Japanese industrial electronics company DAIHEN Corp. makes products that are highly sensitive to environmental conditions during assembly, including dust and temperature changes. To automate the process of tracking conditions and pairing them with each product being assembled, the company has begun using environmental sensors and RFID technology, as well as leveraging edge-intelligence software from FogHorn Systems, at its facility in Osaka, Japan.
The solution monitors the conditions and status of each product under assembly, for greater work-in-progress (WIP) understanding. Since the system was installed last year, the manufacturer projects that it has reduced manual data-entry hours by approximately 1,800 hours annually. During the coming fiscal year, the firm plans to expand the system from 70 percent site coverage to 100 percent, and to deploy the technology to its other factories throughout Japan, after which it expects to save 5,000 man-hours annually. The solution is provided with FogHorn software and IT services from Energia Communications (Enecom).
Founded nearly a century ago, DAIHEN makes power products, industrial robots, welding machines and wireless power-transfer systems. There are multiple challenges when it comes to the manufacturing operations, the firm reports. The work is highly sensitive, not only to dust and temperature fluctuations, but also to moisture. The iron core used to make products must be stored in clean rooms at a consistent temperature and humidity, says Ichiro Yamano, DAIHEN’s executive officer for the Innovation Task Force team, and with a very low dust level. After the cores are used to build coils, he explains, “these coils are then transferred to a drying room to remove moisture from the materials.”
For DAIHEN, ensuring that those conditions met the requirements of both the company and regulators meant considerable manual labor had to be spent checking and recording temperatures. Additionally, there was no automated way to link those conditions with each product under assembly. The assembly process also lacked visibility into the status of each product, so that individual stations could properly anticipate the work they had flowing through that process toward them.
“DAIHEN wanted to reach a higher goal by creating a more automated, accurate and granular monitoring system for each material and fabrication process,” Yamano states. Thus, the company began seeking a system that not only automatically captured data, but also enabled collaboration among company teams.
The solution consists of wireless sensors deployed in specific areas, as well as RFID tags affixed to products. DAIHEN has declined to indicate what kind of RFID technology it is using or the technology provider it utilizes. The sensors capture environmental conditions and forward that data to a server, while the tags transmit their own unique ID numbers, linked to the products to which they are attached, and this information is then paired with the sensor data for each location. As the sensor and RFID data is captured, FogHorn’s Lightning Edge ML provides complex machine learning, says Yuta Endo, FogHorn’s VP, general manager and head of APAC operations.
DAIHEN installed an infrastructure of RFID readers to capture the location of WIP. RFID tags are applied to each product being assembled, and the unique ID number on each tag transmits data to readers to provide a WIP record of every item, based on its location within the assembly process.
The FogHorn Lightning solution is unique, the company reports, in that it manages data “on the edge,” identifies the content of value and then forwards only that useable data to the back-end software. This contrasts with traditional IoT solutions, Endo explains, in which big data goes to a server, where the software must then extract and analyze the valuable information. The FogHorn Lightning system is designed to reduce the level down to only that which is of greatest value. “If you process data at the edge,” he says, “it can immediately be used for operations.” In fact, analytics can be available in near-real time.
The real-time delivery of data is important not only for planning, but also for sharing information between assembly stations or teams. “All of this information is visualized and shared among teams” as it is collected, Yamano says, “so all teams can be in synch without having inaccurate, verbal-based communication.”
The project launched in April 2017, and the FogHorn condition-monitoring feature was fully deployed last May. Two months later, the RFID component for WIP was taken live. Initially, at the first plant, the company reports that it already sees increased accuracy of data, as well as labor reduction when it comes to data input. “Since implementation, DAIHEN expects to save an estimated 1,800 hours in manual logging processes” at the first facility, Yamano says. “DAIHEN can now make faster and more accurate business decisions with better intelligence of the factory manufacturing operation, with detailed, contextualized manufacturing process data.”