Huawei cloud recruits partners in the field of industrial intelligence to provide strong support + commercial realization

Abstract ： Focusing on application development in the field of industrial intelligence ISV, Or equipment manufacturing enterprises seeking intelligent transformation 、 Enterprises in traditional industrial fields , Must not miss .

This article is shared from Huawei cloud community 《 Focus on industrial intelligence , Huawei cloud recruits partners , Help solve transformation challenges 》, author ： Huawei cloud community selection .

Facing the industrial field , Huawei cloud launched DevRun Industrial intelligence industry accelerator , Assist eco partners to integrate key technical capabilities , From a technical 、 resources 、 Business opportunities and other dimensions provide enterprises with full life cycle support , Help enterprises innovate and realize business .

Focusing on application development in the field of industrial intelligence ISV, Or equipment manufacturing enterprises seeking intelligent transformation 、 Enterprises in traditional industrial fields , Must not miss .

Focus on the intelligent scene in the industrial field , Solve transformation problems

The accelerator mainly revolves around Industrial data intelligence as well as Industrial visual intelligence The next few scenes , Provide technology empowerment for enterprises , Help enterprises in the industrial field reduce operation and maintenance costs , Increase of efficiency , Help intelligent transformation .

Industrial data intelligence

＜ Predictive maintenance ＞

Predictive maintenance of the plant can be repaired before equipment failure according to the data received , Avoid serious losses caused by unplanned downtime .

In predictive maintenance , Accelerator provides a general algorithm framework for predictive maintenance CauchyMind Mechanism model , It can be used to detect various signals （ Vibration 、 Electric frequency, etc ） Data processing , Carry out fault feature identification , Combined with the degenerate algorithm , Realize maintenance cycle prediction .

CauchyMind Key technology of mechanism model

CauchyMind The mechanism model was led by chief scientist Hua Weiyun 10 Many doctors of mechanical engineering and algorithms Joint research and development , It has made a breakthrough technological innovation in signal quality judgment and feature extraction module .

Signal quality judgment ： Envelope variation coefficient method 、 Adaptive sliding window average kurtosis method

feature extraction ：

• The working conditions are unified ： Keyless adaptive speed estimation and order tracking
• Early injury ： Harmonic product spectrum technique
• Compact extraction ： Sparse measures represent
• Resonant frequency band ： Dynamic Bayesian wavelet transform
• Monotony is separable ： Construction of generalized adaptive health index

Case study

The annual output of a steel plant is estimated to be millions of tons , If the pre finishing mill 、 Unplanned shutdown of equipment such as cone box , The whole production line shall be shut down , It will take at least several hours to shut down for maintenance , Estimate the value of the affected production 248 ten thousand , In case of safety accident , The loss is immeasurable .

Take mushroom Internet, a partner of Huawei's cloud fertile land start-up program, as an example , They integrate CauchyMind Predictive maintenance solution created by mechanism model , Help an iron and steel plant predict the faults of several rolling equipment in advance and give an alarm , The customer repaired the fault in time . The offline verification phase , Before finding the problem manually 7 The system has generated a prediction alarm for days .

＜ Optimization of process parameters ＞

Industrial intelligent industry accelerator provides process parameter optimization related AI Technical ability , Can help reduce trial production costs , Improve device processing quality .

key technology

• Orthogonal experiment ： According to the orthogonality, the representative process parameters are selected from the overall test to carry out the test . Representative experiments have “ Evenly distributed , Neat and comparable ” Characteristics , Effectively reduce the number of experiments .
• Build a model ： Based on Kriging / Surface response algorithm , Orthogonal experimental data , The fitting equation between the process parameters and the optimization objective is calculated （ The model ）.
• Multi objective optimization algorithm ： Based on multi-objective genetic algorithm NSGA-II/SPEA2, Search out the optimal process parameter set , Can improve productivity , Improve device processing quality , Reduce tool wear .

