Machine Learning Artificial Intelligence Semantic Web Ontology Digital Transformation Web Technology Knowledge Information Processing Anomaly and Change Detection Workflow & Services Navigation of this blog
Fleet case, fault diagnosis system and ontology
Fleet Case is a system for companies that own multiple devices and products to efficiently and accurately diagnose failures and perform maintenance on those devices and products, and is constructed as an ontology-based knowledge base to systematically organize information on the structure, functions, and relationships among parts of the devices and products owned by the company. The information will be organized in a systematic manner. This fleet case has the following characteristics
- Capable of handling large amounts of data: Information on multiple devices and products can be integrated based on an ontology.
- Easy to diagnose failures: The use of an ontology-based knowledge base facilitates the identification of the causes of failures.
- Efficient maintenance: The ontology-based knowledge base enables efficient planning and procedures for maintenance work.
- Easy to reuse: Building an ontology-based knowledge base enables cost reduction through reuse and centralized system management.
- Analyzable: By integrating information on multiple devices and products, data for analysis and improvement can be obtained.
Fleet case is used in the manufacturing, energy, and transportation industries, and is expected to improve productivity and reduce costs by optimizing the equipment and products owned by companies and enabling efficient failure diagnosis and maintenance.
Failure diagnosis systems are designed to automatically detect product and equipment failures and identify the causes. The systems will be able to identify the causes of failures and identify parts that need to be repaired or replaced using machine learning, data mining, and other methods based on sensor data, log data, fault information, and other data. The aforementioned fleet case is a system for companies that own multiple devices and products to efficiently and accurately diagnose failures and perform maintenance on those devices and products, and can be considered a type of failure diagnosis system.
By applying ontology to these systems, it is possible to systematically organize information on the structure, functions, and relationship of parts of the products, which can be used as a knowledge base for failure diagnosis and maintenance, as well as to share failure diagnosis systems for each product line, thereby reducing costs through reuse and enabling centralized system management, This enables cost reduction through reuse and centralized system management.
Ontology-based failure diagnosis systems and fleet cases can also be reused externally. For example, if other companies in the same industry have similar products, sharing the ontology will enable efficient failure diagnosis and maintenance of the products, and will enable joint development with other companies and the establishment of new business models to provide product support.
When reusing ontology-based failure diagnosis systems and fleet cases, the following should be noted
- The scope of application of the ontology should be clarified.
- Industry standard ontologies should be used.
- Ontology updates and maintenance should be performed on an ongoing basis.
- Security measures should be fully considered.
By taking these precautions into account, it will be possible to reuse ontology-based failure diagnosis systems and fleet cases. In this section, we describe the ontology and the ontology of the failure diagnosis system and fleet case based on “Ontology Modeling in Physical Asset Management“.
Ontology-based fault diagnosis system, reuse of fleet cases
「Physical asset integrity is a life cycle concept that is based on system engineering development process in order to maintain equipment functionality in an acceptance level in terms of safety, environment, and cost. Maintenance plays a key role in such aim by improving system availability, performance efficiency, and product quality (Alsyouf 2007). Thinking maintenance as an added value process requires the evo- lution of maintenance strategies. From a “fail and fix,” new maintenance schema must evolve to “predict and prevent” approach. This new vision is supported by an evolution of condition based maintenance (CBM+) and Prognostics and Health Management (PHM) maintenance strategies (Iung et al. 2009). Nevertheless, despite this anticipative approach, failures still occur.
To minimize the effects of unexpected system failures, efficiency of fault diagno- sis has to be improved (Lei et al. 2008; Zio et al. 2008). When considering classical diagnosis techniques, unexpected situation are detected from a local point of view, i.e., the equipment level. However, when considering complex systems, classical techniques may not be useful since the whole system might not be monitored and the interaction between equipment and the environment results in variable performance. Hence, the maintainers are required to come with hypotheses about the causes of failures or the abnormal performance based on the symptoms’ occurrence (Moss et al. 2010). These situations require knowledge about the degradation mechanisms of components built on several technologies, of mechanical, electrical, electronic, or software natures (Verma et al. 2010). The performance can vary over the phases of their life cycle and usage condition (Bonissone and Varma 2005). Such a task may be difficult especially for junior maintainers and even for experts, since it relies on expert’s knowledge. Such knowledge is built first on the theoretical knowledge expert has learned, and second on his experience applying diagnosis on real cases. Due to human nature, Experts can fail at identifying the causes.
