The paper presents the conception of an intelligent system for troubleshooting an aircraft. The intelligent system requirements, its structure (architecture), and the domain conceptual model are described. In particular, the case-based and rule-based expert systems are defined as the main subsystems. The first one is designed to store information about malfunctions that are not accounted for in the current version of the documentation and find a solution by demonstrating similar problem situations. The second one provides the formation of plans to eliminate failures and malfunctions based on information from troubleshooting manuals. The features of their implementation are considered.
One of the factors affecting the time and quality of maintaining and troubleshooting an aircraft is the rapid and operational use of relevant documentation that, in turn, has a large volume and complex structure, and contains texts, diagrams, drawings, and diagrams. For example, about 8 tons of paper documentation is delivered with the A-320 aircraft. The maintenance manual can be a document of up to 50,000 pages. It is quite difficult to use and maintain up-to-dating such a mass of documents.
To solve the task of the rapid and operational use of the documentation the electronic document
management systems have been created, for instance, the AirNav Maintenance [
In this paper, we propose to create an intelligent decision support system for troubleshooting an aircraft, called the AirTech Assistant, designed for use by technical personnel engaged in the maintenance and repair of the Sukhoi Superjet (RRJ-95) aircraft.
The developed software will contain the necessary amount of documentation in electronic form
and provide support for decision-making when forming work plans. Case-based [
The AirTech Assistant will reduce the load on technical personnel during troubleshooting, as well as provide the collection (accumulation) and processing of information about equipment failures and malfunctions. In this paper, the conception of this software is considered in terms of the main functional requirements, the structure (architecture) and domain conceptual model, as well as the features of the implementation of the main subsystems.
The paper is organized as follows. Section 2 presents a state of art. Section 3 contains the description of the conception, including the requirements for the system functions, the structure of software, and the domain model, while Section 4 presents some concluding remarks.
The task of finding and troubleshooting failures and malfunctions of an aircraft is solved based on troubleshooting manuals, often presented either in printed form or in the form of electronic documents. Automation of its using and processing does not lose relevance, while various solutions are offered.
In particular, [
The considered and other solutions have some disadvantages: the lack of implementation details that ensure the reproduction of the proposals, including the models used; in some cases, general solutions are proposed without specifying the type of aircraft, while their adaptation is laborious.
The most famous foreign example of such systems is the AirNav Maintenance [
From our point of view, the following subsystems are the most interesting: • AMM: it contains information in the form of task cards necessary for maintenance, repair, replacement, adjustment, adjustment, inspection, and control of equipment and systems on the aircraft. These works (operations) are usually performed on the ramp or in the maintenance hangar. The information necessary for the maintenance of aircraft equipment is given by the supplier or manufacturer of the component. The AMM also contains information about the inspection and maintenance of aircraft structures. • TSM: is a specific module integrated into the A319/A320/A321, A330, A340, and A380 navigation subsystem. It is designed for finding and eliminating failures.
Further, these works will be considered as a basis for the development of the AirTech Assistant conception, designing, and software implementation.
The conception of our software includes a description of its main elements at the conceptual level, including purpose, requirements, structure, and basic concepts and relationships of the domain that will be used when creating data and knowledge bases. 3.1.
The AirTech Assistant (Figure 1) is designed to support decision-making by technical personnel during maintenance and repair of aviation equipment, in particular when searching and troubleshooting.
The purpose of this stage of the project is to create the first experimental version of the intelligent system.
The object of automation is the activity of personnel for the maintenance and repair of aircraft. The subject of automation is an algorithm for decision support in searching and troubleshooting the aircraft power supply system.
Initial data: 1. OBMS information about malfunctions; 2. Information about new failures and malfunctions that are not accounted for in the current version of the documentation, and their statistical indicators;
3. Information about malfunctions from the troubleshooting manuals. The AirTech Assistant should provide the following main functions: 1. Managing (entering, editing, storing) information about aircraft systems, technical operation, malfunctions (failures), and troubleshooting based on the aircraft documentation package; 2. Search for information about failures and malfunctions based on OBMS information to form a list of possible failed systems; 3. Accumulation (entering, editing, storage) of information about new failures and malfunctions that are not accounted for by the current version of the documentation (supporting of the case base); 4. Formation of a work plan for the search, confirmation, and troubleshooting failures and malfunctions based on the hybrid information contained in the current version of the documentation and the case base; 5. Maintenance of the repair and maintenance process based on a domain-specific interface. 3.3.
