【作者】 张琦；

【导师】 李波；

【作者基本信息】 中国地质大学 ， 机械工程， 2013， 硕士

【摘要】 海洋是重要的能源基地,资源种类繁多,储量较大,世界上已经完成和正在进行的大洋钻探计划包括“深海钻探计划”(DSDP)、"大洋钻探计划”(ODP)和“综合大洋钻探计划”(IODP)等。随着海洋能源勘探开发技术的发展,深水油气勘探开发技术、天然气水合物勘探开发技术成为研发热点,深海矿产资源勘查技术向着大深度、近海底、多功能和原位方向发展。我国海洋勘探尚处于初级阶段,现有的海底勘探取样技术方法比较单一,取样深度浅,样品质量差,设备无深孔取样能力。依托于中央高校基本科研业务费专项资金“海洋勘探设备与仪器”创新团队项目(CUG120107),基于研制的实验室模型样机,提出了构建海底锥探/钻探一体化钻机运动控制/数据采集综合控制系统。通过构建此系统验证海底钻探平台测控系统关键算法的可行性,寻找远程化、自动化、智能化钻进的实现方法。运动控制部分主要完成了对动力头、给进系统、夹持器、机械手、绞车和液压系统的控制。钻进系统和钻具库系统是最关键的部件,本文针对这两个部分的控制做了详细的研究：1.钻进过程主要依靠动力头回转和给进系统给进破碎岩石,模型样机动力头使用交流伺服电机驱动,给进使用液压油缸驱动。动力头回转的负载存在随机性和时变性,普通PID控制很难达到理想的控制要求,采用可在线调整参数的Fuzzy-PID控制算法,加入了人工可控的PID参数调整规则,有效缓解了未知参数、非线性参数、滞后和随机干扰等问题对动力头回转控制系统的影响,在PID控制精度较高的基础上,增加了模糊控制灵活和适应性强的特点。基于钻柱动力学理论模型构建了钻柱固有频率计算分析系统,计算在连续回转钻进过程中需要尽量避免的转速数值；2.钻具库控制系统主要控制对象为机械手。机械手横向运动使用交流伺服电机驱动、齿轮齿条传动,采用Fuzzy-PID控制算法能够较好应对机械手横向运动载荷的非线性问题。机械手抓取动作使用液压油缸驱动。数据采集部分实现了钻进工况数据、冲洗液系统工作参数和下放触底工况状态等的采集监测,并对关键工况参数是否处于安全范围内进行监控。使用所采集到的工况数据对钻探过程中卡钻的风险进行预估和判断,目的在于减小远程钻探的风险和成本,提高钻探取样的效率,保护设备,减小设备损耗,同时保护样品的安全。现有的陆地钻机的卡钻评估主要有扭矩拉力模型,多元统计法,专家系统等几种常见方法。与此同时,神经网络分析法正在被越来越多地运用在这一领域。神经网络分析算法适用性强大,输入和输出参量采用矩阵形式定义,可以方便修改输入输出参量的数量、种类,借助于matlab神经网络工具箱,计算效率高。由于海底钻机平台工况复杂,卡钻风险评估理论模型并不成熟,分析卡钻风险需要用到的的参考变量比较多,在研发过程中需要根据实验情况进行修改验证。在这样的需求下,选择前馈(BP)神经网络算法作为海底钻进平台卡钻风险评估系统的核心算法,可以比较理想地完成海底钻机平台钻进过程中卡钻风险的评估任务。根据海洋环境条件和装备自身参数特点,以研制的实验室模型样机为实际对象,根据系统组成及功能需求,构建了以NI工控机、PCIe6363数据采集卡、NI7350运动控制卡和其他调理电路组成的硬件系统,编写了基于LabVIEW平台的海底锥探/钻探一体化钻机运动控制/数据采集综合控制系统,完成了各功能模块的基本动作控制任务,具备了实现自动化和远程控制钻进和锥探的基本条件,为后续进一步完善系统功能打下了基础。本文的主要创新点有：1.使用电液结合、伺服控制的方式对钻探工程机械进行控制,将模糊控制引入到钻探工程机械的控制中,针对电机驱动动力头的自适应PID控制做了尝试；2.针对海底钻进平台特殊工况下多参量的特点,选择BP神经网络算法作为海底钻探平台卡钻风险评估的核心算法,初步研究了其理论建模和应用方法,完成评估系统的初步构建。基于模型样机对运动控制/数据采集综合控制系统进行了试验验证,试验结果表明,各模块的运动控制程序均能对相应的执行件进行控制,精度达到了使用要求,数据采集系统能够较为理想地采集各传感器的输出量,卡钻风险预测系统能够根据工况数据计算卡钻的风险,钻柱固有频率计算系统能够计算出不同状态下钻柱固有频率对应的转速。系统各模块均工作正常,基本达到系统的预定目标。在后续的工作中,主要需要从以下几个方面对系统进行改进：1.对于动力头回转的模糊PID控制系统来说,完善规则表和各个参量的论域定义需要大量的实验和经验数据作为支撑,从而提高系统控制的可靠性和准确度；2.在卡钻风险进行评估系统中,BP神经网络模型的训练数据决定系统的准确度,训练数据量需要增大；同时,考虑到海底钻探平台在海洋环境中工作,需要在卡钻评估系统中引入海底地形、海况等输入参数；3.控制系统完成对各模块动作的控制,但需要人工操作来完成钻进流程,将各个模块集成应用,设计流程化、标准化和智能化的控制系统是完善本系统的一个需要做的工作。更多还原

