<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Elvina Gindullina, PhD on Medium]]></title>
        <description><![CDATA[Stories by Elvina Gindullina, PhD on Medium]]></description>
        <link>https://medium.com/@elvina.gi?source=rss-c7459d1475aa------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*M4ncKIv2cJEDK0hrlrFmcg.jpeg</url>
            <title>Stories by Elvina Gindullina, PhD on Medium</title>
            <link>https://medium.com/@elvina.gi?source=rss-c7459d1475aa------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 11:06:46 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@elvina.gi/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Benefits of Adapting Sampling Algorithms for energy harvesting sensing device]]></title>
            <link>https://medium.com/@elvina.gi/benefits-of-adapting-sampling-algorithms-for-energy-harvesting-sensing-device-c342a9869bd2?source=rss-c7459d1475aa------2</link>
            <guid isPermaLink="false">https://medium.com/p/c342a9869bd2</guid>
            <category><![CDATA[solar-energy]]></category>
            <dc:creator><![CDATA[Elvina Gindullina, PhD]]></dc:creator>
            <pubDate>Sat, 23 Jun 2018 12:19:34 GMT</pubDate>
            <atom:updated>2020-03-24T09:42:21.805Z</atom:updated>
            <content:encoded><![CDATA[<p>Adaptive sampling algorithms (ASA) serve to provide the estimation of the optimal sampling rate (frequencies) online. Applying of ASAs can benefit the operation of a sensor device in terms of providing sufficient data accuracy with minimization of energy consumption. We tested the data-driven adaptive sampling algorithm (<a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=1&amp;ved=0ahUKEwiw8K37yOnbAhUMQBQKHUrEBYwQFgg4MAA&amp;url=http%3A%2F%2Fwww.mdpi.com%2F1424-8220%2F17%2F11%2F2551%2Fpdf&amp;usg=AOvVaw0dVQKx5FH6nnfoBxvEUpX7">DDASA</a>) to examine the advantages of ASAs in the long term period for an energy-harvesting sensing device.</p><p>LoadSensing (further — LS) powered by a solar panel was chosen for our experiment. LS is a wireless data logger designed for a geotechnical industry that performs measurement and sends the measures via radio transmission to a gateway or concentrator.</p><p>The simulation of the LS operation was forecasted for one year period. LS simulator includes the following components: simulation of the hourly solar radiation in Barcelona (Spain), the solar panel output, LS energy consumption and LS state of the charge based on different energy consumption scenarios.</p><h3><strong>Solar irradiation</strong></h3><p>Amount of solar power that falls on the unit area is called “solar irradiation” [W/m²], that is determined by lots of parameters, for example, Earth’s geometry, location (longitude, latitude, etc..), terrain (elevation, surface orientation, inclination, shadows etc..), and properties of the atmosphere defined by gases, particles (solid and liquid) and clouds. Modelling of solar irradiation includes different combinations of various parameters mentioned above, as well as historical data records. The latter is applied in the most advanced ANN (artificial neural network)-based models and provides the most accurate prediction of solar irradiation. Although requires having solar energy and meteorological data.</p><p>We implemented two models (<a href="https://ieeexplore.ieee.org/document/6850501/">Astronomical</a> and <a href="https://ieeexplore.ieee.org/document/5545754/">Clear-Sky</a> models) and compared with the real data provided by <strong>IREC </strong>(Institut de Recerca en Energia de Catalunya) and <a href="https://power.larc.nasa.gov/data-access-viewer/">NASA</a> for a tilted solar panel (0°, 37°, 90°) oriented on the south. Both models are not as accurate as ANN-models, although meteorological and solar data records are not needed for these models, this provides flexibility and make it easily applicable to any location.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*LdwzFdeH0NIJVLnSV0xUZw.png" /><figcaption>Average daily sums of global irradiation for 0°</figcaption></figure><p>We calculate the sum of squared residuals in order to compare the models, that is the sum of squared differences between real data values and model estimates.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/726/1*YhVAH2lJXdDW5GTg8ksr4w.png" /><figcaption>Squared residuals values</figcaption></figure><p>The clear-sky model demonstrates higher accuracy. Moreover, this model is wider and includes such components as diffusion and reflection solar energy. Due to it, this model can be used to calculate the hourly solar irradiation that is the input for the solar panel.</p><h3>Photovoltaic panel power output</h3><p>Active area and technology of a solar panel, as well as solar irradiation on tilted panel and losses, define the solar panel power output. Lots of factors cause energy losses in a solar panel. For example, meteorological factors as shading can cause a reduction of potential solar power. Soiling is another factor that affects the solar panel performance. The power of the soiling effect is defined by the location, that can differ for desert or a city. Another group of factors that cause reduction of solar panel output are inner factors, such as inverter losses, and AC-DC cable losses.