To implement the AirTech Assistant functions, the architecture, which contains the following main modules, is proposed: 1. Storage of documentation (manuals) in electronic form (in the form of PDF files or HTML pages containing information from operating manuals, maintenance manuals, troubleshooting) and information about an aircraft; 2. A case-based expert system for accumulating, searching, and storing information about failures and malfunctions, including: a. a database (or a case-based knowledge base) with information about failures and malfunctions; b. a search subsystem (solver) for cases retrieval; c. a subsystem for the acquisition (accumulation) of knowledge, which provides an extension of the case base. 3. A rule-based expert system for the defining systems who are the candidate for failure and the formation of plans of work on search, verify and troubleshoot failures, including: a. a knowledge base with information about the failures and task cards are required to maintain, repair, replacement, setup, adjustment, inspection, and control of equipment and systems on the aircraft; b. a search engine (solver) for logical inference; c. a query subsystem; d. a knowledge acquisition subsystem that provides the expansion of the rule base through visual programming and formalisms of event trees, decision tables, and state transition diagrams describing work plans for finding, confirming, and troubleshooting failures and malfunctions. 4. The subsystem for generating reports.
5. The domain-specific user interface.
When implementing the case-based expert system, our experience of creating such systems for
solving diagnostic problems in the petrochemical industry [
Let's look at these issues in more detail.
A case model. A case is a structured representation of the accumulated experience in the form of
data and knowledge, providing its subsequent automated processing with the aid of specialized
software [
From the formal point of view most decision-making tasks can be described by a set of
characteristics (properties), and can be formalized as follows [
MTask_CBR : ProblemCBR → DecisionCBR, where MTask_CBR is a task model in terms of case-based reasoning; ProblemCBR is a task (problem) description, DecisionCBR is a decision of a problem, while:
ProblemCBR = <c*, C>, C = {c1 ... cK}, c*∉ C, where c* is a new case, C is a case base.
DecisionCBR={d1 ,..., dR}, di=(ci,si),ci∈C, s ∈[0;1] where DecisionCBR is a task decision in the form of a set of retrieved cases with similarities si.
ci = {PropiProblem, PropiDecision }, where PropiProblem is an identifying (characterizing) part of a case, PropiDecision is a learning part of a case. In addition, each of these parts contains task properties and the composition of the parts has a problem-specific character:
PropiProblem= {p1,…, pm}, PropiDecision = {pm+1,…, pN }, PropiProblem,UPropiDecision = Prop, PropiProblem,∩PropiDecisio = ∅ .
Thus, when defining the task properties filling a case model, it is necessary to clarify the concepts of the domain that make up these components. For this purpose, we designed the conceptual model (Figure 2).
Case retrieval. Many methods can be used for case retrieval [
N dG (x, y) = i∑=1wihG (xi , yi ) N 1, if xi − yi < x for quantitative 0, otherwise hG (xi , yi ) = 1, xi = yi for qualitative 0, xi ≠ yi , where wi is the information weight, and x is the constraint on the difference between the values of properties. At the same time, the normalization (or standardization) is as follows: . xik → xik − min xik max xik − min xik
k k k
Case reuse. This issue is the most difficult from the point of view of computer-aided processing,
since it requires the involvement of experts for a meaningful interpretation of the solutions obtained
[
In our case, we propose to use the so-called "zero adaptation", based on copying part of the solutions.
Our experience of creating rule-based expert systems [
The structure of rules will be defined based on the domain model (Figure 2), and the prototypes
presented in [
To support the input/acquisition of rules we propose to use the following techniques:
• Describing the rules in the form of decision tables of a specialized type [
The paper considers the conception of an intelligent system for troubleshooting an aircraft, namely the AirTech Assistant. The requirements for its functionality, the structure, and the domain conceptual model are described. Case-based and rule-based expert systems are defined as the AirTech Assistant main subsystems. The features of the implementation are considered.
The AirTech Assistant will be a client-server web application and the following technologies and
means will be used for its realization: the Yii Framework, PHP, MySQL. The prototypes of
knowledge bases will be designed with the aid of the Personal Knowledge Base Designer [
The present study was supported by the Ministry of Education and Science of the Russian Federation (Project no. 121030500071-2 "Methods and technologies of a cloud-based service-oriented platform for collecting, storing and processing large volumes of multi-format interdisciplinary data and knowledge based upon the use of artificial intelligence, model-driven approach and machine learning"). Results are achieved using the Centre of collective usage «Integrated information network of Irkutsk scientific educational complex». 6. References