Aiming at the paraffin-deposition problem of a beam well, the automatic paraffin-controlling device is designed by making use of ratchet-pallet mechanism, cam mechanism and modern designing method. The device has four...Aiming at the paraffin-deposition problem of a beam well, the automatic paraffin-controlling device is designed by making use of ratchet-pallet mechanism, cam mechanism and modern designing method. The device has four main functions： paraffin-controlling, paraffin removal, centralizing the pumping rod, and improving the safety of well tubing. This device integrates the advantages of the paraffin control, such as strong magnetic paraffin controlling and mechanical paraffin-cutting. Theoretical analysis shows that this device has fine working reliability. It turns out to be a new device which can solve the paraffin-deposition problem of a beam well economically and efficiently.展开更多
文摘Aiming at the paraffin-deposition problem of a beam well, the automatic paraffin-controlling device is designed by making use of ratchet-pallet mechanism, cam mechanism and modern designing method. The device has four main functions： paraffin-controlling, paraffin removal, centralizing the pumping rod, and improving the safety of well tubing. This device integrates the advantages of the paraffin control, such as strong magnetic paraffin controlling and mechanical paraffin-cutting. Theoretical analysis shows that this device has fine working reliability. It turns out to be a new device which can solve the paraffin-deposition problem of a beam well economically and efficiently.
文摘目的 建立基于长短期记忆(longshort-termmemory,LSTM)神经网络的时间序列预测模型,对北京地区鲜(冻)肉制品铅含量进行风险评估与预测预警。方法 通过收集2011—2020年国家市场监督管理总局日常食品监督管理抽检数据,筛选出北京地区鲜(冻)肉制品的抽检数据,构建数据集并进行预处理,按照7:2:1的比例划分训练集、验证集和测试集,基于Tensorflow平台构建4层LSTM模型并进行训练,基于重大活动举办前10d的鲜(冻)肉制品铅含量数据,对未来1d的鲜(冻)肉制品铅含量风险进行预测。结果 经过50轮模型迭代训练,训练集和测试集Loss指标收敛至0.084,经过5次训练后的模型评估参数均方根误差(root mean square error,RMSE)为0.192,决定系数(coefficient of determination,R2_score)为0.916,模型误差较小、准确度较高。结论 基于LSTM的鲜(冻)肉制品铅含量风险预测模型整体性能较好,可应用于重大活动举办地的食品风险预测,并精准指导监督抽检。
