技术领域technical field
本发明涉及油田监测技术领域，是一种油井故障智能化分析决策系统及诊断、预测、决策方案使用方法，即油井故障智能化分析决策系统及方法。The invention relates to the technical field of oil field monitoring, and relates to an intelligent analysis and decision-making system for oil well failures and methods for using diagnosis, prediction and decision-making schemes, that is, the intelligent analysis and decision-making system and method for oil well failures.
背景技术Background technique
目前，在油田日常生产管理中大多需要人工巡井，主要依靠值班工人定时、定点去轮巡，靠眼睛、耳朵的感官去发现运行设备的异常。巡视工作量与路线长度、地理环境和气候的变化都有很大的关系。当设备因故障在两次巡视之间停止运行，值班人员就无法及时发现，造成了生产的不安全因素增加。At present, most of the daily production management in oilfields requires manual well patrolling, mainly relying on the workers on duty to patrol regularly and at fixed points, and relying on the senses of eyes and ears to detect abnormalities in operating equipment. The inspection workload has a lot to do with the length of the route, changes in the geographical environment and climate. When the equipment stops running between two inspections due to failure, the personnel on duty will not be able to find out in time, resulting in an increase in unsafe factors in production.
现在对抽油井进行工况分析时，人工不能直接监测到地下杆式泵的工作状况，对于能反映井下抽油泵运行状况的示功图一个月测试一至两次，不能实时看到抽油机井的示功图、电流图，传统的示功图测试技术已经不能满足工作人员的需求。在采集到设备信息数据后，还需现场技术人员花费时间进行故障分析判断，判断出故障之后还需花费大量时间和精力去寻求解决方案，增加了人力、物力、财力。同时，在生产设备管理上，故障预测等方面，各大油田还没有相应的软件系统，对设备进行统一的分类管理，准确查找到每一个设备的型号、生产厂家、性能参数、安装厂家、检修周期等信息，对设备的维护及安装提供参考信息，实现对设备故障进行预测，提前预知抽油机故障，工作人员及时采取相应措施，避免不必要的损失。When analyzing the working conditions of the pumping wells, the working conditions of the underground rod pumps cannot be directly monitored manually. The dynamometer diagrams that can reflect the operating conditions of the downhole pumps are tested once or twice a month, and the status of the pumping wells cannot be seen in real time. Dynamometer diagram, current diagram, and traditional dynamometer diagram testing technology can no longer meet the needs of the staff. After the equipment information data is collected, on-site technicians still need to spend time on fault analysis and judgment. After the fault is judged, it takes a lot of time and energy to find a solution, which increases manpower, material resources, and financial resources. At the same time, in terms of production equipment management, failure prediction, etc., major oil fields do not yet have corresponding software systems to manage equipment in a unified manner, and accurately find the model, manufacturer, performance parameters, installation manufacturer, and maintenance of each equipment. Cycle and other information, provide reference information for equipment maintenance and installation, realize equipment failure prediction, predict pumping unit failure in advance, and staff take corresponding measures in time to avoid unnecessary losses.
发明内容Contents of the invention
本发明提供了一种油井故障智能化分析决策系统及诊断、预测、决策方案使用的方法，克服了上述现有技术之不足，其能有效解决现有技术中存在的不能对油井井下故障实时监测，存在安全隐患以及出现故障后寻找解决方案费时费力，造成工作效率低的问题。The present invention provides an intelligent analysis and decision-making system for oil well faults and a method for diagnosing, predicting, and using decision-making schemes, which overcomes the above-mentioned deficiencies in the prior art, and can effectively solve the problem in the prior art that it cannot monitor oil well downhole faults in real time , there are potential safety hazards and it takes time and effort to find a solution after a fault occurs, resulting in low work efficiency.
本发明的技术方案之一是通过以下措施来实现的：一种油井故障智能化分析决策系统，包括数据采集模块、远程测控模块、服务器、数据库连接模块和功能模块，数据采集模块包括电参数采集模块、温度传感器、压力传感器、动液面测量仪、一体化示功仪、流量传感器和可燃气体检测仪，所述电参数采集模块、温度传感器、压力传感器、动液面测量仪、一体化示功仪、流量传感器和可燃气体检测仪均与远程测控模块通信连接，所述服务器内设有生产数据库和专家数据库，远程测控模块与生产数据库通信连接，生产数据库与数据采集模块单向通信连接，专家数据库与数据库连接模块双向通信连接，所述功能模块包括登录模块、运行监控模块、状态预警模块、故障分析模块、故障预测模块、决策方案模块和设备信息模块，登录模块、运行监控模块、状态预警模块、故障分析模块、故障预测模块、决策方案模块和设备信息模块均与数据库连接模块双向电连接。One of the technical solutions of the present invention is achieved through the following measures: an intelligent analysis and decision-making system for oil well faults, including a data acquisition module, a remote measurement and control module, a server, a database connection module and a function module, and the data acquisition module includes electrical parameter acquisition Module, temperature sensor, pressure sensor, dynamic liquid level measuring instrument, integrated dynamometer, flow sensor and combustible gas detector, the electrical parameter acquisition module, temperature sensor, pressure sensor, dynamic liquid level measuring instrument, integrated display The dynamometer, flow sensor and combustible gas detector are all connected to the remote measurement and control module. The server is equipped with a production database and an expert database. The remote measurement and control module is connected to the production database. The expert database is connected with the database connection module in two-way communication, and the functional modules include a login module, an operation monitoring module, a state early warning module, a failure analysis module, a failure prediction module, a decision-making scheme module and an equipment information module. The login module, the operation monitoring module, the status The early warning module, failure analysis module, failure prediction module, decision-making scheme module and equipment information module are all bidirectionally electrically connected with the database connection module.
下面是对上述发明技术方案的进一步优化或/和改进：Below is the further optimization or/and improvement to above-mentioned technical scheme of the invention:
上述所述专家数据库可包括系统库、数据监测库、用户信息库、设备信息库、专家知识库、案例库和模板库；所述登录模块设有登陆管理单元和数据库连接管理单元，所述运行监控模块设有监控生产数据库单元和更新专家数据库单元，所述状态预警模块设有故障监测预警单元，所述故障分析模块设有故障查询系统单元、故障诊断系统单元和模板库管理单元，故障预测模块设有故障预测单元，所述决策方案模块设有故障管理系统单元和决策方案系统单元，所述设备信息模块设有设备信息管理单元和设备维护系统单元，监控生产数据库单元输出端与更新专家数据库单元输入端相连接，更新专家数据库单元第一输出端与故障监测预警单元输入端相连接，故障诊断系统单元输出端与故障监测预警单元输入端相连接，更新专家数据库单元第二输出端与故障诊断系统单元输入端相连接；所述登录管理单元与用户信息库相连接，数据库连接管理单元与系统库相连接，所述监控生产数据库单元、更新专家数据库单元、故障监测预警单元、故障查询系统单元、故障预测单元、故障管理系统单元均与数据监测库相连接，故障诊断系统单元、模板库管理单元均与模板库相连接，决策方案系统单元分别与专家知识库和案例库相连接，设备信息管理单元和设备维护系统单元均与设备信息库相连接。The above-mentioned expert database may include a system library, a data monitoring library, a user information library, an equipment information library, an expert knowledge library, a case library and a template library; the login module is provided with a login management unit and a database connection management unit, and the operation The monitoring module is provided with a monitoring production database unit and an update expert database unit, the state early warning module is provided with a fault monitoring and early warning unit, and the fault analysis module is provided with a fault query system unit, a fault diagnosis system unit and a template library management unit, and fault prediction The module is provided with a fault prediction unit, the decision-making scheme module is provided with a fault management system unit and a decision-making scheme system unit, and the equipment information module is provided with a equipment information management unit and an equipment maintenance system unit, monitoring the output of the production database unit and updating experts The input terminals of the database unit are connected, the first output terminal of the update expert database unit is connected with the input terminal of the fault monitoring and early warning unit, the output terminal of the fault diagnosis system unit is connected with the input terminal of the fault monitoring and early warning unit, and the second output terminal of the update expert database unit is connected with the input terminal of the fault monitoring and early warning unit. The input terminal of the fault diagnosis system unit is connected; the login management unit is connected with the user information database, the database connection management unit is connected with the system database, the monitoring production database unit, the update expert database unit, the fault monitoring and early warning unit, the fault query The system unit, fault prediction unit, and fault management system unit are all connected to the data monitoring database; the fault diagnosis system unit and the template database management unit are all connected to the template database; Both the equipment information management unit and the equipment maintenance system unit are connected with the equipment information database.
