Hybrid Intelligent System On Coal Roadway Bolting Design

Bolting is one of the most important support in coal roadway. Many scholars and engineers worked hard on both theory and practice of bolting, accumulated with a lot of experience. In the actual production process, technicist will try used support scheme in most cases, which may be not the best chose in that case. That may lead to some lost in both economy and production schedule.Now, there are just some coal mines and research institutes know the advanced bolt technology, while most coal mines don’t know that technology, since the less of communication. With the development of computer technology and artificial intelligence technology, more and more research institutes try to design support scheme with the help of expert system, fuzzy theory, neural network and so on, and some systems were developed. But almost all the existing support design systems are based on single technology such as expert system and neural network, or just use a sample combination of these technologise. At the same time, due to the limitations of technology and the “bottleneck” to obtain knowledge, most existing systems meet several problems in actual use. Such as reasoning performance is unsatisfied, the required parameters are difficult to obtain and system is just useful to certain mine or minning area, and so on. These problems are disadvantaged to the promotiong of system. More and more scholars are interested in how to design support scheme more reasonable and effective with the help of artificial intelligent and computer technology.Supported by the National Natural Science Foundation--rapid excavation and support of large section roadway(51134025), this thesis studied on using HIS technology to help to design support scheme of coal roadway. According to the real situation of support design, a Hybrid Intelligent System on Coal Roadway Bolt Design(CRBD-HIS) was exploited, which contain a lot of technologies, such as fuzzy cluster analysis, fuzzy comprehensive evaluation, expert system and artificial neural network, grey relation analysis, orthogonal test and numerical simulation. The CRBD-HIS will help support design more scientifically.1. This thesis introduced and analyzed the status of HIS research. HIS combine some intelligent and nonintelligent technologies together and is good at solving complicated problem. HIS would improve the effect of data processing by using the advantages of different technologies. HIS choose appropriate technology to deal with different aspects of a complex problem based on their characteristics. Thus using HIS to help solving the complex problem like coal roadway bolt design is feasible. This thesis described HIS as formalized representation from the formal representation of system, and gave a concept of HIS. This thesis introduced the connection, common intilltellectual technologies in HIS and the mixture among them. According to the actual situation of coal roadway bolting design, some technologise were used while building HIS: CBR, RBR, ANN, ES, AHP, FLAC3 D, fuzzy comprehensive evaluation, fuzzy clustering analysis, grey correlation analysis and orthogonal experiment. A new model was builded through the study of CBR-RBR integrated reasoning mechanism, which included six parts, such as case classification, case retrieval, case reuse, case revision, case assessment and case study. Based on this new model, and mixed connection was chose to mix different technologies, the prototype system of CRBD-HIS was builded.2. This thesis study on the analysis model of coal roadway surrounding stability, and a new “global-part” model was proposed. Some traditional single index method, indexes method and modern mechematics and artificial intelligence classification method for engineering rock mass classification was introduced and discussed. Fuzzy mathematics was chose to analyze the surrounding rock stability of roadway, which was a typical kind of fuzziness question. Eight indexes were used to evaluate the surrounding rock stability of roadway. They were roof strength σt, coal strength σc, floor strength σb, depth H, initial caving interval of the immediate roof L, top height ratio N, pollar width X and the maximum horizontal principal σh. Advanced AHP was used to distribute weight for indexes. The sub model of stability classification of surrounding rock in coal roadway was estabed by fuzzy clustering analysis method. The clustering algorithm was programed by C# language. This sub model was used to do clustering analysis on 63 samples roadway from Huozhou mining area and obtain a reasonable result. The 63 samples roadway from Huozhou mining area was devided to 5 types and clustering center of each classification was obtained. The sub model of roadway surrounding rock stability fuzzy comprehensive evaluation was established, by applying the method of fuzzy comprehensive evaluation method. The evaluation object factor