China
Africa
Explore chapters and articles related to this topic
Mobile Applications
Published in Saad Z. Asif, 5G Mobile Communications Concepts and Technologies, 2018
Engineering always brings challenges and heartburns and it is not surprising that Easypaisa gave the same to Telenor Pakistan. One fine evening in the year 2011, the system started to crumple. It was noted in application logs that the application component was unable to establish connection with the database machine. The database machine runs database workloads including online transaction processing and data warehousing. Although an important component, this database machine is a third-party box and not part of the Fundamo system.
A framework of developing machine learning models for facility life-cycle cost analysis
Published in Building Research & Information, 2020
Xinghua Gao, Pardis Pishdad-Bozorgi
The first and most challenging task of an LCC analysis for a building is to determine the economic effects of alternatives and to quantify these effects and express them in monetary amounts (Fuller, 2010). After the cost-related data are extracted from the building systems and stored in one database, machine learning techniques can be implemented on them to forecast each LCC component of a building. The overall process of deriving LCC components is illustrated in Figure 3. The raw data used for deriving the initial design and construction costs, utility costs, and O&M costs are extracted from multiple building systems (discussed in the section entitled ‘Data Requirements and Data Sources’). The data indexed in time order, utility consumption and O&M costs, are analysed using time series methods, and projections are made when necessary. For example, projections are made when there are missing values because sensors were not deployed in the past. Public statistics such as the historical inflation rate, utility price, and labour rate are incorporated into the analysis to calculate the monetary costs and to convert the costs to their present values. These present values of the LCC components are the target attributes of the LCC analysis machine learning models.
A process parameters selection approach for trade-off between energy consumption and polishing quality
Published in International Journal of Computer Integrated Manufacturing, 2018
Haidong Yang, Hongcheng Li, Chengjiu Zhu, Hua Fang, Jun Li
Figure 6 shows the conceptual framework of parameter selection procedure. First of all, based on the polishing machine database, machine parameters (e.g. fickerts size, lateral oscillation amplitude) are determined and regarded as the constant variables. Then, the energy flow characteristics of a polishing machine for defining the boundary of chip formation energy are analysed. According to polishing mechanism, four operational parameters are selected as control variables. Simultaneously, multi-objective optimisation model considering chip formation energy and polishing quality is built. To solve the multi-objective optimisation problem, the improved HGA which can be divided into low-level GA and top GA is adopted. Finally, the design of experiments (DOE) and sensitivity analysis are performed to evaluate the influence of operational parameters on optimal solution. The following work in this section is to introduce the implementation scheme of HGA in detail.
The future of in-field sports biomechanics: wearables plus modelling compute real-time in vivo tissue loading to prevent and repair musculoskeletal injuries
Published in Sports Biomechanics, 2021
Our recent Achilles tendon work is the best representation of real-time biofeedback of estimated tissue stresses and strains. Initially, subject-specific Achilles tendon FEA models from 3D freehand ultrasound were developed (Hansen et al., 2017; Shim et al., 2014, 2018, 2019). Tendon mechanical behaviour was represented as hyper-elastic (Hansen et al., 2017; Shim et al., 2014, 2019) or hyper-elastic fibre reinforced materials (Shim et al., 2018), which predicted physiologically plausible 3D tendon stresses and strains in isometric ankle plantar-flexion tasks. Furthermore, databases were generated encompassing a large range of physiological muscle forces that were applied to the Achilles tendon FEA models thereby producing corresponding 3D stress and strain distributions. From this database machine-learning models were generated that enabled real-time estimation of the Achilles tendon 3D stress and strain distributions from applied muscle forces determined by the real-time OpenSim and CEINMS models and laboratory motion capture data (Pizzolato et al., 2017b, 2017c, 2020). These 3D strains were finally visualised on a smart phone in real-time by linking the cloud-based calculations to the smart phone. To-date this technology is yet to be used for real-time biofeedback, so we are unsure if people can quickly adjust their movement and neuromuscular patterns to alter their Achilles tendon 3D strain distributions in the laboratory. In addition, studies are needed to ascertain the best biofeedback approaches (e.g., visual, haptic and/or auditory) to which people can respond (Booth et al., 2019; Dowling et al., 2010; Genthe et al., 2018; Hunt et al., 2011; Riskowski et al., 2009; Shull et al., 2013). Finally, more complex multi-scale tendon models integrating porous and viscoelastic soft tissue behaviour will raise our understanding of target strains for training, i.e., those which optimise tendon tissue-fluid dynamics and nutrient movement based on mechanobiological models of tendon adaptation (Mehdizadeh et al., 2017; Smith et al., 2013).