上述所述电参数采集模块、温度传感器、压力传感器、动液面测量仪、一体化示功仪分别采集现场的电机电流和功率参数、油温参数、油管压力参数和套管压力参数、动液面深度、抽油机载荷功图和角位移功图并通过无线通信传输至远程测控模块，流量传感器和可燃气体检测仪分别采集现场的油田产液量和可燃气体浓度并通过有线通信传输至远程测控模块；或/和，所述远程测控模块为RTU远程测控终端。The above-mentioned electrical parameter acquisition module, temperature sensor, pressure sensor, dynamic liquid level measuring instrument, and integrated dynamometer respectively collect on-site motor current and power parameters, oil temperature parameters, oil pipe pressure parameters and casing pressure parameters, and dynamic hydraulic pressure parameters. The surface depth, pumping unit load work diagram and angular displacement work diagram are transmitted to the remote measurement and control module through wireless communication. The flow sensor and combustible gas detector collect the on-site oilfield liquid production and combustible gas concentration respectively and transmit them to the remote control module through wired communication. A measurement and control module; or/and, the remote measurement and control module is an RTU remote measurement and control terminal.
本发明的技术方案之二是通过以下措施来实现的：一种油井故障智能化分析决策系统的故障诊断方法，包括以下步骤：The second technical solution of the present invention is achieved through the following measures: a fault diagnosis method for an intelligent analysis and decision-making system for oil well faults, comprising the following steps:
第一步，功图转化，数据采集模块将采集得到设备参数传输至运行监控模块，更新专家数据库后由故障诊断系统单元将地面测的示功图转换为井下有杆泵示功图，完成功图转换，步骤如下：The first step is the conversion of the dynamometer diagram. The data acquisition module will transmit the acquired equipment parameters to the operation monitoring module. After updating the expert database, the fault diagnosis system unit will convert the dynamometer diagram measured on the ground into the dynamometer diagram of the downhole rod pump, and the completion is successful. Graph conversion, the steps are as follows:
建立波动方程，公式如下：Establish the wave equation, the formula is as follows:
式中：u(x,t)为抽油杆柱在x深度处的横截面在t时刻的位移；x为深度；t为时间；a为应力在抽油杆柱上的传播速度；c为阻尼系数；In the formula: u(x,t) is the displacement of the cross section of the sucker rod string at the depth x at time t; x is the depth; t is the time; a is the propagation velocity of the stress on the sucker rod string; c is damping coefficient;
其上、下边界条件的傅里叶级数表示及初始条件公式如下：The Fourier series expression of the upper and lower boundary conditions and the initial condition formula are as follows:
其中：f(x,t)为柱塞在位移x、时刻t时的载荷；D(t)为悬点在t时刻的位移；U(t)为光杆在时刻t时的位移；σn、τn、vn、δn为傅里叶系数；Among them: f(x,t) is the load of the plunger at displacement x and time t; D(t) is the displacement of the suspension point at time t; U(t) is the displacement of polished rod at time t; σn , τn , vn , δn are Fourier coefficients;
通过分离变量及傅里叶变换可以求出u(x,t)：u(x,t) can be found by separating variables and Fourier transform:
根据胡克定律，可以求得According to Hooke's law, we can get
式中：In the formula:
On(x)＝(knchβnx+δnshβnx)sinαnx+(unchβnx+vnshβnx)cosαnx (5)On (x)＝(kn chβn x+δn shβn x)sinαn x+(un chβn x+vn shβn x)cosαn x (5)
Pn(x)＝(knchβnx+δnchβnx)cosαnx+(μnchβnx+vnshβnx)sinαnx (6)Pn (x)＝(kn chβn x+δn chβn x)cosαn x+(μn chβn x+vn shβn x)sinαn x (6)
根据波动方程的解u(x,t)，f(x,t)及其边界条件将井上获取的地面示功图转换为井下泵功图；According to the solution u(x,t) of the wave equation, f(x,t) and their boundary conditions, the surface dynamometer diagram obtained on the well is converted into a downhole pump power diagram;
第二步，对明显特征功图进行分类为需要神经网络分析和不需要神经网络分析，需要神经网络分析的功图进入第三步，不需要神经网络分析的功图进入第四步；The second step is to classify the obvious feature graphs into those that require neural network analysis and those that do not require neural network analysis. The graphs that require neural network analysis enter the third step, and the graphs that do not require neural network analysis enter the fourth step;
第三步，根据不同故障功图变化，需要神经网络分析的功图分别采用几何法、网格法和傅立叶描述子法对功图特征值进行提取，之后进入第五步；In the third step, according to the changes of different fault power diagrams, the geometric method, grid method and Fourier descriptor method are used to extract the characteristic values of the power diagrams that require neural network analysis, and then enter the fifth step;
第四步，根据不同故障功图变化，不需要神经网络分析的功图采用灰度矩阵对功图特征值进行提取，之后进入第六步；In the fourth step, according to the change of different fault power diagrams, the gray scale matrix is used to extract the characteristic values of the power diagrams for the power diagrams that do not require neural network analysis, and then enter the sixth step;
第五步，根据数据采集模块采集的生产数据结合第三步提取的功图特征值，调用单井模板库和区块模板库对故障进行诊断分析，之后进入第七步；In the fifth step, according to the production data collected by the data acquisition module and the characteristic value of the power map extracted in the third step, the single well template library and the block template library are called to diagnose and analyze the fault, and then enter the seventh step;
第六步，根据数据采集模块采集的生产数据结合第三步提取的功图特征值，运用欧式距离分类进行诊断分析，之后进入第七步；In the sixth step, according to the production data collected by the data acquisition module and the characteristic value of the power map extracted in the third step, the Euclidean distance classification is used for diagnosis and analysis, and then the seventh step is entered;
第七步，得出故障诊断结果，对故障诊断结果进行现场验证，若现场验证正确，则将相应故障诊断结果统计数自动加1次；若现场验证错误，则修正故障诊断结果并将该单井功图作为模板存入相应的单井模版库中。The seventh step is to obtain the fault diagnosis result, and carry out on-site verification of the fault diagnosis result. If the on-site verification is correct, the statistics of the corresponding fault diagnosis result will be automatically added once; if the on-site verification is wrong, the fault diagnosis result will be corrected and the list The well work diagram is stored as a template in the corresponding single well template library.
上述在第三步中，几何法特征值提取是基于泵功图进行几何操作获取特征值，对所述泵功图提取的几何特征值为:上冲程始点A、左下面积A1、左上面积A2、右上面积A3、右下面积A4、上行变形结束点B、下行变形结束点D、上死点C、曲线质心G、最小载荷Fmax和最小载荷Fmin。Above-mentioned in the 3rd step, geometrical method eigenvalue extraction is to carry out geometric operation and obtain eigenvalue based on pump work graph, the geometric eigenvalues to described pump work graph extraction are: upstroke starting point A, left lower area A1, left upper area A2, Upper right area A3, lower right area A4, upward deformation end point B, downward deformation end point D, top dead point C, curve centroid G, minimum load Fmax and minimum load Fmin.
上述在第三步中，所述的网格特征值提取法，对功图图像进行网格化处理，将示功图分解成若干个形状、大小相同的网格，对有示功图穿过的网格赋值“1”，对于其它网格，以功图穿过的网格为中心，当其位于功图外部时，每远离一格，灰度值减去1，当位于功图内部时，每远离一格，灰度值增加1，计算灰度矩阵的灰度均值、灰度方差、灰度偏差、灰度峰度、灰度能量、灰度熵六个特征参数，公式如下：In the third step above, the grid eigenvalue extraction method described above performs grid processing on the dynamometer image, decomposes the dynamometer diagram into several grids of the same shape and size, and passes through the dynamometer diagram. The value of the grid is "1". For other grids, the grid passing through the grid is the center. When it is outside the grid, the gray value will be subtracted by 1 for each grid away from it. When it is inside the grid , the gray value increases by 1 for every grid distance away, and calculates six characteristic parameters of the gray matrix, including gray mean, gray variance, gray deviation, gray kurtosis, gray energy, and gray entropy, the formula is as follows:
灰度均值：Grayscale mean:
灰度方差：Gray variance:
灰度偏差：Gray scale deviation:
灰度峰度：Grayscale kurtosis:
灰度能量：Gray energy:
灰度熵：Gray entropy:
以此作为特征值与标准特征值进行比较。Use this as an eigenvalue to compare with the standard eigenvalue.
