Why "Load Profiles" are key product reliability requirements

Do you know how many hours is your heating appliance working per day? Or how many starts it has per week? This type of information, called load profiles, is of extreme importance for EPM (Engineering Product Quality, Processes and Methods) since these are the ones responsible to breakdown the lifetime requirements of a specific product to its several components. One possible way to collect such load profiles is to log data from field appliances. This way the development team can get a reliable prediction of the appliance and components real loads.

A load profile describes the variation of a certain relevant parameter (such as water flow/temperature, tapping duration) over time. The relationship between the strength of the component and the operational load is key to determine the reliability of the overall appliance to perform its function.

In the past load profiles were mainly derived based on assumptions, experience, and worst-case scenarios leading to both over-engineering and/or early failures scenarios. As of today, they are built based on field and simulation data, with the target to get specific load profiles per component or product group over different sales regions.

Load profiles comprise important information to support the product development process such as for the quantification of the components design requirements to meet the loads avoiding overdesign, as well as for definition of adequate verification method and test conditions for our lifetime test procedures.

Statistical analysis techniques such as regression analysis, time series analysis, and machine learning can be used to analyze the data and extract useful insights. The preferred method for load profiles derivation is associated with a percentile which normally comprise 95% of the use cases referring to a certain load type. This 95% is the common value used to cover the real loads but also avoiding potential overdesign due to very extreme users. This 95th percentile must be determined separately for each individual load and results from the application of statistical methods to the field logged data. The figure below shows an example of a load profile for the number of starts of an appliance per day and the respective 95th percentile.

As previously mentioned, field tests are of paramount importance to improve the product's design. By exposing the appliance to a real-world environment, we are able to collect operational data and use it to not only fix any underlying issue that was not visible during lab testing but also use it as input for future component development. Heat pumps are one of the primary examples of such a procedure. As Bosch Thermotechnology (TT) became one of the key players in this business area across Europe, several field tests from different projects have been installed overtime in different countries. The picture on the right showcases the various installation locations where we collected appliance operational data.

Although field tests play a major role in the troubleshooting and diagnosis of design and control issues, the sheer volume of data that is needed to build load profiles makes field tests impractical as the sole source of data. This is where IoT (Internet of Things) enters the picture! Every heat pump that is sold comes installed with a connectivity solution allowing the possibility of remote data logging. This approach allows us to get data from every appliance sold - theoretically speaking of course. In case the customer opts-in and allows its data to be recorded (all legal and General Data Protection Regulation are addressed), we can now build load profiles from a much bigger sample size, which, in turn, provide a greater degree of confidence that our result is representative of the whole population behavior. Indeed, by providing as much accuracy as possible in the load profiles, the design process can be that much more precise, cost-effective and reliable.