Capturing this week's trending industry 4.0 and emerging industrial technology media
Part Level Demand Forecasting at Scale
The challenges of demand forecasting include ensuring the right granularity, timeliness, and fidelity of forecasts. Due to limitations in computing capability and the lack of know-how, forecasting is often performed at an aggregated level, reducing fidelity.
In this blog, we demonstrate how our Solution Accelerator for Part Level Demand Forecasting helps your organization to forecast at the part level, rather than at the aggregate level using the Databricks Lakehouse Platform. Part-level demand forecasting is especially important in discrete manufacturing where manufacturers are at the mercy of their supply chain. This is due to the fact that constituent parts of a discrete manufactured product (e.g. cars) are dependent on components provided by third-party original equipment manufacturers (OEMs). The goal is to map the forecasted demand values for each SKU to quantities of the raw materials (the input of the production line) that are needed to produce the associated finished product (the output of the production line).
Using graph neural networks to recommend related products
In experiments, we found that our approach outperformed state-of-the-art baselines by 30% to 160%, as measured by HitRate and mean reciprocal rank, both of which compare model predictions to actual customer co-purchases. We have begun to deploy this model in production.
The main difficulty with using graph neural networks (GNNs) to do related-product recommendation is that the relationships between products are asymmetric. It makes perfect sense to recommend a phone case to someone who’s buying a new phone but less sense to recommend a phone to someone who’s buying a case. We solve this problem by producing two embeddings of every graph node: one that characterizes its role as the source of a related-product recommendation and one that characterizes its role as the target. We also present a new loss function that encourages related-product recommendation (RPR) models to select products along outbound graph edges and discourages them from recommending products along inbound edges.
How Boeing overcame their on-premises implementation challenges with data & AI
Yield Is Top Issue For MicroLEDs
Early test results indicate yield issues at chip transfer, array-to-driver bonding, and other relatively new processes. High cost for this immature technology is keeping microLED displays from making the prototype-to-production leap. And because probers are not well suited to testing thousands of microLED pixels in densely packed arrays, DFT with self-testing is employed, which enables lifecycle testing — at ATE, post-assembly test, and in the field.
For instance, Dialog Semiconductor, a Renesas Company, developed a testing scheme for a white adaptive headlight module containing a 20,000-microLED array with 40µm pitch. “It’s a very good example of how a DFT circuit is not just overhead and cost to buy quality,” said Hans Martin von Staudt, director of Design-for-Test at Renesas. “Instead, it serves a valuable function over the lifetime of the chip. So we needed a DFT scheme with high-diagnostic coverage of the assembly process for pinpointing process weaknesses while enabling in-field monitoring.”
Inspection and testing methods are improving in their ability to identify and segregate out-of-spec product. Mass transfer methods that remove microLED die from wafers or film carriers and position them on IC drivers (for small AR/VR, watch and headlights) or TFT PCBs (for TVs), must easily separate known good die (KGD) from failures and underperforming die.
Yield targets for most microLED display apps are high (see figure 1) because the human eye can quickly spot missing pixels. To put yield targets in perspective, an 8K TV contains 99 million microLED chips. So if the defectivity rate is 0.5%, 520,000 devices must be removed and replaced. Top Engineering estimates this process would take 144 hours, making it cost-prohibitive until repair cost (removal and replacement of individual microLEDs) can be accelerated.
Keep It Simple and Shined with Automated Surface Finishing
Installing point source carbon capture on industrial sites
A lack of available physical space and the high costs have both been barriers to widespread deployment of carbon capture systems. However, one original equipment manufacturer (OEM) is focused on overcoming these barriers. Carbon Clean is a carbon capture solutions company headquartered in the UK that provides cost-effective carbon capture technologies for hard-to-abate industries such as cement, steel, energy from waste and refineries. Projects include working with CEMEX, a global leader in the building materials industry, on deploying CycloneCC at its cement plant in Victorville, California and Rüdersdorf plant in Germany. Carbon Clean is also seeking to develop a pilot using CycloneCC with Chevron on a gas turbine in San Joaquin Valley, Calif.
Cemvita Factory is an OEM that specializes in the biological conversion of CO2 into value-added products. Cemvita, which is based in Houston, focuses on offering “microbes-as-a-service” to potential clients who are interested in upgrading their CO2 into useful products. One notable Cemvita Factory project is their partnership with Oxy to convert 1.7 million tons per year of captured CO2 (from a cogeneration power plant) into 1 billion pounds per year of bioethylene.