上述在第三步中，傅立叶描述子法特征值提取是对功图进行傅立叶级数展开表征功图形状，即沿示功图轮廓线的一个动点P(l)的坐标变化是一个以功图周长为周期的周期函数，该函数可以用傅里叶级数表示，其系数a(k)称为傅里叶描述子，表达式如下：Above-mentioned in the third step, the Fourier descriptor method eigenvalue extraction is to carry out Fourier series expansion to the work diagram to characterize the shape of the work diagram, that is, the coordinate change of a moving point P(l) along the contour line of the dynamometer diagram is a The perimeter of the graph is a periodic function of the period. This function can be expressed by Fourier series, and its coefficient a(k) is called the Fourier descriptor. The expression is as follows:
进行归一化后公式：After normalization, the formula is:
式中是对傅里叶系数a(k)取模；由于变换后的高频系数较小，因此，通常只选取低频系数作为特征值，根据对功图的傅里叶变换可知，当k>20时，傅里叶描述子变得十分小，故选取k＝20。In the formula is to take the modulus of the Fourier coefficient a(k); since the transformed high-frequency coefficient is small, usually only the low-frequency coefficient is selected as the eigenvalue. According to the Fourier transform of the power diagram, when k>20 , the Fourier descriptor becomes very small, so k=20 is selected.
本发明的技术方案之三是通过以下措施来实现的：一种油井故障智能化分析决策系统的故障预测方法，包括以下步骤：The third technical solution of the present invention is achieved through the following measures: a fault prediction method for an intelligent analysis and decision-making system for oil well faults, comprising the following steps:
第一步，提取单井近期的5组正常示功图的特征值，之后进入第二步；The first step is to extract the eigenvalues of the recent 5 groups of normal dynamometer diagrams of a single well, and then enter the second step;
第二步，将单井近期5组正常示功图的特征值与模板库中正常示功图进行对比，计算相似度，之后进入第三步；The second step is to compare the eigenvalues of the recent 5 groups of normal dynamometer diagrams of a single well with the normal dynamometer diagrams in the template library, calculate the similarity, and then enter the third step;
第三步，建立数据的动态模型，即GM(h,n)模型，其微分方程是时间域的连续函数，h为方程阶数，n为变量个数，微分方程如下公式所示：The third step is to establish a dynamic model of the data, that is, the GM(h,n) model. Its differential equation is a continuous function in the time domain, h is the order of the equation, and n is the number of variables. The differential equation is shown in the following formula:
则微分方程的系数向量为：Then the coefficient vector of the differential equation for:
采用最小二乘法求解，结果为：Using the method of least squares to solve, the result is:
式中为A组成的分块矩阵，之后进入第四步；In the formula is the block matrix composed of A, and then enters the fourth step;
第四步，当发现特征值与正常示功图匹配的越来越低时，对下一组示功图数据进行预测，判断其是否会发生故障；若满足偏离界限，立即通过故障监测预警单元进行预警。In the fourth step, when it is found that the matching between the characteristic value and the normal dynamometer is getting lower and lower, predict the next set of dynamometer data to judge whether it will fail; if it meets the deviation limit, immediately pass the fault monitoring and early warning unit Forewarning.
上述在第三步中，GM(1，1)模型的建模原理为：In the third step above, the modeling principle of the GM(1,1) model is:
(1)设原始序列为X(0)，有：(1) Let the original sequence be X(0) , there are:
X(0)＝[x(0)(1),x(0)(2),...,x(0)(n)] (22)X(0) ＝[x(0) (1),x(0) (2),...,x(0) (n)](22)
对原始序列X(0)进行一次累加生成序列X(1)，得到：Perform an accumulation of the original sequence X(0) to generate a sequence X(1) to obtain:
X(1)＝[x(1)(1),x(1)(2),...,x(1)(n)] (23)X(1) ＝[x(1) (1),x(1) (2),...,x(1) (n)](23)
原始序列X(0)与一次累加生成序列X(1)中的对应数据之间的关系为：The relationship between the original sequence X(0) and the corresponding data in the accumulated sequence X(1) is:
(2)建立GM(1，1)模型的白化微分方程为：(2) Establish the whitening differential equation of the GM(1,1) model as:
GM(1，1)模型的灰微分方程为：The gray differential equation of the GM(1,1) model is:
x(0)(k)+az(1)(k)＝b k＝1,2,..,n (26)x(0) (k)+az(1) (k)=bk=1,2,..,n (26)
公式(26)中，a为发展系数，b为灰作用量，均为待辨识的参数，x(0)(k)为灰导数，z(1)(k)为背景值，背景值的计算公式如式(27)：In formula (26), a is the development coefficient, b is the gray action, both are parameters to be identified, x(0) (k) is the gray derivative, z(1) (k) is the background value, and the calculation of the background value The formula is as formula (27):
(3)设待求参数Y＝[x(0)(2),x(0)(3),...,x(0)(n)]T(3) Set the parameters to be requested Y＝[x(0) (2),x(0) (3),...,x(0) (n)]T
根据最小二乘原理，计算微分方程的解为：According to the principle of least squares, the solution of the differential equation is calculated as:
(4)令x(1)(0)＝x(0)(0)，则原始序列的预测公式为：(4) Let x(1) (0)=x(0) (0), then the prediction formula of the original sequence is:
式(30)中：为x(0)(k)的拟合值，为第k+1个预测值。In formula (30): is the fitted value of x(0) (k), is the k+1th predicted value.
本发明的技术方案之四是通过以下措施来实现的：一种油井故障智能化分析决策系统的决策方案使用方法，包括以下步骤：The fourth technical solution of the present invention is achieved through the following measures: a method for using a decision-making scheme of an intelligent analysis and decision-making system for oil well failures, comprising the following steps:
第一步，在进行故障诊断结果之后，根据故障类型进行问题描述，在案例库中进行案例检索，之后进入第二步；The first step is to describe the problem according to the fault type after the fault diagnosis result, and search the case in the case library, and then enter the second step;
第二步，通过决策方案模块查找该故障解决方案是否在数据库中，若数据库中有解决方案，则得到存在解决方案的结论之后进入第三步，如果未找到对应的解决方案，用户针对诊断结果分析后添加相应的解决方案，补充数据库，结束；The second step is to find out whether the fault solution is in the database through the decision-making solution module. If there is a solution in the database, it will enter the third step after the conclusion that there is a solution. If no corresponding solution is found, the user will check the diagnosis result After the analysis, add the corresponding solution, supplement the database, and end;
第三步，判断解决方案中是否存在满意的解决方案，若没有满意的解决方案用户自行设计适合的解决方案，并将解决结果以案例形式添加到案例库，结束；如寻找到满意的解决方案，之后进入第四步；The third step is to judge whether there is a satisfactory solution in the solution. If there is no satisfactory solution, the user can design a suitable solution by himself, and add the solution result to the case library in the form of a case, and end; if a satisfactory solution is found , and then enter the fourth step;
第四步，如果有满意的方案，根据用户的选择，将该解决方案采纳次数自动加1，在下次寻找方案时，优先显示采纳次数最多的方案，之后结束。Step 4: If there is a satisfactory solution, according to the user's choice, the number of adoptions of the solution will be automatically increased by 1. When looking for a solution next time, the solution with the most adoption times will be displayed first, and then it will end.
