Yoshito Nakashima

Machine learning techniques have attracted attention for lithology identification, which is fundamental for understanding geological structures. However, the efficacy of machine learning-based lithology identification, specifically in deep (several hundred meters) coastal aquifers composed predominantly of unconsolidated sediments, remains underexplored. Therefore, this study developed machine learning models based on decision trees, random forest (RF), logistic regression, and multilayer perceptron (MLP) algorithms that classify lithofacies using geophysical well logging data collected along a borehole located in Suruga Bay, Shizuoka Prefecture, Japan. Two to nine input features were selected for the seven cases to classify four lithofacies: “cobbles and gravel”, “sand with gravel”, “silt with gravel”, and “basaltic lava”. The models showed accuracy ranging from 73.9% to 90.4%, with the RF- and MLP-based models achieving accuracy of 90.4% and 90.1%, respectively, when trained using all available input features. This indicates their effectiveness for lithology identification in the target borehole. Feature importance analysis using the SHapley Additive exPlanations method suggested that natural gamma ray, widely used in groundwater exploration, served as a key input feature in many of the analyzed cases. The incorporation of other input features (i.e., electrical resistivity, P-wave velocity, density, and NMR-derived parameters) generally improved the prediction performance. The results demonstrate the applicability of machine learning algorithms for lithology identification in a coastal aquifer several hundred meters thick, consisting of unconsolidated alluvial sediments and lava flows. These findings suggest that machine-learning techniques have the potential to reduce the time, labor, and costs associated with hydrogeological surveys and groundwater exploration.

地殻には様々な流体が存在するが，塩水中の石油や天然ガスのように不混和2相を形成するものがある．最近，人類が排出したCO2による気候変動を緩和する方法として地下の帯水層にCO2を注入するCCS（Carbon Capture and Storage）が有効であるとされ，塩水と超臨界CO2から成る不混和2相流体の多孔質岩石中での流動に関する研究が行われている．この分野はこれまで地下水，石油，天然ガスなどの資源関連分野で研究がなされてきたが，流動メカニズムの基礎的研究の成果は他の地球科学分野にとっても有用である．特に，岩石に対する流体の濡れ性の違いや，流速の違いが流路に様々な影響を与える．水や塩水は岩石に対する濡れ性流体であるが，石油やガス（超臨界状態CO2も含む）は非濡れ性流体である．孔隙中では濡れ性流体と非濡れ性流体の間に生じる圧力差が不混和流体の流動を支配する．そのため，流体移動を支配する岩石固有の物性値としての浸透率は流動の実態を反映しない．不混和流体の流動に関する研究はCCSだけでなく，火山，地震などの災害分野，地下水，土木の分野，及び地球科学の学術分野にも有用な知見を与えるであろう．本稿では不混和2相流体流動に関する基礎事項と最近の研究動向を紹介する．

Single-sided time-domain proton magnetic resonance (MR) surface scanners are useful for non-destructive measurements of moisture. A mobile single-sided MR sensor unit employing two concentric ring magnets was constructed for the in situ quantification of moisture in slaked lime plaster used in the outer walls and roofs of heritage architecture. This sensor unit allows for non-destructive measurements of water proton relaxation 1.5 to 13.5 mm beneath the surfaces of walls and roofs. The following laboratory experiments on water-saturated porous slaked lime plaster samples were performed. (i) The porosity (water volume fraction) was measured in approximately eight minutes with a root mean square error of 1.9 vol%. (ii) The fundamental MR-related property (i.e. proton surface relaxivity) needed for the estimation of the pore size distribution was also measured successfully. (iii) The pore volume expansion due to frost damage was successfully detected as a significant change in the transverse relaxation time distribution. These results demonstrate that the portable MR surface scanner is a promising non-destructive testing tool for the maintenance and remediation of heritage architecture made of plaster.

For the purpose of the nondestructive measurements of hazardous heavy-element-containing substances in soil samples, the emerging technique, triple-exposure X-ray computed tomography (CT), is proposed. The use of triple exposure (i.e. three different acceleration voltages for the X-ray tube) improves the accuracy of the identification of the atomic number of a heavy element. With lead as the target element, virtual X-ray microtomography simulations utilizing a polychromatic X-ray source accelerated at 70, 100, and 140 kV were performed. CT images for the three voltages were reconstructed on a computer for clayey soil samples with heavy-element-containing natural minerals, namely galena (PbS), witherite (BaCO₃), stibnite (Sb₂S₃), sphalerite (ZnS), and magnetite (Fe₃O₄). Quantitative analysis of the reconstructed CT images was performed to obtain the following results. (i) Galena can be distinguished from other heavy minerals (i.e. witherite, stibnite, sphalerite, and magnetite). (ii) The lead content in the clayey soil can be estimated with reasonable accuracy. These successful results demonstrate that the triple-exposure CT technique is promising for the nondestructive characterization of soil samples contaminated with hazardous heavy elements such as lead.