set of this sub model was U={σt,σc,σb,L,N,X,σh,H}, comment set was V={I(very stable),II(stable), III(marginally stable),IV(unstable),V(very unstable) }, and the clustering center was used as compare template. Then this thesis introduced the Coal Mine Roof Rating(CMRR), proposed by the U.S. Mining Bureau. According to the specific condition of coal mine in China, the CMRR was modified to be more suitable for using in China. In this way. CMRRC(CMRR in China) was builded and its feasibility was proved through an example.3. A new coal roadway bolt parameters reasoning model, A-O-F-G(ANN-Orthogonal Test-FLAC3D-Grey Relational Analysis) model was established. This model would help to study on how to obtain support parameters automaticly and optimal ingerence. Based on the improved BP neural network algorithm, a sub model for coal roadway support prediction was estabilished. There were 7 indexes as input parameters in this sub model: roof strength, coal strength, floor strength, depth, initial caving interval of the immediate roof, roadway weigth and height. Output parameters of this sub model were type, length, diameter and array space of bolt. Support cases were collected as training samples in the way of field research, questionnaire, literature review and so on. Orthogonal experiment, FLAC3 D and grey correlation analysis were used to study on bolt parameters optimization. An eight factors-three levels orthogonal experiment table was design, The eight factors were length, diameter, space and interval of roof bolt and rib bolt. Based on both the support scheme designed by neural network and orthogonal experiment table, 18 compared schemes were obtain. FLAC3 D was used to simulate the effect of these 18 schemes and chose 3 preliminary to optimize schemes, with roof, floor and rib moving as evaluation indexes. Addition to engineering difficulty and support cost, grey correlation analysis method was used to establish sub model of grey evaluation. This sub model can evaluate the above three preliminary optimization scheme again and get the most optimization scheme as reasoning result. In this way, bolt parameters can be generated automatically and optimized by CRBD-HIS.4. This thesis also study on establishment of FLAC3 D intelligent module, to help achieve parameterized modeling and automatic simulation for comparision scheme in A-O-F-G model. Through analysis the process and steps of FLAC3 D that deal with problem like roadway support design, the FLAC3 D intelligent module tried to rewrite command of various stages by FISH language, and built template file to represent the process of numerical simulation. To generate executable scripts through assignment for each template file, this shesis wrote caller programe by C# language. By caller program FLAC3 D software will run the executable script, to realize parametric modeling and simulation analysis automatic of FLAC3 D.5. In the last part of this thesis, Client/Server, Access, Windows Form and C# language are used to expoit the program of CRBD-HIS. The CRBD-HIS program could realize support scheme gengeration and optimization automatic, parametric modeling and simulation analysis automatic of FLAC3 D, case study and support report genetation. The CRBD-HIS program not only contained friendly man-machine interface and easy to use, but also included the “console” setting. Users from different mining areacan could modify the system according to their specific engineering and geological conditions, through the “console” setting. This helped CRBD-HIS to reason particularly base on the situation of each mining area, improved the reasoning precision and universal practicability of CRBD-HIS program. Through lab test, field test and literature review, a geological mechanical parameters database was builded, which included part of geological mechanics parameters of eastern mining area in our country, could provid a reference to field engineers. A case database of coal roadway support was also estabed, which reflected the main support design for coal roadway in the major coal-producing province eastern. There was a application example of CRBD-HIS program, for 2-5061 roadway of Sanjiaohe coal mine. Results show the feasibility and practicability of CRBD-HIS program. And two more application examples for Liyazhuang 6031 coal roadway and Ganhe 2-1121 coal roadway, obtained good results.Study result of this thesis is one result of the National Natural Science Foundation--rapid excavation and support of large section roadway(51134025). CRBD-HIS program could help technician to get optimal support scheme quickly and easily, and reduced the exploration and try of support scheme in the actual production. The research of this thesis shows that with the help of HIS to design support better is feasible, and this can reduce unreasonable support design. The study result of this thesis would help coal mine to product more scientificly and efficiently.