of how the return level and its 95% confidence intervals can be determined and how they vary based on different analysis parameters, such as the block size and extrapolation ratio. The analysis is then tested on simulated wall thickness data that has Gaussian and Exponential distributions. A curve that presents the confidence interval width as a percentage of the actual return level and as a function of the extrapolation ratio is presented. This is valid for the particular scale parameter (σ) that was associated with the simulated data. And for this data it was concluded that, in general, extrapolations to an area the size of 500–1000 times the inspected area result in acceptable return level uncertainties (<20% at 95% confidence). When extrapolating to areas that are larger than 1000 times the inspected area the width of the confidence intervals can become larger than 30–50% of the actual return level. This was deemed unacceptable: for the example of wall thickness mapping that is used throughout this paper, these uncertainties can represent critical defects of nearly through wall extent. The curve that links the confidence interval width to the return value as a function of extrapolation ratio is valid only for a particular scale parameter value of the EVA model. However, it is imagineable that a few of such relations for different scale parameters σ could be simulated. By picking the relation with the closest σ value (based on observation or estimation) for the inspection dataset, the presented approach can then be used to quickly estimate the uncertainty associated with an EVA extrapolation.

APRICOIN: An Adaptive Approach for Prioritizing High-Risk Containers Inspections

Risk evaluation of the containers remains a difficult task, often due to incomplete or ambiguous information on containers. In addition, the evaluation process needs to be adapted on an ongoing basis to cope with emerging risk factors. Furthermore, high-risk container inspection is commonly hindered by a low inspection capacity, which leads to a major issue: how can we prioritize the container inspection if the number of suspect containers exceeds the inspection capacity? Container inspection prioritizing may be the answer. In this paper, we propose a novel approach for adaptively prioritizing container inspection (APRICOIN). First, we enhance the container information flow to alleviate the problem of incomplete information by proposing an enhanced container descriptive. Second, we introduce the APRICOIN algorithm, which combines frequent pattern mining and a fuzzy logic system, to assess the container’s risk score. The frequent pattern growth algorithm is proposed to retrieve the key criteria for evaluating container risk. This is done through mining frequent criteria sets within the historic data set of container inspections by customs. The mined frequent criteria sets are used to assess fuzzy inference rules which are periodically readjusted to integrate new key criteria. Thereafter, the fuzzy logic system uses the obtained fuzzy inference rules to calculate a container’s risk score. Our major contribution consists of providing a new adaptive approach for assessing a container’s risk through combining frequent criteria mining and fuzzy logic. An illustrative study and a comparison with alternative approaches are performed to validate the proposed algorithm.

Evidence-Based Patient Classification for Traditional Chinese Medicine.

Evidence-Based Patient Classification for Traditional Chinese Medicine.

Comparing the utility of decision trees and support vector machines when planning inspections of linear sewer infrastructure

Closed-circuit television inspection technology is traditionally used to identify aging sewer pipes requiring rehabilitation. While these inspections provide essential information on the condition of pipes hidden from day-to-day view, they are expensive and often limited to small portions of an entire sewer system. Municipalities may benefit from utilizing predictive analytics to leverage existing inspection datasets so that reliable predictions of condition are available for pipes that have not yet been inspected. The predictive capabilities of data mining systems, namely support vector machines (SVMs) and decision tree classifiers, are demonstrated using a case study of sanitary sewer pipe inspection data collected by the municipality of Guelph, Ontario, Canada. The modeling algorithms are implemented using open-source software and are tuned to counteract the negative impact on predictive performance resulting from class imbalance common within pipe inspection datasets. The decision tree classifier outperforms SVM for this classification task – achieving an acceptable area under the receiver operating characteristic curve of 0.77 and an overall accuracy of 76% on a stratified test set. Although predicting individual pipe condition is a notoriously difficult task, decision trees are found to be a useful screening tool for planning future inspection-related activities.

Predicting the structural condition of individual sanitary sewer pipes with random forests

Closed-circuit television inspections of sewer condition deterioration as required for proactive management are expensive and hence limited to portions of a sewer network. The data mining approach presented herein is shown capable of unlocking information contained within inspection records and enhances existing pipe inspection practices currently used in the wastewater industry. Predictive models developed using the random forests algorithm are found capable of predicting individual sewer pipe condition so that uninspected pipes in a sewer network with the greatest likelihood of being in a structurally defective condition state are identified for future rounds of inspection. Complications posed by imbalance between classes common within inspection datasets are overcome by first establishing the classification task in a binary format (where pipes are in either good or bad structural condition) and then using the receiver-operating characteristic (ROC) curve to establish alternative cutoffs for the predicted class probability. The random forests algorithm achieved a stratified test set false negative rate of 18%, false positive rate of 27% and an excellent area under the ROC curve of 0.81 in a case study application to the City of Guelph, Ontario, Canada. The novel inclusion of condition information of pipes attached at either the upstream or downstream manholes of an individual pipe enhances the predictive power for bad pipes representing the minority class of interest (reducing the false negative rate to 11%, reducing the false positive rate to 25% and increasing the area under the ROC curve to 0.85). An area under the ROC curve &gt;0.80 indicates random forests are an “excellent” choice for predicting the condition of individual pipes in a sewer network.

