摘要：【中文】本发明涉及一种净水机降噪冲洗结构，包括滤芯组及RO滤芯，滤芯组的进水口与外界原水连通；还包括冲洗阀、进水阀、增压水泵及控制模块；滤芯组的出水口与进水阀的进水口连通，进水阀的出水口与增压水泵的进水口连通，增压水泵的出水口与RO滤芯的进水口连通，RO滤芯的纯水出水口与外界水路连通，RO滤芯的废水出水口通过冲洗阀与外界浓水出水口连通；控制模块分别与增压水泵、进水阀及冲洗阀电连接用于驱动增压水泵、进水阀及冲洗阀工作。其具有冲洗时，噪音小，平稳，冲洗方便等优点。 【EN】The invention relates to a noise reduction flushing structure of a water purifier, which comprises a filter element group and an RO (reverse osmosis) filter element, wherein a water inlet of the filter element group is communicated with external raw water; the device also comprises a flushing valve, a water inlet valve, a booster water pump and a control module; the water outlet of the filter element group is communicated with the water inlet of the water inlet valve, the water outlet of the water inlet valve is communicated with the water inlet of the booster water pump, the water outlet of the booster water pump is communicated with the water inlet of the RO filter element, the pure water outlet of the RO filter element is communicated with an external water path, and the wastewater outlet of the RO filter element is communicated with an external concentrated water outlet through the flushing valve; the control module is respectively electrically connected with the booster water pump, the water inlet valve and the flushing valve and is used for driving the booster water pump, the water inlet valve and the flushing valve to work. It has the advantages of low noise, stability, convenient flushing and the like when in flushing.

摘要：【中文】本发明公开了一种基于烟花算法和支持向量机的恶意软件检测方法，利用烟花算法自适应进化特性，实现对支持向量机参数的自适应调节，使得模型利用软件特征数据集进行训练学习过程中能够有效实现自适应提高恶意软件检测准确率，能够用于计算机信息安全领域中；本发明采用正确率均值作为评价函数有效地避免了训练模型在某一数据集上出现过拟合的情况，可以很好地、客观地提升模型整体性能，提高鲁棒性，同时采用烟花算法优化求解最优参数，可以满足快速自适应增强恶意软件检测精度的要求。 【EN】The invention discloses a malicious software detection method based on a firework algorithm and a support vector machine, which utilizes the adaptive evolutionary characteristic of the firework algorithm to realize the adaptive adjustment of parameters of the support vector machine, so that a model can effectively realize the adaptive improvement of the malicious software detection accuracy rate in the training and learning process by utilizing a software characteristic data set, and can be used in the field of computer information security; the method adopts the correct rate mean value as an evaluation function to effectively avoid the over-fitting condition of a training model on a certain data set, can well and objectively improve the overall performance of the model and improve the robustness, and simultaneously adopts the firework algorithm to optimize and solve the optimal parameters, so that the requirement of rapidly and adaptively enhancing the detection precision of the malicious software can be met.

摘要：【中文】本发明涉及植物生态与栽培引种领域，公开了利用气候数据预测金钱松适宜引种地的方法，包括：将金钱松目标引种地的经度、纬度以及海拔高度数据代入预测方程中，得到预测年均温度和预测年均降水量，预测方程为：T＝44.481－0.077E－0.613N－0.004H，P＝1545.865+22.192E－93.973N+0.298H，其中T、P、E、N、H分别为预测年均温度、预测年均降水量，目标引种地的经度、纬度以及海拔高度；将预测年均温度和预测年均降水量分别与目标引种地的历史年均温度和历史年均降水量进行比较，误差在10％以内则认为适宜引种。该预测方法，可快速、准确预测目标引种地是否适合引种。 【EN】The invention relates to the field of plant ecology and cultivation introduction, and discloses a method for predicting a suitable introduction place of a Jinqian pine by using climate data, which comprises the following steps: substituting longitude, latitude and altitude data of the loosestrife target indexing land into a prediction equation to obtain a predicted annual average temperature and a predicted annual average precipitation, wherein the prediction equation is as follows: t is 44.481-0.077E-0.613N-0.004H, P is 1545.865+ 22.192E-93.973N +0.298H, wherein T, P, E, N, H respectively represents the predicted annual average temperature, the predicted annual average precipitation and the longitude, latitude and altitude of the target introduction place; and comparing the predicted annual average temperature and the predicted annual average precipitation with the historical annual average temperature and the historical annual average precipitation of the target index land respectively, and determining that the seeds are suitable for introduction when the error is within 10%. The prediction method can quickly and accurately predict whether the target introduction place is suitable for introduction.