Design of a Ni-based superalloy for laser repair applications using probabilistic neural network identification
A neural network framework is used to design a new Ni-based superalloy that surpasses the performance of IN718 for laser-blown-powder directed-energy-deposition repair applications. Current high-performance engineering alloys commonly suffer from issues when processed using additive manufacturing methods. These include cracking, porosity, elemental segregation, and anisotropy. The computational method reported here enables the identification of new alloy compositions that have the highest likelihood of simultaneously satisfying a range of target properties, including criteria specific to additive manufacturing. The efficacy of this method is demonstrated with the design of a new alloy more amenable to laser-blown-powder direct-energy-deposition. The method may be readily extended to the optimization of other alloy types and process methods.
Tracking this week's major mergers, partnerships, and funding events in manufacturing and supply chain
Katana Closes $35M Series B to Scale an Industry-Leading Manufacturing Ecosystem for SMBs
Katana, an industry-leading manufacturing ERP software has closed a $35 million Series B round, bringing the total funding to date to $51 million. Funds will be used to scale up Katana’s features for small to medium-sized manufacturers, allowing them to manage D2C and B2B sales in an intuitive, industry-leading software solution.
Foxglove's $15M Series A and the Missing Data Stack for Robotics
I’m happy to announce our $15M Series A funding led by Eclipse Ventures. It’s an exciting time for Foxglove – thousands of robotics engineers use our software, and our user base has grown by over 8x in the past 12 months alone! But before we jump into details, I want to spend a minute reflecting on our mission and why it is so important.
Foxglove is building the data stack for robotics. Our platform combines open-source robot logging, data lake management, and visualization to streamline common robotics development workflows, increase collaboration, and ultimately help companies get robots to market faster. Our products are used by thousands of engineers, product managers, and operations teams, and we’re transforming development at some of the top companies across a range of industries, from autonomous vehicles (NVIDIA), logistics (6 River Systems), and autonomous forklifts (Third Wave Automation), to lawn mowing, sidewalk delivery, undersea exploration, fulfillment, and defense.
Sustainable 3D print-on-demand fine jewellery service provider Cloud Factory raises €2m in seed funding
Cloud Factory, a precious metal 3D printing company, announced this week that it has raised a 2 million EUR seed round led by Charge Ventures, a leading early-stage fund in the Baltics. The company claims to be the world’s first company to use direct metal 3D printing to “scalably” manufacture fine jewellery in a cost-effective and sustainable way.
Cloud Factory aims to use the investment to increase its manufacturing capacities and further develop its online platform for product creation and order management integrations. It says that through introducing green tech to the jewellery industry, it has become the first zero waste jewellery factory in the world using resource-efficient 3D printing technology. The type of machine in use at the company is the SLM 125 metal 3D printer.
trinckle Scores €3M for 3D Printing Design Configuration Software
3D printing software startup trinckle allows customers to create configurators for customizable 3D printed goods. This enables businesses to save time by automating their workflows. Its Paramate software make it possible to parametrically design customizable orthotics, grippers, production jigs, copper inductors, surf fins, optics, and jewelry.
Now, the company has raised €3 million to beef up its efforts in this area. trinckle hopes to use the money to grow its team and roll out its software worldwide. Interestingly, the round was led by HZG Group the investment fund run by the founders of Concept Laser (now owned by GE Additive), Kerstin and Frank Carsten Herzog.
Siemens and Desktop Metal begin partnership with aim of accelerating sustainable additive manufacturing at scale
Siemens and Desktop Metal have announced a multi-faceted partnership aimed at accelerating the adoption of additive manufacturing for production applications with a focus on the world’s largest manufacturers.
The collaboration will touch multiple aspects of the Desktop Metal business. This includes increased integration of Siemens technology in Desktop Metal’s AM 2.0 systems, including operational technology, information technology and automation. Desktop Metal says its solutions will be fully integrated into Siemens simulation and planning tools for machine and factory design. Siemens Digital Twin tools will now be used for designing certain machines, and Siemens Advanta can simulate all levels of the binder jetting process and global plant planning, which Siemens says enables fast and reliable decisions for factory planning.
Mercedes-Benz and Microsoft collaborate to boost efficiency, resilience and sustainability in car production
Mercedes-Benz production staff gets access to the MO360 Data Platform via a self-service portal available on any company device including tablets, smartphones and laptops. Its visualization with Microsoft Power BI provides a what-you-see-is-what-you-get experience, allowing employees to become data workers with the ability to model and correlate data. The Teams Walkie Talkie app provides workers with an instant push-to-talk (PTT) communication on their business phones — no extra device needed.
With the MO360 Data Platform, teams at Factory 56 have shortened their daily shop floor meeting by 30%. In addition, they identify priority tasks to optimize production workflows within two minutes, which took up to four hours prior to the introduction of the platform. From team leads and process engineers to shop and plant managers, employees are encouraged to contribute new use cases to drive process innovation with Microsoft Power Platform.