本发明通过建立油田智能化分析与决策专家系统，对油井的生产参数进行实时监控、实时分析、实时故障诊断、实时预测，同时还对油田的各类生产设备进行分类管理，从而真正意义上的实现了油田的数字化、智能化，对油田节约成本，节省人力物力。油井逐渐趋于数字化、智能化，采用准确、高效、实时的抽油机井工况自动监控技术，实现实时生产数据采集、生产环境实时监控、设备运行状态实时监控、实时故障诊断、业务单元预警分析、智能决策及自动生产解决方案，可以降低无功消耗及设备故障率，提高开井时效。The present invention implements real-time monitoring, real-time analysis, real-time fault diagnosis and real-time prediction on the production parameters of oil wells by establishing an oilfield intelligent analysis and decision-making expert system, and at the same time classifies and manages all kinds of production equipment in the oilfield, so that the real-time Realized the digitization and intelligence of the oil field, saved the cost of the oil field, and saved manpower and material resources. Oil wells are gradually becoming digitalized and intelligent. Accurate, efficient and real-time automatic monitoring technology of pumping well working conditions is adopted to realize real-time production data collection, real-time monitoring of production environment, real-time monitoring of equipment operation status, real-time fault diagnosis, and early warning analysis of business units , Intelligent decision-making and automatic production solutions can reduce reactive power consumption and equipment failure rate, and improve the timeliness of well opening.
附图说明Description of drawings
附图1为本发明实施例1的整体结构框图。Accompanying drawing 1 is the overall structural block diagram of embodiment 1 of the present invention.
附图2为本发明实施例1的数据库与功能模块的关系框图。Accompanying drawing 2 is the relation block diagram of the database and function module of embodiment 1 of the present invention.
附图3为本发明实施例2的故障诊断流程图。Accompanying drawing 3 is the fault diagnosis flowchart of embodiment 2 of the present invention.
附图4为本发明实施例2的功图转化示意图。Accompanying drawing 4 is the schematic diagram of the conversion of the work diagram of the embodiment 2 of the present invention.
附图5为本发明实施例2的示功图几何特征值提取示意图。Accompanying drawing 5 is the schematic diagram of extracting the geometric feature value of the dynamometer diagram according to Embodiment 2 of the present invention.
附图6为本发明实施例2的正常示功图的灰度矩阵示意图。Accompanying drawing 6 is a schematic diagram of the gray scale matrix of the normal dynamometer diagram of Embodiment 2 of the present invention.
附图7为本发明实施例2的BP神经网络训练结果图。Accompanying drawing 7 is the BP neural network training result figure of embodiment 2 of the present invention.
附图8为本发明实施例3油井故障预测方法流程图。Accompanying drawing 8 is the flow chart of the oil well fault prediction method in Embodiment 3 of the present invention.
附图9为本发明实施例3功图异常预测示意图。Accompanying drawing 9 is a schematic diagram of prediction of abnormality of power map in Embodiment 3 of the present invention.
附图10为本发明实施例4决策方案流程图。Accompanying drawing 10 is the flowchart of the decision-making scheme of Embodiment 4 of the present invention.
具体实施方式Detailed ways
本发明不受下述实施例的限制，可根据本发明的技术方案与实际情况来确定具体的实施方式。The present invention is not limited by the following examples, and specific implementation methods can be determined according to the technical solutions of the present invention and actual conditions.
下面结合实施例及附图对本发明作进一步描述：Below in conjunction with embodiment and accompanying drawing, the present invention will be further described:
实施例1：如附图1、2所示，油井故障智能化分析决策系统包括数据采集模块、远程测控模块、服务器、数据库连接模块和功能模块，数据采集模块包括电参数采集模块、温度传感器、压力传感器、动液面测量仪、一体化示功仪、流量传感器和可燃气体检测仪，所述电参数采集模块、温度传感器、压力传感器、动液面测量仪、一体化示功仪、流量传感器和可燃气体检测仪均与远程测控模块通信连接，所述服务器内设有生产数据库和专家数据库，远程测控模块与生产数据库通信连接，生产数据库与数据采集模块单向通信连接，专家数据库与数据库连接模块双向通信连接，所述功能模块包括登录模块、运行监控模块、状态预警模块、故障分析模块、故障预测模块、决策方案模块和设备信息模块，登录模块、运行监控模块、状态预警模块、故障分析模块、故障预测模块、决策方案模块和设备信息模块均与数据库连接模块双向电连接。Embodiment 1: As shown in accompanying drawings 1 and 2, the oil well failure intelligent analysis and decision-making system includes a data acquisition module, a remote measurement and control module, a server, a database connection module and a function module, and the data acquisition module includes an electrical parameter acquisition module, a temperature sensor, Pressure sensor, dynamic liquid level measuring instrument, integrated dynamometer, flow sensor and combustible gas detector, the electrical parameter acquisition module, temperature sensor, pressure sensor, dynamic liquid level measuring instrument, integrated dynamometer, flow sensor and the combustible gas detector are connected to the remote measurement and control module. The server is equipped with a production database and an expert database. The remote measurement and control module is connected to the production database. The production database is connected to the data acquisition module in one-way communication. Two-way communication connection of modules, the functional modules include login module, operation monitoring module, state early warning module, fault analysis module, fault prediction module, decision-making scheme module and equipment information module, login module, operation monitoring module, state early warning module, fault analysis The module, the fault prediction module, the decision-making scheme module and the equipment information module are all bidirectionally electrically connected with the database connection module.
这里的登录模块在运行时，首先会检测是否与服务器内的专家数据库和生产数据库连接，当连接成功时，用户才能登录，否则提示用户修改数据库连接；运行监控模块能实时显示油井的检测数据，包括：油井常规参数及近期数据变化信息，按照一定的时间间隔循环显示不同油井的设备数据信息，也可以根据个人需求查看单井信息。本发明通过建立油田智能化分析与决策专家系统，对油井的生产参数进行实时监控、实时分析、实时故障诊断、实时预测，同时还对油田的各类生产设备进行分类管理，从而真正意义上的实现了油田的数字化、智能化，对油田节约成本，节省人力物力。油井逐渐趋于数字化、智能化，采用准确、高效、实时的抽油机井工况自动监控技术，实现实时生产数据采集、生产环境实时监控、设备运行状态实时监控、实时故障诊断、业务单元预警分析、智能决策及自动生产解决方案，可以降低无功消耗及设备故障率，提高开井时效。When the login module here is running, it will first check whether it is connected to the expert database and production database in the server. When the connection is successful, the user can log in, otherwise the user will be prompted to modify the database connection; the operation monitoring module can display the detection data of the oil well in real time. Including: conventional parameters of oil wells and recent data change information, cyclic display of equipment data information of different oil wells according to a certain time interval, and single well information can also be viewed according to individual needs. The present invention implements real-time monitoring, real-time analysis, real-time fault diagnosis and real-time prediction on the production parameters of oil wells by establishing an oilfield intelligent analysis and decision-making expert system, and at the same time classifies and manages all kinds of production equipment in the oilfield, so that the real-time Realized the digitization and intelligence of the oil field, saved the cost of the oil field, and saved manpower and material resources. Oil wells are gradually becoming digitalized and intelligent. Accurate, efficient and real-time automatic monitoring technology of pumping well working conditions is adopted to realize real-time production data collection, real-time monitoring of production environment, real-time monitoring of equipment operation status, real-time fault diagnosis, and early warning analysis of business units , Intelligent decision-making and automatic production solutions can reduce reactive power consumption and equipment failure rate, and improve the timeliness of well opening.