摘要：【中文】本发明公开了一种甚高分辨率红外探测器芯片与读出电路芯片倒装互连方法，该方法包括：通过倒装焊设备预压红外探测器芯片与读出电路芯片，使红外探测器芯片与读出电路芯片初步互相连接；通过超大压力键合设备压焊经过预压后初步互相连接的红外探测器芯片与读出电路芯片。本发明采用先预压后二次压焊的方式克服了超高互连精度和超大互连压力之间的矛盾，解决了现有技术中传统的工艺难以保证芯片良好的混成互连耦合效果的问题。 【EN】The invention discloses a flip interconnection method of a very high resolution infrared detector chip and a reading circuit chip, which comprises the following steps: prepressing the infrared detector chip and the reading circuit chip by flip chip welding equipment to preliminarily connect the infrared detector chip and the reading circuit chip; and (3) bonding the infrared detector chip and the reading circuit chip which are preliminarily connected with each other after pre-pressing through ultra-high pressure bonding equipment. The invention overcomes the contradiction between the ultrahigh interconnection precision and the ultrahigh interconnection pressure by adopting a mode of pre-pressing and then secondary pressure welding, and solves the problem that the traditional process in the prior art cannot ensure the good mixed interconnection coupling effect of the chip.

摘要：【中文】本发明公开了一种基于深度学习的地下管道异常自动化定位与识别方法，包括：地下管道机器人获取的图像进行数据标注形成训练、验证及测试数据集；管道异常主要包括破裂、变形、障碍物、异物插入以及沉积共5类异常情况；构建深度学习网络并训练模型以实现地下管道异常自动化定位与识别；最终输出带有识别标注结果的图像，其中包括：识别到的缺陷位置、种类以及相应的类别概率。本发明利用大量已经标注好的地下管道图像数据进行深度模型训练，生成具有定位与识别功能的深度模型，能够对单幅图像中的多种缺陷进行定位识别，识别率和准确率高，识别速度快。 【EN】The invention discloses an underground pipeline abnormity automatic positioning and identifying method based on deep learning, which comprises the following steps: carrying out data annotation on an image acquired by the underground pipeline robot to form a training, verifying and testing data set; the pipeline abnormity mainly comprises 5 types of abnormal conditions of fracture, deformation, obstacles, foreign body insertion and deposition; constructing a deep learning network and training a model to realize automatic positioning and identification of underground pipeline abnormity; and finally outputting the image with the identification annotation result, wherein the image with the identification annotation result comprises the following steps: the identified defect location, category, and corresponding category probability. According to the invention, a depth model is trained by using a large amount of marked image data of the underground pipeline to generate a depth model with positioning and identifying functions, so that various defects in a single image can be positioned and identified, the identification rate and accuracy are high, and the identification speed is high.