To realize nondestructive in-situ quantification of intramuscular fat in fresh beef meat, a lightweight (5.0-kg) hand-held sensor that consists of a planar radio-frequency coil and a unilateral (i.e., single-sided) magnetic circuit was constructed as a subunit of a time-domain proton magnetic resonance (MR) scanner system. The MR scanner based on proton relaxometry enables the nondestructive quantification of the fat content of meat by using the difference in the spin-spin relaxation times (T2) of water molecules in muscle and of fat molecules. The investigation depth of the sensor unit was 7 mm, which is sufficient to probe the meat beneath a subcutaneous fat layer less than a few millimeters thick. The scanner was successfully applied in a laboratory at a meat temperature of 30.5°C to quantify the fat content at 30 locations in fresh beef samples. The required measurement time was 22 s for each location. Reasonable agreement with a root-mean-square error as small as 5.4 wt% was obtained for fat quantification compared with the conventional destructive food analysis (Soxhlet extraction method). Thus, the portable MR scanner with a hand-held sensor unit is a promising nondestructive and noninvasive tool for the in-situ fat quantification of fresh beef (e.g., carcasses) with thin subcutaneous fat layers at meat processing factories.

For the in-situ nondestructive fat quantification of fresh tuna meat, an original lightweight (5.7 kg) hand-held sensor that consists of a planar radio-frequency coil and a single-sided magnetic circuit was developed as a subunit of a time-domain proton magnetic resonance (MR) scanner system. The investigation depth of the sensor unit is 12 mm, which is sufficient to probe the meat section beneath thick skin with scales and the underlying subcutaneous fat layer of large fish such as tuna. The scanner was successfully applied in a laboratory to a fillet of a bluefin tuna (Thunnus thynnus) to measure meat sections 12 mm beneath the skin. The required measurement time was 100 s for each section. The results of MR scan at 11 locations on the fillet were compared with those of conventional destructive food analysis. Reasonable agreement with an error (root-mean-square residual) of as small as 1.8 wt% was obtained for fat quantification. The time-domain MR relaxometry for the same tuna fillet also allowed lean meat quantification with a small root-mean-square residual of 6.7 wt%.

We present a theoretical background for heavy element quantification through the intensive analysis of beam hardening (cupping artifacts) in X-ray computed tomography (CT) images. Cupping artifacts resulting from X-ray CT using a polychromatic X-ray source are quantitatively analyzed with an analytical solution for a cylindrical sample of a homogeneous aqueous solution/suspension containing a heavy element. The theoretical solution reveals that the severity of cupping artifacts is strongly dependent on the sample chemistry and the acceleration voltage of the X-ray tube. Careful analysis of this dependency enabled simultaneous determination of the atomic number and molar concentration of the heavy element within a particular estimation error range. Significant improvement in terms of the accuracy of determining the atomic number was achieved by employing double-exposure X-ray CT.

Moisture content in coal powder is a critical index because it controls the difficulty of coal handling and ultimately its combustion efficiency in coal-fired power plants. A unilateral proton nuclear magnetic resonance (NMR) scanner was applied in a laboratory to nondestructively quantify moisture content in powdered samples of brown and subbituminous coals. Due to the open geometry of the sensor unit, the scanner allows the nondestructive moisture quantification of a portion several millimeters below the surface of a large sample. Proton transverse relaxation was measured by the unilateral NMR scanner, and obtained moisture content values were compared with those by a conventional bilateral NMR apparatus. The comparison showed reasonable agreement with a root mean square residual of 4.1 wt.%H₂O for a moisture content range of > 12 wt.%H₂O (wet basis). This moisture-quantification performance by unilateral NMR is sufficient for the quality control of wet coal having a critical value of approximately 30 wt.%H₂O in terms of the handling property (flow-ability). Because unilateral NMR requires no specific sample preparation, such as sample insertion into tubes, the system is potentially applicable to in-situ nondestructive monitoring of the coal moisture in power plants to contribute to improved efficiency of handling and combustion.