Optimization of Predictive Maintenance Solutions across Companies
Leading manufacturers and machine providers can use advanced analytics and AI to predict machine failures and improve predictive maintenance solutions. Via the Apheris Platform, data from multiple companies can be leveraged to train models without putting their trade secrets and intellectual property at risk.
Situation
An aerospace manufacturer produces critical components of new airplanes. The production and assembly of the individual parts is carried out by sophisticated and expensive machines. The process is constantly monitored and data of different process parameters, performance parameters and undesired failure events are collected. Reliable and accurate models that predict failure events and enable better predictive maintenance solutions can save millions of dollars.
Problem
The manufacturer needs large amounts of data to create robust prediction models that provide actionable insights. The machine provider is only able to offer useful models if multiple customers pool their machine usage data so that the training dataset is sufficiently large. The other customers, however, are the manufacturer’s competitors. None are willing to share their machine usage data because this would allow competitors to deduce when there were downtimes, which parts were produced, and reveal insights into the utilization of the machines. The data represents trade secrets and intellectual property of the manufacturer, and hence cannot be pooled with other parties’ data.
Apheris solution
The Apheris Platform for federated and privacy preserving data science enables training of robust models for machine failure prediction on data from multiple manufacturers without pooling the data. Machine providers can offer superior products without putting their customer’s secrets at risk as the intellectual property and privacy of each manufacturers’ proprietary data is fully preserved.
Advantages of using Apheris
Reduce risk of machine downtime
Predictive maintenance solutions that rely on high-quality models reduce the risk of machine downtime and save millions of dollars
Improve product quality
Superior predictive maintenance solutions that are integrated into devices are adding significant value to the machine provider’s products and are likely to become the future standard
Protect intellectual property
The Apheris Platform allows manufacturers to benefit from superior predictive maintenance and significantly reduced costs, while protecting full data privacy and intellectual property
Improved Process Conditions along the Supply Chain
Modern supply chains consist of several suppliers working in circular networks. These can be optimized towards reduced resource inputs, increased revenue or reduced CO2 emissions. The Apheris Platform for federated and privacy preserving data science helps suppliers engage in data collaborations to develop new models and work towards their shared goals.
Situation
A supplier is dependent on two third-party suppliers. All three companies are attempting to use their in-house data to identify patterns for optimized resource usage and minimal CO2 emissions, but the resulting models are not applicable along the entire supply chain. The siloed optimization efforts lead to efficiency loss and suboptimal use of resources (e.g., truck queuing due to misaligned schedules between shipper and receiver).
Problem
Despite their common goal of optimized resource consumption and a minimized carbon footprint, the three suppliers do not pursue a joint approach for optimal processes along the supply chain. As their data contains valuable trade secrets, they refrain from openly sharing it. Diverse data sets representing the workflows along the entire supply chain are needed for AI model training and optimized supply chain processes, which would be beneficial for all parties involved.
Apheris solution
Using the Apheris Platform, the three suppliers can employ federated and privacy preserving data analyses to train machine learning models. Each company connects its data to the Platform where shared models are developed to compute on all datasets, without the underlying data being revealed. These models then generate key insights which can be used to optimize resource usage and reduce CO2 emissions. For example, a shipping supplier would be able to better understand when to schedule trucks, reduce idle times, and make positive changes to their business.
Advantages of using Apheris
Reduce CO2 emission
The suppliers are future-proofed and able to meet increasing regulatory and customer demands with regards to emissions. They can propose this as an additional value proposition
Improve processes along the supply chain
Optimized processes along the supply chain provide saving potentials for each supplier (e.g., by reduced trucking idle times and more load balanced operations)
Comply with regulation
Governmental regulations on resource consumption and CO2 emission can be implemented without restricting revenue streams
Improving Workplace Safety within Facilities
Workplace safety procedures and applications are indispensable for high-risk workplace environments to protect employees, machinery and materials. Companies with multiple facilities need to ensure that their unified security standards are met in each facility. By employing AI models based on video data from cameras within the facilities, this can be done accurately and efficiently. The Apheris Platform for federated and privacy preserving data science enables companies to leverage and combine the facilities’ video data in a compliant manner, while protecting sensitive employee data.
Situation
A company running three production sites across continents has recently made investments into internet connected cameras to improve safety procedures and reduce accident-caused employee compensations and machinery down-times. They use camera video data to train AI models for hazard recognition (e.g., misplaced hazardous materials such as oil barrels).
Problem
To train robust AI models that reliably detect critical workplace situations and set preventative rules, a large set of video data is needed which cannot be provided by each site separately. Combining the data from several sites is restricted because of compliance and regulatory requirements related to data that contains personally identifiable information of employees.
Apheris solution
The Apheris Platform for federated and privacy preserving data science allows for joint model training on video data from all facilities in a privacy preserving way. Therefore, reliable and accurate workplace safety applications can be developed and deployed within the facilities.
Advantages of using Apheris
Ensure legal compliance
By reducing workplace injuries, the manufacturing company improves employee health, minimizes production delays, and ensures legal compliance
Improve employee safety
Anomalous events are rapidly identified and safety procedures are continuously improved
Save time and money
Employee health & safety inspections are automated and standardized, saving time and money