摘要：【中文】本发明公开一种制备超薄碳/碳复合材料双极板的方法，其采用超细碳纤维网胎成型工艺、石墨烯改性树脂/沥青料浆分散技术、碳纤维网胎/料浆流场压延成型工艺、大规模双极板CVI(Chemical Vapor Infiltration化学气相渗透法)致密化工艺。本发明通过设计和优化每一步工艺，获得了低成本超薄高强高电导C/C复合材料双极板，其具有良好的强度、弹性和导热，密度<2.2g/cm³、弯曲强度超过150MPa、腐蚀电流密度<0.016mA/cm²、体积电导率达到300S/cm、面电阻<0.01Ωcm²、平均气体透过率<2.0e‑6cm³/s cm²(△P＝2atm)。 【EN】The invention discloses a method for preparing an ultrathin carbon/carbon composite bipolar plate, which adopts an ultrafine carbon fiber mesh blank forming process, a graphene modified resin/asphalt slurry dispersing technology, a carbon fiber mesh blank/slurry flow field calendaring forming process and a large-scale bipolar plate CVI (Chemical Vapor Infiltration) densification process. The invention obtains the low-cost ultrathin high-strength high-conductivity C/C composite material bipolar plate by designing and optimizing each step of process, and the bipolar plate has good strength, elasticity, heat conduction and density<2.2g/cm³Bending strength of over 150MPa, corrosion current density<0.016mA/cm²The volume conductivity reaches 300S/cm and the surface resistance<0.01Ωcm²Average gas permeability<2.0e‑6cm³/s cm²(△P＝2atm)。

摘要：【中文】本发明涉及一种钢瓶旋转换向装置，其特征在于包括机架，气缸固定设置在机架顶端，芯轴的上端固定连接在气缸的活塞杆端，芯轴外套同轴套装在芯轴外并可转动连接芯轴；上安装板固定设置在机架的顶端，下安装板固定设置在上安装板下方，上导向套及下导向套上下设置并同轴套装在芯轴外套外，上导向套的上端固定连接上安装板，下导向套的下端固定连接下安装板，上导向套与下导向套之间的轴向间隙构成若干个“人”字形的连通的导向槽；芯轴外套两侧对称设置有2个径向的销钉，销钉延伸进入所述导向槽；芯轴外套的下端连接钢瓶抓取机构。本发明结构简单紧凑，制造成本低，操作方便快捷，快速实现钢瓶旋转换向。 【EN】The invention relates to a steel cylinder rotary reversing device which is characterized by comprising a rack, wherein a cylinder is fixedly arranged at the top end of the rack, the upper end of a mandrel is fixedly connected with the end of a piston rod of the cylinder, and a mandrel outer sleeve is coaxially sleeved outside the mandrel and can be rotatably connected with the mandrel; the upper mounting plate is fixedly arranged at the top end of the rack, the lower mounting plate is fixedly arranged below the upper mounting plate, the upper guide sleeve and the lower guide sleeve are vertically arranged and coaxially sleeved outside the mandrel outer sleeve, the upper end of the upper guide sleeve is fixedly connected with the upper mounting plate, the lower end of the lower guide sleeve is fixedly connected with the lower mounting plate, and a plurality of herringbone communicated guide grooves are formed in axial gaps between the upper guide sleeve and the lower guide sleeve; 2 radial pins are symmetrically arranged on two sides of the mandrel outer sleeve and extend into the guide groove; the lower end of the mandrel outer sleeve is connected with a steel cylinder grabbing mechanism. The invention has simple and compact structure, low manufacturing cost, convenient and fast operation and fast realization of the rotation reversing of the steel cylinder.