</p><p>Another factor that determines the effectiveness of a solar panel operation is a cell temperature. Raising of the cell temperature negatively affects the performance of a solar panel, which is defined by a heat dissipation capabilities of a solar panel. Due to it, we chose the formula, that takes into account all listed factors above: <a href="https://ieeexplore.ieee.org/document/8089345/"><em>P = η · A · G· [1–0.005(Tcell-25)]</em></a>, where <em>A</em> is a total solar panel area, η is photoelectric conversion efficiency, <em>G</em> is solar irradiation on the inclined surface of the solar panel and <em>Tc</em> is an operational temperature of a solar cell and, that is defined as <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=1&amp;cad=rja&amp;uact=8&amp;ved=0ahUKEwjr9ve-0OnbAhUrIcAKHZO2A2AQFggnMAA&amp;url=http%3A%2F%2Fwww.ht.energy.lth.se%2Ffileadmin%2Fht%2FKurser%2FMVK160%2F2012%2FGail-Angee_Migan.pdf&amp;usg=AOvVaw0-B85TMhb40VsoxX_WZF2S"><em>Tcell=Tair + 0.035 · G</em></a>.</p><p>We simulated the values of the power output for an amorphous silicon solar panel with size 138.8x90 mm and theoretical conversion efficiency — 12.7%, inclined on 37°(south), average losses due to shading, dust, wiring — 0,75. The hourly temperature was extracted for Barcelona for 1 year for 01.01.2017–31.12.2017 from <a href="http://www.soda-pro.com/"><em>SoDa database</em></a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nBLsc2atRI3YPDatsburgg.png" /><figcaption>Air temperature and PV power output</figcaption></figure><h3>State of the charge calculation</h3><p>To calculate batteries SoC we adopt and modify the model presented <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=2&amp;cad=rja&amp;uact=8&amp;ved=0ahUKEwiVwMCszOnbAhUFvhQKHQ9ICx8QFggwMAE&amp;url=https%3A%2F%2Fresearch.utwente.nl%2Ffiles%2F28201914%2FA_comprehensive_model_for_battery_State_of_Charge_prediction.pdf&amp;usg=AOvVaw0I_K9cb7omduOGI3uRDeq0">here</a>. The model is determined by charging energy, energy demand and capacity of the battery. Battery capacity is defined by nominal voltage and current. But it is necessary to stress, that the battery performance is affected by ageing, temperature variations, charge-discharge cycle. Therefore, it is a challenging task to calculate an accurate SoC. We enhanced the model by including such factors as battery ageing and temperature effect into the model.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/846/1*tmoOMq8rz1c9rvt32ws1lg.png" /></figure><p>The battery present in LS is LG18650B4 with nominal capacity 2600 mAh, nominal voltage 3.6 V and amount of 8. After 300 cycles the battery loses its capacity on 20%. Moreover, the battery capacity can vary from 59% (-20C) to 104% (&gt;40C), depending on the air temperature.</p><h3>LS energy consumption</h3><p>Ls has lots of different possible configurations that affect battery consumption. LS can be configured to different measure duty cycles: from once per 30 seconds to on measure per day. Here we consider the worst-case energy consumption scenario of LS, that includes: 3 sec of warming up (60 mA, 12 V), 3 sec of measurement (60 mA, 12 V), the transmission of 3 pulses (900 ms, 120 mA, 3.6 V each) and time between pulses (2 sec, 15 mA, 3.6 V). Background consumption between cycles is 30 µAh, 3.6 V. The total time for one cycle is 8.86 sec with the average current — 40 mA. If LS performs a complete cycle once per hour, then the total energy consumption for one hour is 2.84 W.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/834/1*84O0mab1o4waPrmZFFRRRQ.png" /><figcaption>Radio consumption profile</figcaption></figure><h3>Constant sampling rate vs. DDASA</h3><p>DDASA changes the sampling frequency based on the values of the collected data. Briefly, DDASA adjusts the sampling frequency based on the calculation of the Sigmond function. The complete description of the algorithm can be found <a href="http://www.mdpi.com/1424-8220/17/11/2551">here</a>. DDASA was tested and implemented for tiltmeter data, extracted from LS.</p><p>Tiltmeters are used to measure the surface inclination. One of the application areas is objects under construction (buildings, bridges etc..).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/742/1*znwD4m-MsvRBD1ess84k0g.png" /><figcaption>LS tiltmeter installation</figcaption></figure><p>There is exist the trade-off between the accuracy of the data and the battery state. Therefore, comparing the results for constant sampling rate we will take into account these parameters: number of transmitted packets, battery discharge events, total energy consumption and average energy level during a whole period. More precise data requires more frequent measurements that drains the battery faster. In particular, if the LS samples once per 6 hours, then the total energy consumption is 4141.9 W, while for DDASA is 2583.3 W, that is defined by the number of collected and transmitted packets: for the constant sampling rate, strategy number of transmitted packets is 1456, while for DDASA is 858. Constant sampling rate caused the battery to go out of charge much more frequent — 2689 vs. 795. In average, DDASA keeps the average level of the battery higher than a constant rate — 57.8 W vs. 38.1 W.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C9ahEEQNXsY02a75IZmCSg.png" /><figcaption>DDASA vs. Constant rate SoC</figcaption></figure><p>Implementation of DDASA algorithm demonstrated, that the algorithm can be improved in terms of energy consumption and accuracy. Improvements of the DDASA and more results for the topic is given <a href="https://onlinelibrary.wiley.com/doi/abs/10.1002/ett.3754">here</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c342a9869bd2" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>