49results about How to "Effective sorting" patented technology

The invention relates to a water-purifying process and water-purifying equipment thereof. Tap water is treated into flushing water, clean water and direct drinking water in a grading and classifying manner as required through a first-stage purification treatment device, a second-stage purification treatment device and a third-stage purification treatment device which are connected in sequence from bottom to top. By adopting the water-purifying process and the water-purifying equipment, water resources are utilized more optionally and reasonably, the problems of secondary pollution and the like since heavy metal pollutants in tap water cannot be removed in an efficient and environment-friendly way in the prior art are solved effectively, heavy metal ions, residual chlorine and other harmful particles in tap water are removed efficiently, bacteria are killed, the water quality is improved, sewage is disposed automatically, and the water yield and the utilization ratio of water are increased greatly; the health of a user is ensured; tap water is classified and treated effectively and is utilized in a shunting way according to the water quality, water amount and water pressure requirements of household water, and water is saved; energy is saved efficiently, space is saved, and remarkable energy-saving and environmental protection effects are achieved. The water-purifying process and the water-purifying equipment are suitable to be popularized and applied in relevant tap water treatment field.

The invention discloses a quality inspection method based on laser welding, comprising the following steps: collecting welding training information within a standard time period; dividing abnormal information in distribution into: abnormal training information in distribution and abnormal detection information in distribution; Carry out the marking process of quality classification with the abnormal training information in the distribution to obtain the marked welding information; construct the initial neural network, and train the initial neural network through the marked welding information to obtain a standard neural network; through the abnormal detection information in the distribution and The out-of-distribution abnormal information detects the standard neural network, extracts the standard neural network whose accuracy rate is greater than the accurate threshold, and obtains the target neural network; obtains the welding target information within the target time period, and inputs it into the target neural network for classification processing, and obtains the welding target neural network. The quality type of the target information. The invention can detect and classify the welding quality of products, and improves the efficiency and accuracy of quality detection.