Cycle Data Research Articles

Online data-driven battery life prediction and quick classification based on partial charging data within 10 min

Accurate online battery life prediction is critical for the health management of battery powered systems. This study develops a moving window-based method for in-situ battery life prediction and quick classification. Five features are extracted from the partial charging data within 10 min to indicate battery aging evolution. The machine learning techniques are used to connect the features and battery end of life (EOL), with the Gaussian process regression (GPR) and support vector machine (SVM) used to predict and classify battery life, respectively. The performance of the developed methods is validated based on experimental data of 121 battery cells. Results show that GPR predicts accurate battery EOL and knees with the root mean square errors and mean absolute percentage errors being within 100 cycles and 10 %, respectively. SVM classifies battery life quickly and accurately based on only one cycle's data, with the classification accuracy close to 92 %. In summary, the developed methods show comparable in-situ life prediction and classification accuracies to the benchmark that needs offline calibrations.

Early prediction of battery life by learning from both time-series and histogram data

Due to dynamic operating conditions, random user behaviors, and cell-to-cell variations, accurately predicting battery life is challenging, especially using information from only a few early cycles. This work proposes a data-driven battery early prediction pipeline using both time-series, measurement-related features, and usage-related histogram features. We first investigate the prediction performance of using these two feature sources individually, then two methods of systematically combining these two feature sources are devised. Additionally, four machine learning algorithms with different characteristics are applied to compare their performances on battery prognostic problems. We show that the prediction accuracy of using these two feature sources individually is comparable. Moreover, a systematic combination of these two features considerably improves the prediction performance in terms of accuracy and robustness, achieving excellent prediction results with a root mean square error of around 150 cycles using only the first 100 cycle's data. Finally, experimental data of different cell types and cycling conditions are used to verify the developed method's effectiveness and generality.

Non-specific effects and clusters of women with painful TMD responders and non-responders to LLLT: double-blind randomized clinical trial.

The aim was to analyze the non-specific effects (placebo, spontaneous remission, and regression to the mean) of the low-level laser therapy (LLLT) in women with myofascial pain (painful temporomandibular disorder (TMD)), as well as to differentiate between responders and non-responder clusters to active and placebo LLLT according to the anxiety levels, salivary cortisol, use of oral contraceptives, and premenstrual period. Sixty-four women diagnosed with myofascial pain (Research Diagnostic Criteria for Temporomandibular Disorders (RDC/TMD)) were included, divided into laser (n = 20), placebo group (n = 21), and 23 controls (without treatment (WT)). The LLLT applied was 780nm, masseter and temporal = 5J/cm2 (20mW-0.5W/cm2), and TMJ area = 7.5J/cm2 (30mW-0.8W/cm2), eight sessions, twice a week. The pain intensity (visual analogue scale (VAS)), anxiety (Beck Anxiety Inventory), salivary cortisol, and menstrual cycle's data at the baseline, T1-T8, and 30days after LLLT (follow-up) were evaluated. The laser group showed 80% of pain reduction, placebo 85%, and WT 43% in T8. Women with severe anxiety and at the premenstrual period did not reduce pain with any LLLT. Active and placebo LLLT had similar effectiveness during the treatment period; however, women with moderate anxiety, cortisol levels above 10ng/ml, and without contraceptive use maintain analgesia longer with active LLLT than placebo (follow-up 30days). Women with low levels of anxiety, salivary cortisol below 10ng/ml, and with contraceptive use showed the higher pain reduction. The analgesia promoted by LLLT in women with myofascial pain is a result of non-specific effects during the treatment period, although active LLLT is more effective in maintaining the analgesia after treatment (30days) for the cluster of women with moderate anxiety, salivary cortisol above 10ng/ml, and without contraceptive use.

Modelling natural drying of European beech (Fagus sylvatica L.) logs for energy based on meteorological data

ABSTRACTA key factor for improving the whole fuel wood supply chain is moisture management. Moisture affects material quality, transport costs and revenue. The aim of this study was to develop a drying model for piled European beech (Fagus sylvatica L.) logs based on meteorological data, employing the continuous weighing approach. Two drying cycles with three piles altogether were conducted. In the first one (February to October 2013) the moisture content of the pile dropped from 43.5% to 22.5%, whereas in the second one (December 2013 to March 2015), the moisture content of the 2 piles dropped from 42.1% and 40.5% to 20.6% and 19.8%, respectively. Model performance (R2 adj = 0.41) was validated by applying the model developed from the second cycle's data on the first cycle's data. Model's mean deviation in moisture content from the observed curve was −0.41% ± 0.72%. Economic benefits and environmental impact were highlighted. Calorific value increased from 2.65 to 3.57 kWh kg–1 from February to October 2013. Truckload volume utilization increased by 25.1% and transported calorific value by 48.0%. The number of truck trips for transportation was reduced from nine to seven.

Study on Universal Parameter of Power System for Urban Bus in Cold Region Based on Driving Cycles

In northern China, temperature is below zero degrees Celsius more than six months in a year. The buses lacked power and temperature in winter. In order to study the universal parameter of power system, 40 bus lines were selected. The bus lines came from the cities of Changchun, Jilin, Tonghua, and Songyuan. Driving cycle's data was collected by an experiment with the real urban bus in two years. After defining the kinematics fragments and characteristic parameter, 13660 divided kinematics fragments and 13 selected kinematic eigenvalues composed a matrix. After reducing the matrix by principal component analysis, the matrix has been reconstructed including the kinematics fragments and the corresponding principal component score. The 13660 kinematics fragments were divided into three categories by fuzzy C-means clustering. On this basis, abnormal data were removed by Pauta criterion. Then according to the characteristic parameter of kinematic eigenvalues, a model on engine power and driving cycles was built, which included the heating power in the winter. Finally, three sets of power-weight ratios were calculated by the weight and volume of the traditional local bus. The buses which are configured with this method have enough power to provide warm carriage.

Modelling of HEV Lithium-Ion High Voltage Battery Pack using Dynamic Data

A new approach is proposed for developing a model for High Voltage (HV) Lithium-ion (Li-Ion) battery pack using Hybrid Electrical Vehicle (HEV) drive cycles data. A variety of drive cycle's data have been collected from the test vehicle at various operating conditions. The equivalent electrical circuit for the HV battery is formulated using mathematical equations and discrete state space equations. The Open Circuit Voltage (OCV), Offset Voltage at Zero Load conditions (VZL), battery pack internal resistance (Rbatt) are considered as various components of the mathematical equations and battery pack RC filter states are represented using discrete state space equations at various battery temperature (Batt Temp) conditions (-70C to 450C). The parameter estimation problem is then considered as the problem of simultaneously estimating the parameters of all the coefficients formulated in the battery pack mathematical and state space equations. This is mathematically posed as a constrained optimization problem and a variant of genetic algorithm (GA) is used to solve it. Modelling is done using MATLAB®/Simulink® tools. The developed model is validated on actual vehicle drive cycle data. The results obtained by the proposed approach are closely matching with the actual battery response under different drive cycles. The distinction of the approach is that it does not call for any lab tests, additional instrumentation and cell level measurements yet serves the purpose of fair fidelity model for the design, analysis and control of the HEV powernet.