Peter Matthews

Wind turbine pitch faults prognosis using a-priori knowledge-based ANFIS

The fast growing wind industry has shown a need for more sophisticated fault prognosis analysis in the critical and high value components of a wind turbine (WT). Current WT studies focus on improving their reliability and reducing the cost of energy, particularly when WTs are operated offshore. WT Supervisory Control and Data Acquisition (SCADA) systems contain alarms and signals that could provide an early indication of component fault and allow the operator to plan system repair prior to complete failure. Several research programmes have been made for that purpose; however, the resulting cost savings are limited because of the data complexity and relatively low number of failures that can be easily detected in early stages. A new fault prognosis procedure is proposed in this paper using a-priori knowledge-based Adaptive Neuro-Fuzzy Inference System (ANFIS). This has the aim to achieve automated detection of significant pitch faults, which are known to be significant failure modes. With the advantage of a-priori knowledge incorporation, the proposed system has improved ability to interpret the previously unseen conditions and thus fault diagnoses are improved. In order to construct the proposed system, the data of the 6 known WT pitch faults were used to train the system with a-priori knowledge incorporated. The effectiveness of the approach was demonstrated using three metrics: (1) the trained system was tested in a new wind farm containing 26 WTs to show its prognosis ability; (2) the first test result was compared to a general alarm approach; (3) a Confusion Matrix analysis was made to demonstrate the accuracy of the proposed approach. The result of this research has demonstrated that the proposed a-priori knowledge-based ANFIS (APK-ANFIS) approach has strong potential for WT pitch fault prognosis.

The introduction of a design heuristics extraction method

Abstract This paper introduces a novel method for analyzing conceptual design data. Given a database of previous designs, this method identifies relationships between design components within this database. Further, the method transforms these relationships into explicit design knowledge that can be used to generate a ‘heuristic-based’ model of the design domain for use at the conceptual stage. This can be viewed as a knowledge extracting method for the conceptual design stage. Such a method is particularly interesting, as the conceptual stage typically lacks explicit models to describe the trade-offs that must be made when designing. The method uses either principal components analysis or self-organizing maps to identify the relationships, and this paper describes all the elements required by the method to successfully extract and verify design knowledge from design databases.

Advanced product development integration architecture: an out-of-box solution to support distributed production networks

This paper presents novel collaboration methods implemented using a centralized client/server product development integration architecture, and a decentralized peer-to-peer network for smaller and larger companies using open source solutions. The product development integration architecture has been developed for the integration of disparate technologies and software systems for the benefit of collaborative work teams in design and manufacturing. This will facilitate the communication of early design and product development within a distributed and collaborative environment. The novelty of this work is the introduction of an ‘out-of-box’ concept which provides a standard framework and deploys this utilizing a proprietary state-of-the-art product lifecycle management system (PLM). The term ‘out-of-box’ means to modify the product development and business processes to suit the technologies rather than vice versa. The key business benefits of adopting such an approach are a rapidly reconfigurable network and minimal requirements for software customization to avoid systems instability.

Classification and Detection of Electrical Control System Faults Through Scada Data Analysis

The development of electrical control system faults leads to increased mechanical component degradation, severe reduction of asset performance, and a direct increase in annual maintenance costs. This paper presents a highly accurate data driven classification system for the diagnosis of electrical control system faults, in particular, wind turbine pitch faults. Early diagnosis of these faults can enable operators to move from traditional corrective or time based maintenance towards a predictive maintenance strategy, whilst simultaneously mitigating risks and requiring no further capital expenditure. Our approach provides transparent, human-readable rules for maintenance operators which have been validated by an independent domain expert. Data from 8 wind turbines was collected every 10 min over a period of 28 months with 10 attributes utilised to diagnose pitch faults. Three fault classes are identified, each represented by 6,000 instances in each of the testing and training sets. Of the turbines, 4 are used to train the system with a further 4 for validation. Repeated random sampling of the majority fault class was used to reduce computational overheads whilst retaining information content and balancing the training and validation sets. A classification accuracy of 85.50 % was achieved with 14 human readable rules generated via the RIPPER inductive rule learner. Of these, 11 were described as “useful and intuitive” by an independent domain-expert. An expert system was developed utilising the model along with domain knowledge, resulting in a pitch fault diagnostic accuracy of 87.05 % along with a 42.12 % reduction in pitch fault alarms.

