Research On Dispatching Method Of Open-pit Mine Production System Based On Random Process

Large open-pit mines typically need a great deal of equipment investments and consume lots of energy.It has been a key to improve the economic benefit how to reduce the energy consuption while ensuring equipments utilization.The mines always address the issue at three levels: the mine development planning,mining technique optimizarion and,production scheduling and manegement.From the perspective of production scheduling and manegement,the thesis tries to give a analysis and optimization on the equipment maintainance and scheduling problem in the open-pit mining.The objective is to increase the utilization time of equipments and reduce the production cost per unit.The Open-pit mining system involves a great materials transportation proess and typically needs to make online decision on the equipments scheduling based on the realtime production data.It is,actually,a hybird complex system,including decrete and continual processes and combining the qualitative and quantitative issues.To resolve the hybrid complexity,the stochastic process theory,grey theory,mathematical statistical analysis approach and hybrid integer programming model are all applied in the thesis.It investigates and quatitatively analysing a great deal of production data from the practical mining to discovers a series of stochastic features involved in the mine production system.It also addresses the scheduling problem on the main prodution equipments including electric shavels,trucks and so on.The main content includes:(1)This paper combined with the theoretical basis and practical application requirements of the open-pit mine production process,describes the background and research significance of choosing the topic of the dissertation,and summarizes the application status of open-pit mine production technology in China and the research methods at home and abroad.Since the production of open-pit mine in the future will be affected by many uncertain factors such as equipment condition,external obstacle,natural factors and market factors,it holds great uncertainty.The gray neural network integrated model is constructed for the randomness of production,which improves the prediction accuracy of the production material consumption model,and provides a basis for scientifically formulating production scheduling plans and achieving reasonable resource allocation.Markov chain model is widely used in the ineffectiveness stochastic process analysis with limited state.The stochastic process is extracted from the measured time series,and the Markov process of multivariate time-space series is used to analyze the output of open-pit mine,which not only reveals the macro rule of the process of output advancing with time,but also serves as the basis for clarifying the micro mechanism of output index development.In terms of the multiple index of stripping and mining yield,such as stripping amount,coal rate,distance,transportation distance and hoisting height,the growth rate fluctuations is divided into five kinds of states.According to the 23 months production of Anjialing coal mine,the state transition probability was calculated,so the production in the December 2017 was accurately forecasted through the principle of maximum probability.(2)Based on the analysis of the uncertainty in the open-pit mine production system,the time parameter of everyday fault duration in the process of production was performed with the time-series statistics,and then the BP-ARIMA combination model for the random sequence with unstable time was analyzed.The continuous 100 days fault duration was obtained by extracting the database,and the time sequence diagram of the fault duration shows the random sequence with unstable time.The mean and autocorrelation coefficient estimation of this model both pass the test of significance,and the model passes through the residual autocorrelation test.Because the equipment failure often presents nonlinear behavior,the advantage of the kernel principal component analysis in nonlinear feature extraction is applied to the pattern identification and analysis of equipment fault.The mechanical maintenance time,electrical maintenance time,alternative injection time,welding time,and outer barrier time constitute the fault time feature library.Through the kernel function the original space is implicitly mapped to the feature space where the linear relationship can be found,thus achieving the efficient solution to nonlinear problem.The simulation experiment results show that the model can reduce the complexity of the calculation,and possesses good generalization ability.It can realize the dimensionality reduction of the equipment failure features,accurately identify the equipment affected drastically by random factors,and effectively reduce the computational complexity.(3)The failure of key equipment such as electric shovel and truck will seriously affect the production of coal mine.In order to keep the equipment in good performance and ensure the completion of production tasks,it is necessary to monitor and analyze the performance status of the equipment,so as to adopt reasonable preventive maintenance strategy in advance.At present,the general idea about the monitoring and analysis of equipment reliability is to monitor the dynamic performance signal of equipment,and extract the key performance characteristic parameters after processing and analyzing the signal,thus identifying the equipment running status and further analyzing the equipment reliability.However,due to the complexity of the structure and operating environment of the equipment,the performance characteristic parameter that was detected in the use of the equipment,namely the observation sequence,cannot directly correspond to the state.The Hidden Markov Chain(HMC)model contains a kind of dual random process mechanism,which can link the observation sequence with the hidden state through a set of probability distribution,and describe the actual engineering situation more truly.This article identifies the implicit state transition process of the equipment with the discrete multiple observable sequences.The equipment running performance status was judged by the established state change model,and then the state change of the important equipment 730 E,a kind of mining truck,was predicted by means of example analysis.Through the eigenvalue extraction and scalar quantization to the training failure data signal of 730 E,the Hidden Markov model was established with the corresponding state numbers and observation numbers.Then,the model is used to calculate the similar probability between the undiagnosed fault data signal and the training data signal.The differences in similar probability can be used to judge the change of fault state,and achieve the goal of fault pattern classification.The monitoring,analysis and prediction of equipment running performance status can help coal mining enterprises find the potential faults of equipment in time,and make the reasonable maintenance plan,so as to improve equipment reliability.This work plays an important role in improving equipment utilization,reducing equipment maintenance costs,prolonging the service life of equipment,and ensuring the coal production plan.(4)A kind of truck scheduling problem,taking into consideration the impact of regular maintenance of equipment and vehicle on production,is extracted and studied from the perspective of actual production and scheduling practices in open-pit mines.In view of the capacity constraints of the actual transportation vehicles and the production sequence requirements of the loading point,a mixed integer programming model is established to optimize the robustness of equipment scheduling.The objective of the model includes two parts: the expected value of the actual total transportation value and the expected value of the vehicle scheduling caused by the temporary fault.In production scheduling,the actual total transportation value depends on shovel and truck capacity and effective operation time.By analyzing the relationship between production capacity and equipment capacity,an inequality on the upper bound of total transportation capacity is proposed.Based on the characteristics of the problem,a heuristic algorithm is designed to solve the scheduling problem.Data experiments show that the proposed algorithm can solve the problem in a feasible time.From the robustness and upper limit comparison of the solution,we can see that the solution quality is satisfactory.