Administrators mainly design and supervise the fault diagnosis ontology model. The results, including fault causes, fault locations and maintenance measures, can be obtained through UI. The UI application program was developed with the Visual Studio software platform and C language under the Windows 7 system.
The ontology proposed in this paper is mainly used in the field of loaders fault diagnosis; thus, it can be classified as a highly specialized domain ontology. Currently, there are many kinds of approaches for constructing domain ontologies, i. Toshihiro Uchibayashi et al. This method is suitable for large-scale ontology domains and will be applied in this paper, when fault diagnosis ontology increases gradually.
Hu qingxi et al. Zhou Yong [ 50 ] constructed an improved ontology model based on machine learning BP neural network which changed the mode of multi strategy merging in ontology mapping. Chaware et al. However, this integrated method does not consider the collaborative and distributed construction of ontologies. The Seven-Step Method is among the top choices for building a domain ontology [ 22 ].
In addition, it not only enjoys detailed technical support and advantages in model creation, but also has advantages in the detailed modelling process [ 22 ]. Hence, this paper chose the Seven-Step Method, developed by the Medical Information Center of Stanford University [ 52 ], to construct the ontology model in the fault diagnosis domain, and it is shown in Figure 2. Next, the ontology modeling steps of the Seven-Step Method are described in detail.
The research described in this paper is in the field of loaders fault diagnosis. The keywords of loaders fault diagnosis are engine, gearbox, oil temperature, gear pump, main valve, etc. The components of the loaders can be divided into four levels, from top to bottom, which are the device level, system level, assembly level, and part level. When the loader fails, a single fault can be aroused by various fault causes rather than one fault cause. Faults in low level components affect not only faults at same level, but also faults in higher levels.
In order to handle the complexity of faults, the fault diagnosis ontology model has five defined classes: FaultMode , FaultEquipment , FaultMaintenance , Parameters and FaultPhenomenon , as shown in Figure 3. The FaultEquipment indicates the location of faults [ 43 ].
How can I improve my fault finding techniques?
The parameters class expresses running data collected by sensors. The FaultPhenomenon indicates the phenomena when a failure occurs.
- Submit documents to WikiLeaks.
- Solving Audio Hum Problems – Electronic Product Design?
- mesopotamia hanging vine shrines location.
- how to find htc device id.
- Navigation menu.
- Information.
The FaultMaintenance means that faults can be repaired by some measure. As shown in Table 1 , each object property has its corresponding domains and ranges.
If an object property has an inverse property, then the inverse property has inverse domains and ranges. Property 3 indicates the relationship between the fault components and their sub-components. Property 4 shows the relationship between the fault mode and fault effects. Property 5 and 6 are sub-properties of property 4, and property 5 indicates the relationship fault mode and its same level effects, and property 6 indicates the relationship fault mode and its higher level effects.
Property 7 indicates that fault locations provide information support for maintenance methods.
- little rock arkansas marriage records.
- florida waiting period for divorce.
- free public death records arizona.
- ip address searcher v2 12.
- obama fake birth certificate nigger forum.
- when people see the track marks on her arms.
- getting marriage license in louisiana.
Property 8 indicates that the fault mode is able to be repaired through using maintenance methods. As shown in Table 2 , each data property, similar to the object properties, also has its own domains and ranges. After the former six steps, the structure of the fault diagnosis ontology model is built, as shown in Figure 4. Instances of fault diagnosis ontology model are divided into two parts.
For the second part, system administrators and experts are responsible for instantiating other classes and properties according to the actual condition of the loaders. So far, the loaders fault diagnosis ontology model has been established after the instantiation. A pellet is used for checking exceptions, to verify the correctness of ontology model [ 22 ]. As introduced in the second section, the Jena inference engine was used to parsethe ontology model in this paper, and the SELECT mode of the Sparql was selected to query the properties of the ontology model.