Case study

In the establishment of new devices for a traditional CNC machine tool enterprise / New machine tool processing parameters optimization model process , By way of AI Technology and numerical control system , Realize intelligent adjustment of various process parameters , Increased productivity 26%. Process parameter optimization of incubated digital machine tools Root model , Support rapid expansion , The process optimization model for a specific machine tool class can be formed by adding orthogonal experiments .

besides ,CauchyMind The mechanism model supports Supply chain optimization , Energy saving optimization Such as the scene , If ISV For example, the accuracy of the algorithm needs to be improved , The proportion of negative samples is too large ,AI Model development is difficult , Click the link   Join the industrial intelligence industry accelerator , Hua Wei Yun Provide core algorithm model , Joint delivery solutions , Solve technical problems , Achieve business breakthrough .

Industrial visual intelligence

Accelerators focus on industrial visual intelligence , Focus on Visual inspection 、 Intelligent sorting Such as the scene , Provide Huawei cloud industrial vision solutions for enterprises .

key technology

• Cloud side collaborative development ： The cloud provides AI Model & application development 、 Operation management capability , On the side AI Reasoning services , The end side provides business logic arrangement and debugging capabilities , And cooperate with cloud edge , Realize data reporting 、 It's hard to excavate 、 Model iteration .
• Zero code AI workflow ： For specific industrial visual inspection scenarios , Provide zero code AI Development workflow ：
• Arrange the station detection application business logic and parameters as required ： Through cloud platform 、 The end side software configures the application logic and parameters of each station hook .

Case study

Visual inspection scene ： Support steel surfaces 、 Automobile appearance assembly 、 Defect of auto parts 、PCB Device defect 、 Dozens of industrial visual inspection scenes, such as high-speed rail underbody bolt inspection , Take automobile appearance assembly quality inspection as an example ,AI Can quickly identify the wheel hub 、 Auto Logos 、 Models of headlamp and other assembly parts , The recognition accuracy is as high as 99%, Recognition speed is lower than 300ms.

Intelligent sorting scenario ： Take the brake disc sorting as an example , An auto parts manufacturer mainly depends on stamping steel marks on the brake disc to distinguish different models . Steel stamp printing 、 The identification and erasure process is complex , The identification process requires manual intervention , Affect productivity . Use deep learning techniques , The model can be determined according to different appearance , Identify multiple types , Average recognition accuracy >98%, Recognition speed <100ms.

The accelerator will be ISV Provide the end edge cloud collaborative development capability , Help solve the problem of AI The model is not universal , Difficult technical problems such as continuous iteration , Improve the accuracy of model recognition , Improve the competitiveness of solutions .

The one-stop workbench brings better application services to the industrial field

The industrial intelligent industry accelerator will combine the industrial software capabilities of ecological partners , For industrial enterprises aPaaS The workbench IMC（Intelligent Manufacturing Cloud）.

IMC It can provide R & D and design for enterprises 、 manufacturing 、 Marketing and daily office services , Realize one-stop work experience .

It mainly includes

• Industrial cloud mall ： Buy apps , Support one click purchase of industrial applications and solutions 、 Place an order .
• Enterprise workbench ： Use the app , Realize account opening between different applications , One account to access all industrial applications .
• Integrated workbench ： Connect applications , Provide enterprise unified services APIG, By applying the connector 、 Connection flow realizes inter application API Integrate , Support enterprises to quickly carry out data analysis services .
• Zero code platform ： Custom applications , Preset rich industrial application templates , Quickly customize lightweight industrial applications , Support drag and drop to develop visual reports .

Join now Accelerator , Enterprises can Get value for free 5 Million yuan of digital diagnosis services , Provide suggestions and personalized solutions for their digital transformation planning ; And enjoy Free data discount experience service .

Three rights and interests of industrial intelligent industry accelerator

Click the link Registration Accelerator , Become a cloud partner of Huawei .

Click to follow , The first time to learn about Huawei's new cloud technology ~