可根据实际需要，对上述油井故障智能化分析决策系统作进一步优化或/和改进：According to actual needs, the above-mentioned oil well failure intelligent analysis and decision-making system can be further optimized or/and improved:
如附图1、2所示，所述专家数据库包括系统库、数据监测库、用户信息库、设备信息库、专家知识库、案例库和模板库；所述登录模块设有登陆管理单元和数据库连接管理单元，所述运行监控模块设有监控生产数据库单元和更新专家数据库单元，所述状态预警模块设有故障监测预警单元，所述故障分析模块设有故障查询系统单元、故障诊断系统单元和模板库管理单元，故障预测模块设有故障预测单元，所述决策方案模块设有故障管理系统单元和决策方案系统单元，所述设备信息模块设有设备信息管理单元和设备维护系统单元，监控生产数据库单元输出端与更新专家数据库单元输入端相连接，更新专家数据库单元第一输出端与故障监测预警单元输入端相连接，故障诊断系统单元输出端与故障监测预警单元输入端相连接，更新专家数据库单元第二输出端与故障诊断系统单元输入端相连接；所述登录管理单元与用户信息库相连接，数据库连接管理单元与系统库相连接，所述监控生产数据库单元、更新专家数据库单元、故障监测预警单元、故障查询系统单元、故障预测单元、故障管理系统单元均与数据监测库相连接，故障诊断系统单元、模板库管理单元均与模板库相连接，决策方案系统单元分别与专家知识库和案例库相连接，设备信息管理单元和设备维护系统单元均与设备信息库相连接。As shown in accompanying drawings 1 and 2, the expert database includes a system library, a data monitoring library, a user information library, an equipment information library, an expert knowledge library, a case library, and a template library; the login module is provided with a login management unit and a database Connect the management unit, the operation monitoring module is provided with a monitoring production database unit and an update expert database unit, the state early warning module is provided with a fault monitoring and early warning unit, and the fault analysis module is provided with a fault query system unit, a fault diagnosis system unit and template library management unit, the fault prediction module is provided with a fault prediction unit, the decision-making scheme module is provided with a fault management system unit and a decision-making scheme system unit, and the equipment information module is provided with a equipment information management unit and an equipment maintenance system unit to monitor production The output end of the database unit is connected to the input end of the update expert database unit, the first output end of the update expert database unit is connected to the input end of the fault monitoring and early warning unit, the output end of the fault diagnosis system unit is connected to the input end of the fault monitoring and early warning unit, and the update expert The second output end of the database unit is connected with the input end of the fault diagnosis system unit; the login management unit is connected with the user information database, the database connection management unit is connected with the system library, the monitoring production database unit, the update expert database unit, The fault monitoring and early warning unit, fault query system unit, fault prediction unit, and fault management system unit are all connected to the data monitoring database. The fault diagnosis system unit and template database management unit are all connected to the template database. The library is connected with the case library, and the equipment information management unit and the equipment maintenance system unit are connected with the equipment information library.
如附图1、2所示，所述电参数采集模块、温度传感器、压力传感器、动液面测量仪、一体化示功仪分别采集现场的电机电流和功率参数、油温参数、油管压力参数和套管压力参数、动液面深度、抽油机载荷功图和角位移功图并通过无线通信传输至远程测控模块，流量传感器和可燃气体检测仪分别采集现场的油田产液量和可燃气体浓度并通过有线通信传输至远程测控模块；或/和，所述远程测控模块为RTU远程测控终端。这里的RTU远程控制终端提供了多个与现场测试端的接口，可分别采集各生产数据，并进行数据转换后存储在临时寄存器中，再通过无线接口发送给远端的无线服务器。这里的无线通信传输采用ZigBee通信，有线通信传输采用RS485有线通信。As shown in Figures 1 and 2, the electrical parameter acquisition module, temperature sensor, pressure sensor, dynamic liquid level measuring instrument, and integrated dynamometer respectively collect on-site motor current and power parameters, oil temperature parameters, and oil pipe pressure parameters. And the casing pressure parameters, dynamic liquid surface depth, pumping unit load work diagram and angular displacement work diagram are transmitted to the remote measurement and control module through wireless communication, and the flow sensor and combustible gas detector collect the on-site oilfield liquid production and combustible gas The concentration is transmitted to the remote measurement and control module through wired communication; or/and, the remote measurement and control module is an RTU remote measurement and control terminal. The RTU remote control terminal here provides multiple interfaces with the on-site test terminal, which can collect each production data separately, convert the data and store it in the temporary register, and then send it to the remote wireless server through the wireless interface. The wireless communication transmission here adopts ZigBee communication, and the wired communication transmission adopts RS485 wired communication.
实施例2：如附图1、2、3、4、5、6、7所示，一种上述油井故障智能化分析决策系统的故障诊断方法，包括以下步骤：Embodiment 2: as shown in accompanying drawing 1,2,3,4,5,6,7, a kind of fault diagnosis method of above-mentioned oil well fault intelligent analysis decision-making system comprises the following steps:
第一步，功图转化，数据采集模块将采集得到设备参数传输至运行监控模块，更新专家数据库后由故障诊断系统单元将地面测的示功图转换为井下有杆泵示功图，完成功图转换，步骤如下：The first step is the conversion of the dynamometer diagram. The data acquisition module will transmit the acquired equipment parameters to the operation monitoring module. After updating the expert database, the fault diagnosis system unit will convert the dynamometer diagram measured on the ground into the dynamometer diagram of the downhole rod pump, and the completion is successful. Graph conversion, the steps are as follows:
建立波动方程，公式如下：Establish the wave equation, the formula is as follows:
式中：u(x,t)为抽油杆柱在x深度处的横截面在t时刻的位移；x为深度；t为时间；a为应力在抽油杆柱上的传播速度；c为阻尼系数；In the formula: u(x,t) is the displacement of the cross section of the sucker rod string at the depth x at time t; x is the depth; t is the time; a is the propagation velocity of the stress on the sucker rod string; c is damping coefficient;
其上、下边界条件的傅里叶级数表示及初始条件公式如下：The Fourier series expression of the upper and lower boundary conditions and the initial condition formula are as follows:
其中：f(x,t)为柱塞在位移x、时刻t时的载荷；D(t)为悬点在t时刻的位移；U(t)为光杆在时刻t时的位移；σn、τn、vn、δn为傅里叶系数；Among them: f(x,t) is the load of the plunger at displacement x and time t; D(t) is the displacement of the suspension point at time t; U(t) is the displacement of polished rod at time t; σn , τn , vn , δn are Fourier coefficients;
通过分离变量及傅里叶变换可以求出u(x,t)：u(x,t) can be found by separating variables and Fourier transform:
根据胡克定律，可以求得According to Hooke's law, we can get
式中：In the formula:
On(x)＝(knchβnx+δnshβnx)sinαnx+(unchβnx+vnshβnx)cosαnx (5)On (x)＝(kn chβn x+δn shβn x)sinαn x+(un chβn x+vn shβn x)cosαn x (5)
Pn(x)＝(knchβnx+δnchβnx)cosαnx+(μnchβnx+vnshβnx)sinαnx (6)Pn (x)＝(kn chβn x+δn chβn x)cosαn x+(μn chβn x+vn shβn x)sinαn x (6)
根据波动方程的解u(x,t)，f(x,t)及其边界条件将井上获取的地面示功图转换为井下泵功图；这里转化之后，对功图进行特征值提取、识别，识别结果显示示功图经过转化之后，识别率大大提高。According to the solution u(x,t) of the wave equation, f(x,t) and their boundary conditions, the ground indicator map obtained on the well is converted into a downhole pump work map; after the transformation, the eigenvalue extraction and identification of the work map are carried out , the recognition results show that after the dynamometer diagram is converted, the recognition rate is greatly improved.
第二步，对明显特征功图进行分类为需要神经网络分析和不需要神经网络分析，需要神经网络分析的功图进入第三步，不需要神经网络分析的功图进入第四步；The second step is to classify the obvious feature graphs into those that require neural network analysis and those that do not require neural network analysis. The graphs that require neural network analysis enter the third step, and the graphs that do not require neural network analysis enter the fourth step;
第三步，根据不同故障功图变化，需要神经网络分析的功图分别采用几何法、网格法和傅立叶描述子法对功图特征值进行提取，之后进入第五步；对于需要神经网络分析故障如“结蜡、气体影响、供液不足、稠油影响、砂卡、漏失、气锁”等故障结合所有的特征值相关参数以及生产数据(如产液量、动液面、理论最大最小载荷)、模版库，通过BP神经网络算法实现复杂功图的自动诊断。In the third step, according to the change of different fault power diagrams, the geometric method, grid method and Fourier descriptor method are used to extract the characteristic values of the power diagrams that require neural network analysis, and then enter the fifth step; for those that require neural network analysis Faults such as "wax deposition, gas impact, insufficient liquid supply, heavy oil impact, sand stuck, leakage, gas lock" and other faults combined with all characteristic value related parameters and production data (such as liquid production, dynamic liquid level, theoretical maximum and minimum Load), template library, and automatic diagnosis of complex work diagrams through BP neural network algorithm.