Managing Conceptual Design Objects

The set of conceptual entities generated during the early stages of a design form an integral part of a product’s final definition (whether accepted or rejected) because they provide an implicit record of the rationale that led to the final configuration. Many of these alternative concepts are created from a non-spatial perspective and cannot be adequately captured using conventional CAD systems which still require a geometric framework. This paper describes a product data model that supports the capture and characterization of alternative concepts from both a product and process viewpoint by using non-geometric objects to structure a multilayered project entity. A simple management tool that uses this model is presented and demonstrated in the context of an undergraduate design project to build a semiautonomous guided vehicle.

Establishing agile supply networks through competence profiling

Concurrent engineering has been seen as the main means to respond to the increased complexity of engineering systems in sectors such as automotive, aerospace and defence. However, concurrent methods fail to respond to unpredicted events such as late customer requests or technological advances that dictate unexpected product or process alterations. Focusing on the organizational competence notion, the current paper introduces an agile supply network establishment process that enables evaluation of the impact of an event and suggests rapid restructuring of supply networks. The ideas described have been tested with over two hundred engineering small and medium-sized enterprises (SMEs) in the North East region of England proving to be of significant benefit to them. Participant companies are mainly engaged in the defence and aerospace industry.

Prognosis of wind turbine gearbox failures by utilising robust multivariate statistical techniques

In this paper we present a new methodology for the prognosis of a wind turbine gearbox. The statistically robust Mahalanobis distance was used to determine multivariate outliers within low frequency SCADA data without the need for manual labelling. Domain knowledge (meta-knowledge) was used to determine the multivariate vectors which encapsulate the condition of the wind turbine gearbox, providing a means to model anomalous gearbox behaviour whilst quantifying the severity of a monitored fault. A prognostic horizon of over 146 days was achieved using a new 3 degrees of freedom model, with a strong trend observed within the presented prognostic. This allowed for the quantification of fault severity, an estimation of the rate of fault development and also a means to quantify the quality and effectiveness of maintenance. In order to reduce noise inherent within SCADA data, an expert system was developed to transform the prognostic capability into actionable intelligence. This reduced the potential cognitive load placed upon the maintenance operator, whilst providing the knowledge required to optimise available maintenance resources. Due to the statistically robust nature of the approach, no gearbox fault data was required for training, enabling prognostic capability without the capital expense incurred through destructive testing. Furthermore, no additional capital expenditure is required due to data being collected from the pre-existing SCADA system available on all of the latest generation of wind turbines.

A methodology for quantitative estimates for the work and disturbance transformation matrices

Modern design projects are typically undertaken concurrently in a virtual enterprise network of expert design and manufacture agents. The general need for agile response in turbulent environments is well documented and has been analysed at the manufacture phase. This paper proposes a framework to enable the simulation and analysis of an agile design methodology. This framework models the occurrence of an unexpected local event in a concurrent design project and how it propagates to the global project. The redistribution of the design work can be controlled within the virtual enterprise and the total redistribution impact can be measured. A four-level classification scheme for the severity of unexpected events is proposed. A trial design experiment is conducted, and a first-order quantitative analysis is performed based on Work Transformation Matrices (WTM) and a novel Disturbance Transformation Matrix (DTM). A design negotiation process based on the WTM/DTM is proposed.

Linking design and manufacturing domains via web-based and enterprise integration technologies

The manufacturing industry faces many challenges such as reducing time-to-market and cutting costs. In order to meet these increasing demands, effective methods are need to support the early product development stages by bridging the gap of communicating early design ideas and the evaluation of manufacturing performance. This paper introduces methods of linking design and manufacturing domains using disparate technologies. The combined technologies include knowledge management supporting for product lifecycle management systems, Enterprise Resource Planning (ERP) systems, aggregate process planning systems, workflow management and data exchange formats. A case study has been used to demonstrate the use of these technologies, illustrated by adding manufacturing knowledge to generate alternative early process plan which are in turn used by an ERP system to obtain and optimise a rough-cut capacity plan.

Learning inexpensive parametric design models using an augmented genetic programming technique

Previous applications of genetic programming (GP) have been restricted to searching for algebraic approximations mapping the design parameters (e.g., geometrical parameters) to a single design objective (e.g., weight). In addition, these algebraic expressions tend to be highly complex. By adding a simple extension to the GP technique, a powerful design data analysis tool is developed. This paper significantly extends the analysis capabilities of GP by searching for multiple simple models within a single population by splitting the population into multiple islands according to the design variables used by individual members. Where members from different islands “cooperate,” simple design models can be extracted from this cooperation. This relatively simple extension to GP is shown to have powerful implications to extracting design models that can be readily interpreted and exploited by human designers. The full analysis method, GP heuristics extraction method, is described and illustrated by means of a design case study.