摘要：【中文】本发明公开了一种基于全卷积神经网络的图像裂缝分割方法，大致包括数据集的构建、模型训练和结果预测三个阶段，利用训练好的全卷积深度学习模型对待检测的图像进行语义分割，完成对图像中裂缝区域的特征提取，识别图像中的裂缝。本发明方法中全卷积深度学习模型的设计构思为：以VGG16网络为基础，将其最后的全连接层替换为卷积层，使用上采样操作将低分辨率的特征图恢复为高分辨率的特征图，并利用跳级结构融合多层深度特征,以提高裂缝分割精度。本发明方法节省人力成本，并排除人为主观因素的干扰，具有效率高，准确率高，实用性强的优点。 【EN】The invention discloses an image crack segmentation method based on a full convolution neural network, which roughly comprises three stages of data set construction, model training and result prediction, and the trained full convolution deep learning model is utilized to carry out semantic segmentation on an image to be detected, so that the feature extraction of a crack region in the image is completed, and cracks in the image are identified. The design concept of the full convolution deep learning model in the method is as follows: based on the VGG16 network, the last full connection layer is replaced by a convolutional layer, the low-resolution feature map is restored to be a high-resolution feature map by using an up-sampling operation, and a multi-layer depth feature is fused by using a skip level structure so as to improve the fracture segmentation precision. The method saves labor cost, eliminates interference of artificial subjective factors, and has the advantages of high efficiency, high accuracy and strong practicability.

摘要：【中文】本发明提供一种高效清除细胞培养过程中微生物污染的组合物及方法，该组合物由喹诺酮类抗生素、氨基糖苷类抗生素和至少两种抗菌剂组成；本发明提供的组合物及方法在不灭活、无需判断污染源且保证细胞活性的基础上，可强效清除大部分革兰氏阳性菌、革兰氏阴性菌、真菌、支原体，解除细胞系污染状态并使其能够再次被利用。 【EN】The invention provides a composition and a method for efficiently eliminating microbial contamination in a cell culture process, wherein the composition consists of quinolone antibiotics, aminoglycoside antibiotics and at least two antibacterial agents; the composition and the method provided by the invention can strongly remove most gram-positive bacteria, gram-negative bacteria, fungi and mycoplasma without inactivation, judging a pollution source and ensuring the activity of cells, and can remove the pollution state of a cell line and ensure that the cell line can be reused.

摘要：【中文】本发明涉及一种多工位气瓶翻转机，其特征在于包括机架，翻转架中部的两端设有转轴并可转动连接机架，翻转链轮固定连接翻转架一端的转轴，双轴磁性气缸水平设置在翻转链轮的下方并固定连接机架，链条连接架固定设置在双轴磁性气缸的缸体上，链条连接架的两端设有2个活接螺栓，链条绕过翻转链轮及2个过渡链轮并两端连接活接螺栓；翻转架中部设有若干个水平方向均匀分布的气瓶定位座，气瓶定位座的一侧设有与气瓶定位座数量相对应的用于水平方向夹紧气瓶的夹瓶机构；翻转架的上部设有用于垂直方向压紧气瓶的压瓶机构。本发明操作方便快捷，快速为多个气瓶翻转倒水，减轻操作工人的劳动强度，提高工作效率，降低生产成本。 【EN】The invention relates to a multi-station gas cylinder overturning machine which is characterized by comprising a rack, wherein rotating shafts are arranged at two ends of the middle part of an overturning frame and are rotatably connected with the rack, an overturning chain wheel is fixedly connected with the rotating shaft at one end of the overturning frame, a double-shaft magnetic cylinder is horizontally arranged below the overturning chain wheel and is fixedly connected with the rack, a chain connecting frame is fixedly arranged on a cylinder body of the double-shaft magnetic cylinder, 2 loose joint bolts are arranged at two ends of the chain connecting frame, and a chain bypasses the overturning chain wheel and 2 transition chain wheels and is connected with the loose joint bolts at two ends; the middle part of the roll-over stand is provided with a plurality of gas cylinder positioning seats which are uniformly distributed in the horizontal direction, and one side of each gas cylinder positioning seat is provided with a gas cylinder clamping mechanism which is used for clamping the gas cylinders in the horizontal direction and corresponds to the number of the gas cylinder positioning seats; the upper part of the roll-over stand is provided with a bottle pressing mechanism for pressing the gas bottle in the vertical direction. The turnover water pouring device is convenient and quick to operate, can be used for quickly turning over and pouring water for a plurality of gas cylinders, reduces the labor intensity of operators, improves the working efficiency and reduces the production cost.