第四步，根据不同故障功图变化，不需要神经网络分析的功图采用灰度矩阵对功图特征值进行提取，之后进入第六步；In the fourth step, according to the change of different fault power diagrams, the gray scale matrix is used to extract the characteristic values of the power diagrams for the power diagrams that do not require neural network analysis, and then enter the sixth step;
第五步，根据数据采集模块采集的生产数据结合第三步提取的功图特征值，调用单井模板库和区块模板库对故障进行诊断分析，之后进入第七步；这里的第五步采用BP神经网络算法对故障进行诊断分析，使用三层神经网络来实现，BP神经网络的输入层为29个特征值点；BP神经网络隐含层选择Sigmoid型函数进行内部学习及处理；输出节点设为7个，如设置正常功图对应的输出为[000 0001]，碰泵功图对应的输出为[000 0010]，供液不足功图对应输出为[000 0100],气体影响功图对应输出为[000 1000]。在实验时，选取现场搜集的500幅功图数据作为训练数据，网络训练好之后，选取部分数据进行测试，训练结果如附图7所示。In the fifth step, according to the production data collected by the data acquisition module combined with the characteristic value of the power map extracted in the third step, call the single well template library and the block template library to diagnose and analyze the fault, and then enter the seventh step; the fifth step here The BP neural network algorithm is used to diagnose and analyze the fault, and the three-layer neural network is used to realize. The input layer of the BP neural network is 29 eigenvalue points; the hidden layer of the BP neural network selects the Sigmoid function for internal learning and processing; the output node Set to 7, such as setting the output corresponding to the normal work diagram to [000 0001], the corresponding output to the bump pump work diagram to [000 0010], the corresponding output to the insufficient liquid supply diagram to [000 0100], and the corresponding output to the gas impact work diagram The output is [000 1000]. During the experiment, 500 pieces of power map data collected on site were selected as training data. After the network was trained, some data were selected for testing. The training results are shown in Figure 7.
第六步，根据数据采集模块采集的生产数据结合第三步提取的功图特征值，运用欧式距离分类进行诊断分析，之后进入第七步；对于不需要神经网络分析故障如“抽油杆断脱、泵卡、连抽带喷、活塞出泵筒”等故障直接结合生产数据(如产液量、动液面、理论最大最小载荷等)使用欧式距离分类方法实现故障诊断。In the sixth step, according to the production data collected by the data acquisition module combined with the characteristic value of the power map extracted in the third step, the Euclidean distance classification is used to perform diagnostic analysis, and then enter the seventh step; for faults that do not require neural network analysis such as "sucker rod break Faults such as disengagement, pump stuck, continuous pumping and spraying, and piston out of the pump barrel" are directly combined with production data (such as liquid production volume, dynamic liquid level, theoretical maximum and minimum loads, etc.) and use the Euclidean distance classification method to realize fault diagnosis.
第七步，得出故障诊断结果，对故障诊断结果进行现场验证，若现场验证正确，则将相应故障诊断结果统计数自动加1次；若现场验证错误，则修正故障诊断结果并将该单井功图作为模板存入相应的单井模版库中。The seventh step is to obtain the fault diagnosis result, and carry out on-site verification of the fault diagnosis result. If the on-site verification is correct, the statistics of the corresponding fault diagnosis result will be automatically added once; if the on-site verification is wrong, the fault diagnosis result will be corrected and the list The well work diagram is stored as a template in the corresponding single well template library.
如附图1、2、3、4、5、6、7、表1所示，在第三步中，几何法特征值提取是基于泵功图进行几何操作获取特征值，对所述泵功图提取的几何特征值为:上冲程始点A、左下面积A1、左上面积A2、右上面积A3、右下面积A4、上行变形结束点B、下行变形结束点D、上死点C、曲线质心G、最小载荷Fmax和最小载荷Fmin。As shown in accompanying drawing 1, 2, 3, 4, 5, 6, 7, table 1, in the 3rd step, geometrical method eigenvalue extraction is to carry out geometric operation and obtain eigenvalue based on pump power diagram, to described pump power The geometric feature values extracted from the graph are: starting point A of upstroke, lower left area A1, upper left area A2, upper right area A3, lower right area A4, upward deformation end point B, downward deformation end point D, upper dead point C, and curve centroid G , minimum load Fmax and minimum load Fmin.
如附图1、2、3、4、5、6、7所示，在第三步中，所述的网格特征值提取法，对功图图像进行网格化处理，将示功图分解成若干个形状、大小相同的网格，对有示功图穿过的网格赋值“1”，对于其它网格，以功图穿过的网格为中心，当其位于功图外部时，每远离一格，灰度值减去1，当位于功图内部时，每远离一格，灰度值增加1，计算灰度矩阵的灰度均值、灰度方差、灰度偏差、灰度峰度、灰度能量、灰度熵六个特征参数，公式如下：As shown in accompanying drawings 1, 2, 3, 4, 5, 6, and 7, in the third step, the grid eigenvalue extraction method performs grid processing on the gong diagram image, and decomposes the gong diagram into several grids of the same shape and size, and assign a value of "1" to the grid passing through the dynamometer diagram, and for other grids, take the grid passing through the dynamometer diagram as the center, and when it is located outside the dynamometer diagram, Every time you move away from a grid, the gray value is subtracted by 1. When it is inside the power map, every time you get away from it, the gray value increases by 1. Calculate the gray mean, gray variance, gray deviation, and gray peak of the gray matrix degree, gray energy, gray entropy six characteristic parameters, the formula is as follows:
灰度均值：Grayscale mean:
灰度方差：Gray variance:
灰度偏差：Gray scale deviation:
灰度峰度：Grayscale kurtosis:
灰度能量：Gray energy:
灰度熵：Gray entropy:
以此作为特征值与标准特征值进行比较；该网格法得到的特征值具有较低的维数，只能识别特征较为明显的示功图。Take this as the eigenvalue and compare it with the standard eigenvalue; the eigenvalue obtained by the grid method has a lower dimension, and can only identify the dynamometer diagram with more obvious characteristics.
如附图1、2、3、4、5、6、7所示，在第三步中，傅立叶描述子法特征值提取是对功图进行傅立叶级数展开表征功图形状，即沿示功图轮廓线的一个动点P(l)的坐标变化是一个以功图周长为周期的周期函数，该函数可以用傅里叶级数表示，其系数a(k)称为傅里叶描述子，表达式如下：As shown in accompanying drawings 1, 2, 3, 4, 5, 6, and 7, in the third step, the Fourier descriptor method eigenvalue extraction is to perform Fourier series expansion on the work graph to characterize the shape of the work graph, that is, along the The coordinate change of a moving point P(l) on the graph contour is a periodic function whose period is the circumference of the graph. This function can be expressed by Fourier series, and its coefficient a(k) is called Fourier description. , the expression is as follows:
进行归一化后公式：After normalization, the formula is:
式中是对傅里叶系数a(k)取模；由于变换后的高频系数较小，因此，通常只选取低频系数作为特征值，根据对功图的傅里叶变换可知，当k>20时，傅里叶描述子变得十分小，故选取k＝20。为了使傅里叶描述子与功图的旋转、平移和尺度无关，所以对a(k)进行归一化计算，该方法能对功图的细节特征进行描述，能识别较为复杂的功图。In the formula is to take the modulus of the Fourier coefficient a(k); since the transformed high-frequency coefficient is small, usually only the low-frequency coefficient is selected as the eigenvalue. According to the Fourier transform of the power diagram, when k>20 , the Fourier descriptor becomes very small, so k=20 is selected. In order to make the Fourier descriptor irrelevant to the rotation, translation and scale of the work graph, a(k) is normalized and calculated. This method can describe the detailed features of the work graph and can identify more complex work graphs.