In order to improve the diagnosis process for large and complex systems such as power plants, ships, and aircrafts at the maintainer’s level, one possible approach is to take advantage of the “fleet” dimension. This dimension provides more knowl- edge and data to help maintainers (Monnin et al. 2011a).
A fleet shall be viewed as a set of systems, subsystems, and components. As we address the naval domain, in the following, a unit of a fleet will be considered as a system (e.g., ship), a subsystem (e.g., propulsion or electric power generation), or a component (e.g., diesel engine or a shaft) depending on the target object.
The individual knowledge of each unit is capitalized and reused in order to improve PHM activities such as predictive diagnosis, which will serve as example in the case study. To take advantage of the individual knowledge at the fleet level, a semantic model, using ontology, is proposed for the PHM activities in the naval domain. A specificity of the approach is that the ontology represents and structures the knowledge arising from the PHM processes as well as from the domain of inter- est, i.e., the naval domain. Usually, applications using ontology in the PHM domain only focus on knowledge of a specific PHM process and for a single component (i.e., knowledge for the diagnostic of an electric motor). This way several categories of concepts related to the PHM and system domains, called contexts, are modeled and relations between these concepts are explicit. Such a semantic model enables to reuse particular data, such as maintenance records, reliability analysis, failure anal- ysis, monitoring data analysis at a fleet level in order to provide more knowledge. As data become available, PHM activities benefit from more contextual information.」
5.1 Introduction
5.2 Context of Diagnosis of Complex System
5.3 PHM Vs. Fleet-Wide Approach
5.3.1 Fleet Integrated PHM Review
5.3.2 Predictive Diagnosis Using Fleet-Wide Knowledge
5.3.3 Sub-fleet Characterization
5.4 Ontology for Fleet-Wide Semantic Knowledge Modeling
5.4.1 Providing Semantic Through Ontology
5.4.2 Ontology-Based PHM Knowledge Modeling Rules
5.4.2.1 To Define the Key Concepts of the Domain
5.4.2.2 To Define a Class Hierarchy (Subsumption)
5.4.2.3 To Describe and Define Classes
5.4.2.4 To Define Properties of Classes
5.4.2.5 To Define the Value Type, the Cardinality and the Allowed Values of Classes
5.4.2.6 To Create Instances
5.4.3 Ontology for PHM and Marine Domains
5.4.3.1 Technical Context
5.4.3.2 Dysfunctional Context
5.4.3.3 Operational Context
5.4.3.4 Service Context
5.4.3.5 Application Context
5.4.3.6 Relations Between the Contexts
5.5 Application
5.5.1 Fleet-Wide Diagnosis Software
5.5.2 Case Study
5.6 Conclusions and Perspectives
Application of AI Technology to Fleet Cases
AI technology can bring a variety of benefits in its application to fleet cases. The following are some examples of how AI technology can be applied to fleet cases
- Automatic case classification and retrieval: AI techniques can be used to automatically classify fleet cases and enable retrieval of similar problems or cases. Natural language processing (NLP) and machine learning algorithms can be used to extract case content and keywords, calculate similarities, and automatically find relevant cases.
- Case-Based Reasoning: AI techniques can be used to automatically reason about solutions to fleet cases. It will be able to learn information about previous cases and their associated solutions and suggest the best solution for a new problem or case. Here, case-based reasoning is a type of machine learning, an empirical problem-solving technique.
- Automatically update and extend knowledge: AI techniques can be used to automatically collect information on new problems and cases, update and extend the knowledge of the fleet of cases, utilize web scraping and information extraction techniques to collect relevant information and add cases and solutions, and analyze user feedback and evaluations to improve the knowledge base. feedback and evaluations can be analyzed to improve the knowledge base.
- Leverage chatbots and virtual assistants: AI technology can be used to develop chatbots or virtual assistants that provide access to fleet cases and assistance in problem solving. These systems would use natural language processing and interactive systems to provide appropriate support when users report problems or search for solutions.
The application of AI technology will enable more efficient use of fleet cases and faster problem resolution. In addition, AI technology is constantly evolving and is expected to build systems with more advanced reasoning and self-learning capabilities.
コメント