【Abstract】 As an important energy base, the ocean contains a wide variety of resources, Marine resources survey rely on the development of marine geological resources exploration equipment. The completed and ongoing Ocean Drilling Programs include "Deep Sea Drilling Project (DSDP), the Ocean Drilling Program (ODP) and Integrated Ocean Drilling Program (IODP). Relying on the Central Universities Fundamental Research special funds innovation team project Marine Exploration equipment and apparatus (CUG120107), the motion control/data acquisition integrated control system of submarine cone Exploration/Downhole drilling rig laboratory model prototype was built, in order to meet the requirements of the submarine cone Exploration/Downhole drilling. motion control systems of key components and data acquisition system of main operating parameters was built in order to verify the feasibility of the key algorithm for remote, automated, intelligent drilling.Control of the power head, feed system, the gripper, robots, winch and the hydraulic system is completed. As a kind of rig drilling system, the key components include the drilling system and library system.1. Drilling process mainly rely on the rotation of power head and feeding of feeding system to broke rock. AC servo motor was chosen as the rotation power. The feeding system was driven by hydraulic power. There is randomness and time-varying load during rotation, ordinary PID control is difficult to achieve the desired control requirements. The Fuzzy-PID control algorithm, with which parameters can be adjusted online, was chosen as the core control algorithm for the power head. The Fuzzy-PID control algorithm can effectively alleviate the unknown parameters and nonlinear parameters, as well as hysteresis and random interference problem of power control system. It combines the characteristics of flexible, adaptable and high precision by the combination of Fuzzy control and PID control.2. The main control object of the library system is the carrying manipulator. It is driven by AC servo motor using gear rack transmission on transverse direction. The grape motion is driven by hydraulic power. The Fuzzy-PID control algorithm is used in the transverse controlling to solve the nonlinear problems.As a remote control drilling platform, real-time working condition monitoring is helpful to understand the working state of the drilling platform. Firstly, the system needs to monitor whether the key parameters in the safe range. On the other hand the system needs to forecast and judgment the risk of sticking in the drilling process. It can greatly reduce the overall risk and cost of drilling, so as to improve the efficiency of the drilling sampling, at the same time protect equipment, reduce equipment loss and protect the safety of the sample. Due to the complexity of working condition of ocean drilling platform, the BackPropagation neural network (BP) algorithm is used in the sticking risk forecast system. The flexibility and accuracy of the BackPropagation neural network algorithm can meet the demand of research in evaluation system. The input parameters and output parameters can be defined in a matrix form and can be modified easily. The efficiency is very high with the help of matlab neural network toolbox.With the help of NI industrial control platform, the motion control card7350and data acquisition card6363, and the LabVIEW environment, the cone/drilling integration laboratory model prototype motion control and data acquisition integrated system is built. Basic motion control task of each function module is completed. In this system, fuzzy control is introduced into the control of drilling engineering, to realize adaptive PID control of motor driven power head for complex load. The preliminary sticking risk evaluation system is built on the base of BP neural network data model, to meet the evaluation need of complicated working condition for offshore drilling platform. The innovation points include:1. The fuzzy control is introduced into the control of drilling engineering machinery, adaptive PID control for electrical motor driven power head.2. Considering the special working condition the seafoor drill platform, the BP neural network algorithm is chosen as the core algorithm of the sticking risk assessment system. The theory of modeling and the application method is developed.Experiments show the intergrated system can work properly.Following points are needed in the follow-up work:1. In fuzzy PID control system of power head rotation, the rule table and the parameter domain definition requires a lot of experiments and empirical data as support, so as to improve the reliability and accuracy of system control.2. In sticking risk assessment system, the quantity of the training data for BP neural network model is insufficient. At the same time, input parameters such as sea state and submarine topography should be taken into consideration.3. The control module should be integrated to realize a standardization and intelligent drilling control system.更多还原