实施例3：如附图1、2、8、9所示，一种上述油井故障智能化分析决策系统的故障预测方法，包括以下步骤：Embodiment 3: As shown in accompanying drawing 1,2,8,9, a kind of failure prediction method of above-mentioned oil well failure intelligent analysis decision-making system, comprises the following steps:
第一步，提取单井近期的5组正常示功图的特征值，之后进入第二步；The first step is to extract the eigenvalues of the recent 5 groups of normal dynamometer diagrams of a single well, and then enter the second step;
第二步，将单井近期5组正常示功图的特征值与模板库中正常示功图进行对比，计算相似度，之后进入第三步；The second step is to compare the eigenvalues of the recent 5 groups of normal dynamometer diagrams of a single well with the normal dynamometer diagrams in the template library, calculate the similarity, and then enter the third step;
第三步，建立数据的动态模型，即GM(h,n)模型，其微分方程是时间域的连续函数，h为方程阶数，n为变量个数，微分方程如下公式所示：The third step is to establish a dynamic model of the data, that is, the GM(h,n) model. Its differential equation is a continuous function in the time domain, h is the order of the equation, and n is the number of variables. The differential equation is shown in the following formula:
则微分方程的系数向量为：Then the coefficient vector of the differential equation for:
采用最小二乘法求解，结果为：Using the method of least squares to solve, the result is:
式中为A组成的分块矩阵，之后进入第四步；In the formula is the block matrix composed of A, and then enters the fourth step;
第四步，当发现特征值与正常示功图匹配的越来越低时，对下一组示功图数据进行预测，判断其是否会发生故障；若满足偏离界限，立即通过故障监测预警单元进行预警。In the fourth step, when it is found that the matching between the characteristic value and the normal dynamometer is getting lower and lower, predict the next set of dynamometer data to judge whether it will fail; if it meets the deviation limit, immediately pass the fault monitoring and early warning unit Forewarning.
如附图1、2、8、9所示，上述第三步中，GM(1，1)模型的建模原理为：As shown in accompanying drawings 1, 2, 8, and 9, in the third step above, the modeling principle of the GM (1, 1) model is:
(1)设原始序列为X(0)，有：(1) Let the original sequence be X(0) , there are:
X(0)＝[x(0)(1),x(0)(2),...,x(0)(n)] (22)X(0) ＝[x(0) (1),x(0) (2),...,x(0) (n)](22)
对原始序列X(0)进行一次累加生成序列X(1)，得到：Perform an accumulation of the original sequence X(0) to generate a sequence X(1) to obtain:
X(1)＝[x(1)(1),x(1)(2),...,x(1)(n)] (23)X(1) ＝[x(1) (1),x(1) (2),...,x(1) (n)](23)
原始序列X(0)与一次累加生成序列X(1)中的对应数据之间的关系为：The relationship between the original sequence X(0) and the corresponding data in the accumulated sequence X(1) is:
(2)建立GM(1，1)模型的白化微分方程为：(2) Establish the whitening differential equation of the GM(1,1) model as:
GM(1，1)模型的灰微分方程为：The gray differential equation of the GM(1,1) model is:
x(0)(k)+az(1)(k)＝b k＝1,2,..,n (26)x(0) (k)+az(1) (k)=bk=1,2,..,n (26)
公式(26)中，a为发展系数，b为灰作用量，均为待辨识的参数，x(0)(k)为灰导数，z(1)(k)为背景值，背景值的计算公式如式(27)：In formula (26), a is the development coefficient, b is the gray action, both are parameters to be identified, x(0) (k) is the gray derivative, z(1) (k) is the background value, and the calculation of the background value The formula is as formula (27):
(3)设待求参数Y＝[x(0)(2),x(0)(3),...,x(0)(n)]T(3) Set the parameters to be requested Y＝[x(0) (2),x(0) (3),...,x(0) (n)]T
根据最小二乘原理，计算微分方程的解为：According to the principle of least squares, the solution of the differential equation is calculated as:
(4)令x(1)(0)＝x(0)(0)，则原始序列的预测公式为：(4) Let x(1) (0)=x(0) (0), then the prediction formula of the original sequence is:
式(30)中：为x(0)(k)的拟合值，为第k+1个预测值。In formula (30): is the fitted value of x(0) (k), is the k+1th predicted value.
实施例4：如附图1、2、10所示，一种油井故障智能化分析决策系统的决策方案使用方法，包括以下步骤：Embodiment 4: As shown in accompanying drawings 1, 2, and 10, a method for using a decision-making scheme of an intelligent analysis and decision-making system for oil well failures comprises the following steps:
第一步，在进行故障诊断结果之后，根据故障类型进行问题描述，在案例库中进行案例检索，之后进入第二步；The first step is to describe the problem according to the fault type after the fault diagnosis result, and search the case in the case library, and then enter the second step;
第二步，通过决策方案模块查找该故障解决方案是否在数据库中，若数据库中有解决方案，则得到存在解决方案的结论之后进入第三步，如果未找到对应的解决方案，用户针对诊断结果分析后添加相应的解决方案，补充数据库，结束；The second step is to find out whether the fault solution is in the database through the decision-making solution module. If there is a solution in the database, it will enter the third step after the conclusion that there is a solution. If no corresponding solution is found, the user will check the diagnosis result After the analysis, add the corresponding solution, supplement the database, and end;
第三步，判断解决方案中是否存在满意的解决方案，若没有满意的解决方案用户自行设计适合的解决方案，并将解决结果以案例形式添加到案例库，结束；如寻找到满意的解决方案，之后进入第四步；The third step is to judge whether there is a satisfactory solution in the solution. If there is no satisfactory solution, the user can design a suitable solution by himself, and add the solution result to the case library in the form of a case, and end; if a satisfactory solution is found , and then enter the fourth step;
第四步，如果有满意的方案，根据用户的选择，将该解决方案采纳次数自动加1，在下次寻找方案时，优先显示采纳次数最多的方案，之后结束。Step 4: If there is a satisfactory solution, according to the user's choice, the number of adoptions of the solution will be automatically increased by 1. When looking for a solution next time, the solution with the most adoption times will be displayed first, and then it will end.
以上技术特征构成了本发明的实施例，其具有较强的适应性和实施效果，可根据实际需要增减非必要的技术特征，来满足不同情况的需求。The above technical features constitute the embodiment of the present invention, which has strong adaptability and implementation effect, and non-essential technical features can be increased or decreased according to actual needs to meet the needs of different situations.