Time‐Series Models for Border Inspection Data

We propose a new modeling approach for inspection data that provides a more useful interpretation of the patterns of detections of invasive pests, using cargo inspection as a motivating example. Methods that are currently in use generally classify shipments according to their likelihood of carrying biosecurity risk material, given available historical and contextual data. Ideally, decisions regarding which cargo containers to inspect should be made in real time, and the models used should be able to focus efforts when the risk is higher. In this study, we propose a dynamic approach that treats the data as a time series in order to detect periods of high risk. A regulatory organization will respond differently to evidence of systematic problems than evidence of random problems, so testing for serial correlation is of major interest. We compare three models that account for various degrees of serial dependence within the data. First is the independence model where the prediction of the arrival of a risky shipment is made solely on the basis of contextual information. We also consider a Markov chain that allows dependence between successive observations, and a hidden Markov model that allows further dependence on past data. The predictive performance of the models is then evaluated using ROC and leakage curves. We illustrate this methodology on two sets of real inspection data.

性能規定に基づく寒冷地トンネル覆工の劣化予測のためのマルコフ遷移確率行列の同定

In this paper propose, a methodology to identify the Markovian transition probability matrix, which is presented to forecast the degradation process of cold region tunnel lining for performance based design. Reliability, effect of rehabilitation, and steady state probability of tunnel linings, can be evaluated by this process. The degradation states of the tunnel linings are categorized into several ranks, and their degradation processes are characterized by the tunnel management system. The Markovian transition probability matrix of the degradation states which are defined for fix intervals between the inspection points in time. And, Markovian transition probability matrix have to carry out conversion to a discrete value from a continuation value of tunnel lining degradation. The applicability of the identification methodology is investigated by the inspection data set of tunnel linings in Hokkaido.

A multilevel information fusion approach for visual quality inspection

For visual quality inspection systems to be applicable in industrial settings, it is mandatory that they are highly flexible, robust and accurate. In order to improve these characteristics a multilevel information fusion approach is presented. A first fusion step at the feature-level enables the system to learn from an undefined number of potential defects which might be segmented from the images. This allows for the quality control operators to label the data at the image-level and the sub-image-level, and use this information during the learning process. Additionally, the operators are allowed to provide a confidence measure for their labelling. The additional information obtained from the increased flexibility of the operator inputs allows to build more accurate classifiers. A second fusion step at the decision-level combines the classifications of different classifiers, making the system more accurate and more robust with respect to the classification method chosen. The experimental results, using various artificial and real-world visual quality inspection data sets, show that each of these fusion approaches can significantly improve the classification accuracy. If both information fusion approaches are combined the accuracy increases even further, significantly outperforming each of the fusion approaches on their own.

Regulator Heterogeneity and Endogenous Efforts to Close the Information Asymmetry Gap

The now standard principal-agent model of regulator-firm interactions typically assumes the presence of a single regulator and an exogenously determined information asymmetry between the principal and the agent. In this paper we draw upon a unique data set of regulatory inspections conducted by the U.S. Food and Drug Administration (FDA) to explore the consistency of these assumptions with the actual practice of regulators. We find that the canonical assumptions of the agency paradigm are strained by, if not altogether inconsistent with, the key practical realities of regulation by the FDA. Our analysis uncovers several dimensions along which regulators actively and endogenously seek to close the information asymmetry gap. We also find considerable regulator heterogeneity, which in turn depends in part upon the specific training and experience of individual regulators.

The geographic distribution of environmental inspections

Models of the enforcement of environmental regulations regarding point source pollution suppose that the probability of inspection or audit is independent across facilities. However, there are a number of reasons why regulators may choose to inspect many sites in a particular geographic area at one time. If the probability a site is inspected also depends on its compliance behavior, the expected payoff from choosing to violate will depend upon the compliance decisions of neighboring sites, creating a game of strategic interdependence between firms. In this paper, we use a dataset of inspections at petroleum storage sites in Manitoba between 1981 and 1998 to consider to what extent inspections are spatially correlated and whether inspection probabilities are a function of the inspection and violation history of the site and its neighbors. Further, we examine to what extent firms take into account whether their neighbors have been previously found in violation in determining compliance.