CBR, proposed by Schank [ 53 ], can simulates human cognitive processes and integrates empirical knowledge of different fields into a unified format [ 54 ]. CBR can achieve effective and accurate fault diagnosis of loaders. This is because that CBR can obtain causes of failure, fault locations and maintenance methods, as long as cases are able to be matched successfully in the case library.
On the other hand, CBR can exploit historical cases which have actually happened and have been solved, to guarantee accuracy. The process of CBR based loaders fault diagnosis is shown in Figure 5. Loader type, fault phenomena and parameters are selected as the feature indexes for calculating case similarities in the case library. When the similarity value is greater than or equal to the set value, it indicates that the case-matching has been successful, and the fault diagnosis results of the matched case are displayed directly; Then case-updating is evaluated, and the case library will be updated if the evaluation is satisfactory, or else the CBR method will end.
When the similarity value is less than the set value, the CBR based fault diagnosis will fail and the diagnosis process will end. The fault diagnostic process is described in detail below. Although there are a large number of loader types, repair methods are often similar for similar faults of the same loader type; therefore, the loader type is selected as an essential feature index. The maintenance methods used in the same fault phenomena can also be used as guidance to the maintenance person; thus, the fault phenomena index is another alternative.
In addition, parameters obtained directly from the data acquisition device will change when faults occur, therefore, meeting the needs of the feature index. Hence, the three feature indexes are loader type, fault phenomena and operating parameters.
1st Edition
Case-retrieval refers to searching the fault diagnosis case library based on the above three feature indexes. Hence, the construction of the case library is of great significance. As shown in Figure 6 , experts select the classes and properties of fault diagnosis ontology manually, and the corresponding data is filed manually regarding historical cases and events, and then fault diagnosis cases are built and stored in the case library. At this point, the case library is constructed. Moreover, the established ontology model with its semantic description can be retrieved with cases that are also built semantically in the above process of constructing the case base.
Acronyms Finder
The established ontology model with its semantic description can be retrieved with cases which are also built semantically, as mentioned above. Hence, case matching can be realized by matching instances of hasNumber , FaultPhenomenon , Parameters corresponding to three feature indexes between ontology models and cases.
Similarity values are used to judge the similarity between to-be-diagnosed faults and the cases in the case library. The nearest neighbor algorithm [ 55 ] was used to calculate the similarity value in this paper, and its formula is shown in Equation 1. The value, w j , is directly defined by system experts to simplify the designing of this value. S i m C i j is determined as follows in this paper:.
If fault phenomena are identical to a case in the case base, then S i m C i 2 equals 1, or else S i m C i 2 equals 0.
Fundamentals Of Electronic Circuit Design Pdf
Operating parameters collected by sensors are calculated with Equation 2. When case-matching is achieved successfully, the actual fault diagnosis results can be saved into the case library. However, too many similar cases will waste the storage resource and decrease the efficiency of case-retrieval and case-matching. Therefore, it is necessary to develop case-updating strategy to eliminate redundancy in the case library. Derbinsky et al. By controlling suitable values of Mt , we can ensure that the case base does not become too narrowed down, oversized or skewed.
In this way, a suitable number of cases can be saved in the case library. Additionally, cases in the case base can also be improved by different experts at any time to improve the effectiveness of case management. Based on the fault diagnosis ontology model, RBR was used to process the loaders fault diagnosis when the CBR method failed due to the lack of proper cases. The standard SWRL language is used to express the fault diagnosis rules. Based on the analysis of classes and properties in the fault diagnosis ontology model, as shown in Table 3 , the constructed SWRL rules are represented in this paper, and these rules are the basis of the ontology-based RBR method.
To understand SWRL easily, two basic elements of syntax are introduced, as follows [ 57 ]:. In order to simplify the discussion here, we will only explain the following representative rules. Rule 1: If fault mode x occurs with the occurrence of fault phenomena y, then the direct fault locations of fault mode x are z.
Rule If the fault mode x happens at equipment y, and the fault locations provide the method for troubleshooting z, then failure maintenance z can repair fault mode x.