表1几何特征值Table 1 Geometric eigenvalues
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611029035.3ACN106761668B (en) | 2016-11-19 | 2016-11-19 | Oil well failure intelligent analysis decision system and method |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611029035.3ACN106761668B (en) | 2016-11-19 | 2016-11-19 | Oil well failure intelligent analysis decision system and method |
| Publication Number | Publication Date |
|---|---|
| CN106761668A CN106761668A (en) | 2017-05-31 |
| CN106761668Btrue CN106761668B (en) | 2019-11-26 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611029035.3AActiveCN106761668B (en) | 2016-11-19 | 2016-11-19 | Oil well failure intelligent analysis decision system and method |
| Country | Link |
|---|---|
| CN (1) | CN106761668B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108122075A (en)* | 2017-12-22 | 2018-06-05 | 胜利油田鲁明油气勘探开发有限公司 | A kind of sucker rod pump operating mode intelligent monitoring method based on convolutional neural networks intensified learning |
| CN108386182B (en)* | 2018-01-16 | 2021-08-13 | 中国海洋石油集团有限公司 | Oil well production abnormity early warning method based on SPC control chart and weighted decision tree |
| CN108729902B (en)* | 2018-05-03 | 2021-09-10 | 西安永瑞自动化有限公司 | Online fault diagnosis system and method for oil pumping unit |
| CN108412482A (en)* | 2018-05-23 | 2018-08-17 | 大庆市亿动科技有限公司 | Oil well comprehensive parameters TT＆C system based on LORA technologies |
| CN108952673B (en)* | 2018-06-22 | 2023-09-26 | 中国石油天然气股份有限公司 | Method and device for checking working condition of oil pumping well |
| CN108915668A (en)* | 2018-07-17 | 2018-11-30 | 东北大学 | A kind of Diagnosing The Faults of Sucker Rod Pumping System method based on gray level co-occurrence matrixes |
| CN109507924B (en)* | 2018-11-28 | 2021-06-29 | 北京迈斯康特测控技术有限公司 | Remote monitoring system for oil field operation equipment |
| CN109491341B (en)* | 2018-11-28 | 2021-04-13 | 北京迈斯康特测控技术有限公司 | Monitoring and data analysis system for oil field operation equipment |
| CN109977156A (en)* | 2019-03-22 | 2019-07-05 | 常州大学 | A kind of real-time analysis and management system of oil well condition enhancing study and management method |
| CN110552685B (en)* | 2019-08-19 | 2022-08-19 | 大庆油田有限责任公司 | Method for calculating working fluid level of oil well by utilizing ground indicator diagram in wax precipitation well |
| CA3060937C (en)* | 2019-11-05 | 2023-06-13 | Weatherford Technology Holdings, Llc | System and method for controlling artificial lift units |
| CN111520114B (en)* | 2020-03-27 | 2022-07-08 | 上海复泉工程技术有限公司 | Intelligent management system for pump efficiency adjustment oil field |
| CN111594146A (en)* | 2020-05-28 | 2020-08-28 | 中国石油集团渤海钻探工程有限公司 | Liquid level monitoring and early warning system for well drilling |
| US20220010668A1 (en)* | 2020-07-10 | 2022-01-13 | Halliburton Energy Services, Inc. | Wellbore isolation barrier monitoring |
| CN112508105B (en)* | 2020-12-11 | 2024-03-19 | 南京富岛信息工程有限公司 | Fault detection and retrieval method for oil extraction machine |
| CN112878989A (en)* | 2021-02-04 | 2021-06-01 | 东北大学 | Real-time monitoring method and system for hot washing process of sucker rod pump oil well based on electric power |
| CN113112639A (en)* | 2021-05-20 | 2021-07-13 | 宏华油气工程技术服务有限公司 | Parameter recording system and method suitable for petroleum drilling |
| CN115434669A (en)* | 2021-06-02 | 2022-12-06 | 中国石油化工股份有限公司 | Intelligent closed-loop control frequency conversion system and method for wellhead of oil pumping unit |
| CN113279812B (en)* | 2021-07-05 | 2024-04-09 | 太原科技大学 | Method and system for state monitoring and residual life prediction of mine main drainage equipment |
| CN115749737A (en)* | 2021-09-02 | 2023-03-07 | 中国石油天然气股份有限公司 | Artificial lifting oil well indicator diagram acquisition system and acquisition method |
| CN114048998A (en)* | 2021-11-10 | 2022-02-15 | 山东寿光市坤隆石油机械股份有限公司 | Oil extraction equipment management system |
| CN114263458B (en)* | 2021-12-27 | 2023-04-07 | 西安健尚智能科技有限公司 | Method and system for full-perception intelligent diagnosis automatic processing of oil well working condition |
| CN114459766B (en)* | 2022-01-13 | 2024-06-11 | 武汉理工大学 | Method for monitoring working state of oil head of crude oil generator set on ocean platform |
| CN114578160B (en)* | 2022-02-28 | 2024-12-10 | 国网陕西省电力有限公司电力科学研究院 | Casing condition monitoring and fault diagnosis system and method based on digital twin model |
| CN115082255A (en)* | 2022-05-20 | 2022-09-20 | 中国石油化工股份有限公司 | Coal gasification device remote intelligent service method and intelligent expert system |
| CN114837654A (en)* | 2022-05-30 | 2022-08-02 | 杭州瑞利超声科技有限公司 | Multi-terminal monitoring system for oil well dynamic liquid level based on Internet of Things and cloud platform |
| CN117173434B (en)* | 2023-09-05 | 2024-05-14 | 大庆石油管理局有限公司 | Comprehensive working condition data processing method and system based on tower electromechanical parameter indicator diagram |
| CN117449833B (en)* | 2023-12-07 | 2024-05-10 | 北京华晖恒泰能源科技有限公司 | Oil field oil-well pump running state analysis monitoring system |
| CN117404072B (en)* | 2023-12-15 | 2024-02-23 | 山东新云鹏电气有限公司 | Drilling site management system based on artificial intelligence |
| CN117722173B (en)* | 2024-02-06 | 2024-04-30 | 灵知科技(大庆)有限公司 | Intelligent diagnosis measurement and control system and device for monitoring dynamic parameters of multiple scenes |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6751555B2 (en)* | 2001-10-17 | 2004-06-15 | Schlumberger Technology Corporation | Method and system for display of well log data and data ancillary to its recording and interpretation |
| CN102184414B (en)* | 2011-05-16 | 2013-03-13 | 中国石油天然气股份有限公司 | Method and system for identifying and judging pump indicator diagram |
| CN103806893B (en)* | 2012-11-14 | 2016-07-06 | 中国石油天然气股份有限公司 | Intelligent detection method, device and detection system |
| CN103953328A (en)* | 2014-04-18 | 2014-07-30 | 中国石油天然气股份有限公司 | Oil well testing and analyzing system |
| CN104731079A (en)* | 2015-04-09 | 2015-06-24 | 沈阳日明自动化工程系统有限公司 | Oil field data monitoring system and method |
| Publication number | Publication date |
|---|---|
| CN106761668A (en) | 2017-05-31 |
| Publication | Publication Date | Title |
|---|---|---|
| CN106761668B (en) | Oil well failure intelligent analysis decision system and method | |
| CN109272123B (en) | A working condition early warning method of rod pump based on convolutional neural network | |
| CN101799359B (en) | A power equipment failure monitoring and forecasting method and system thereof | |
| CN106321072B (en) | A fault diagnosis method for pumping wells based on pump power map | |
| CN105257277B (en) | Dlagnosis of Sucker Rod Pumping Well underground failure prediction method based on Multi-variable Grey Model | |
| CN111963116B (en) | An intelligent gas field system and method for adaptive and intelligent analysis and decision making | |
| CN106121622B (en) | A Multiple Fault Diagnosis Method for Rod Pumped Oil Wells Based on Indicator Diagram | |
| CN104405365A (en) | Pumping unit indicator diagram liquid production capacity measurement technology | |
| CN106682815B (en) | Monitoring system and monitoring method for pumping well | |
| CN103061749B (en) | Soft measuring method of dynamic liquid level under sucker rod oil well | |
| CN106837301A (en) | The rod-pumped well diagnostic method of working condition that a kind of electric work figure is integrated with indicator card | |
| CN101059130A (en) | On-line remote state monitoring and fault analysis diagnosis system of reciprocating compressor | |
| CN103147972A (en) | Reciprocating-type compressor fault diagnosis method based on multi-sensor information fusion | |
| CN104794283A (en) | Diesel fault grey prediction method based on outlier characteristic analysis | |
| Wei et al. | Fault diagnosis of sucker rod pump based on deep-broad learning using motor data | |
| CN104295373B (en) | Diesel engine fault Forecasting Methodology based on three point models | |
| CN118820862B (en) | Knowledge graph-based fault diagnosis method for indicator diagram | |
| CN112112631A (en) | Intelligent identification method for indicator diagram of oil pumping unit | |
| CN108678728A (en) | A kind of oilwell parameter analysis combinational algorithm based on k-means | |
| CN115146690A (en) | A kind of intelligent analysis method and device for pumping unit well | |
| CN117454271A (en) | Sucker rod pump oil well fault diagnosis method and system | |
| CN109236277A (en) | A kind of oil well fault diagnostic expert system based on production rule | |
| CN118088152A (en) | Electric pump well working condition intelligent analysis and optimization system and method based on deep learning | |
| Ying et al. | Integrated fault diagnosis method for down-hole working conditions of the beam pumping unit | |
| CN2122232U (en) | Intelligent test analyzer of multi-parameter beam-pumping unit |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | Effective date of registration:20241025 Address after:No. 26, Yupan Road, Karamay District, Karamay City, Xinjiang Uygur Autonomous Region 834000 Patentee after:Karamay Hongzhi Oilfield Technology Service Co.,Ltd. Country or region after:China Address before:No. 12-1, Jinxing Road, Karamay District, Karamay City, Xinjiang Uygur Autonomous Region 834000 Patentee before:XINJIANG HUALONG OILFIELD TECHNOLOGICAL LLC Country or region before:China | |
| TR01 | Transfer of patent right |