Applying statistical methodology to monitor the piping conditions in petroleum process

In most chemical plants and refineries, the process piping in the process unit suffers from metal thinning, especially the carbon steel components such as elbow, tee, direct piping, branches, and reducers. Therefore, the inspectors should have accurate knowledge about the deterioration conditions of the piping system in order to maintain the personal safety and the structural integrity of the chemical plants and refineries. Basically, ultrasonic thickness spot measurement is an efficient way to check the thinning condition of the piping spools. However, the variations of the thickness readings always influence the evaluation of the corrosion rate of the piping spools. In this study, in order to waive the influence of the uncertainties of the thickness measurements, a statistical methodology is proposed to calculate the corrosion rate of each piping spool. With the proposed statistical approach, the corrosion rate of the piping spools in the process unit can be easily understood by plotting the histogram and using the upper limit of calculated corrosion rates. By utilizing the trend shown in the histogram, one can get knowledge about the deterioration condition of the piping system in the process unit. Finally, in order to discuss the feasibility of the proposed methodology, two sets of inspection data in a selected process unit from a lubricant plant were adopted as the examples. Also, a site survey result performed in the lubricant rerun unit is presented and shows the effectiveness of this statistical methodology.

Citizen Complaints, Regulatory Violations, and Their Implications for Swine Operations in Illinois

Using a unique Environmental Protection Agency (EPA) livestock inspection data set, this paper presents an extensive empirical investigation of relationships between citizen complaints, swine production and community characteristics, EPA inspections, and regulatory violations. Our results suggest that facility and community characteristics have a relationship with citizen complaints. Also, complaint-initiated inspections are more efficient than regularly scheduled ones in terms of regulatory violation detection rates. Additionally, EPA inspectors and inspections influence a facility's compliance behavior. Except for building type and swine inventory intensity, facility and community characteristics are not associated with a facility's probability of regulatory violation.

Bayesian network classifiers versus selective [formula omitted]-NN classifier

In this paper Bayesian network classifiers are compared to the k -nearest neighbor ( k -NN) classifier, which is based on a subset of features. This subset is established by means of sequential feature selection methods. Experimental results on classifying data of a surface inspection task and data sets from the UCI repository show that Bayesian network classifiers are competitive with selective k -NN classifiers concerning classification accuracy. The k -NN classifier performs well in the case where the number of samples for learning the parameters of the Bayesian network is small. Bayesian network classifiers outperform selective k -NN methods in terms of memory requirements and computational demands. This paper demonstrates the strength of Bayesian networks for classification.

Using machine learning for estimating the defect content after an inspection

We view the problem of estimating the defect content of a document after an inspection as a machine learning problem: The goal is to learn from empirical data the relationship between certain observable features of an inspection (such as the total number of different defects detected) and the number of defects actually contained in the document. We show that some features can carry significant nonlinear information about the defect content. Therefore, we use a nonlinear regression technique, neural networks, to solve the learning problem. To select the best among all neural networks trained on a given data set, one usually reserves part of the data set for later cross-validation; in contrast, we use a technique which leaves the full data set for training. This is an advantage when the data set is small. We validate our approach on a known empirical inspection data set. For that benchmark, our novel approach clearly outperforms both linear regression and the current standard methods in software engineering for estimating the defect content, such as capture-recapture. The validation also shows that our machine learning approach can be successful even when the empirical inspection data set is small.

Inspections, warnings, and compliance: the case of petroleum storage regulation

This paper examines the use of inspections and warnings to enforce environmental regulations. An empirical model of petroleum storage inventory reconciliation regulations is developed based on a state-dependent penalty system. I estimate a two-stage probit model on a dataset of inspections and warnings at petroleum storage sites in the province of Manitoba, Canada between 1983 and 1998. The results suggest that, even in the absence of frequent prosecutions, inspections deter future violations but that this effect is small.

GEOMETRICAL ANALYSIS OF SEGMENTED ULTRASONIC FLAW REGIONS

Abstract The automated or manual evaluation of ultrasonic NDT data involves the initial identification of suspect flaw regions from different probe inspections, and then the comparison and correlation of the physical parameters associated with each extracted region. This paper describes a general technique for deriving a series of geometrical attributes for a suspect indication which has been segmented from three-dimensional ultrasonic images. The attributes are position, orientation, approximate size, and surface texture. This approach, which may be applied to any form of imaged flaw indication, involves the calculation of a best-fit cuboid, or bounding-box”, for the contained data points. A number of real ultrasonic inspection data-sets are employed to illustrate the method proposed.

Nuclear reactor pressure vessel-specific flaw distribution development

Vessel intergrity predictions performed through fracture mechanics analysis of a pressurized thermal shock event have been shown to be significantly sensitive to the overall flaw distribution input. It has been shown that modern vessel in-service inspection (ISI) results can be used for development of vessel flaw distribution(s) that are more representative of U.S. vessels. This paper describes the development and application of a methodology to analyze ISI data for the purpose of flaw distribution determination. The resultant methodology considers detection reliability, flaw sizing accuracy and flaw detection threshold in its application. Application of the methodology was then demonstrated using four recently acquired U.S. PWR vessel inspection data sets. Throughout the program, new insight was obtained into several key inspection performance and vessel integrity prediction practice issues that will impact future vessel integrity evaluation. For example, the potential application of a vessel-specific flaw distribution now provides at least one method by which a vessel-specific reference flaw size applicable to pressure-temperature limit curves determination can be estimated. This paper will discuss also the development and application of the methodology and the impact to future vessel integrity analyses.