Two-dimensional Network Model Research Articles

Deep learning-assisted local resonance strategy for accurate internal damage imaging in composites

In this paper, we propose a deep neural network-assisted strategy to accurately and efficiently identify local defect resonance (LDR) modes and accurately image the internal damage in composites. A two-dimensional convolutional neural network (2D-CNN) model was constructed to identify LDR modes. The frequency-domain contour maps were used as input data, given that the LDR phenomenon exhibits discernible physical attributes in the frequency domain that are conducive to deep neural network assimilation. The obtained results demonstrate effective training outcomes and transferability, even with a limited number of samples. The LDR modes are efficiently extracted by the developed 2D-CNN model and used to obtain the accurate imaging of internal damages in composites.

Disrupting glioblastoma networks with tumor treating fields (TTFields) in in vitro models

PurposeThis study investigates the biological effect of Tumor Treating Fields (TTFields) on key drivers of glioblastoma’s malignancy—tumor microtube (TM) formation—and on the function and overall integrity of the tumor cell network.MethodUsing a two-dimensional monoculture GB cell network model (2DTM) of primary glioblastoma cell (GBC) cultures (S24, BG5 or T269), we evaluated the effects of TTFields on cell density, interconnectivity and structural integrity of the tumor network. We also analyzed calcium (Ca2+) transient dynamics and network morphology, validating findings in patient-derived tumoroids and brain tumor organoids.ResultsIn the 2DTM assay, TTFields reduced cell density by 85–88% and disrupted network interconnectivity, particularly in cells with multiple TMs. A “crooked TM” phenotype emerged in 5–6% of treated cells, rarely seen in controls. Ca2+ transients were significantly compromised, with global Ca2+ activity reduced by 51–83%, active and periodic cells by over 50%, and intercellular co-activity by 52% in S24, and almost completely in BG5 GBCs. The effects were more pronounced at 200 kHz compared to a 50 kHz TTFields. Similar reductions in Ca2+ activity were observed in patient-derived tumoroids. In brain tumor organoids, TTFields significantly reduced tumor cell proliferation and infiltration.ConclusionOur comprehensive study provides new insights into the multiple effects of Inovitro-modeled TTFields on glioma progression, morphology and network dynamics in vitro. Future in vivo studies to verify our in vitro findings may provide the basis for a deeper understanding and optimization of TTFields as a therapeutic modality in the treatment of GB.

Tipping points, multistability, and stochasticity in a two-dimensional traffic network dynamics

Mitigating traffic jams is a critical step for the betterment of the urban transportation system, which comprises a large number of interconnected routes to form an intricate network. To understand distinct features of vehicular traffic flow on a network, a macroscopic two-dimensional traffic network model is proposed incorporating intra-nodal and inter-nodal vehicular interaction. Utilizing the popular techniques of nonlinear dynamics, we investigate the impact of different parameters like occupancy, entry rates, and exit rates of vehicles. The existence of saddle-node, Hopf, homoclinic, Bogdanov–Takens, and cusp bifurcations have been shown using single or biparametric bifurcation diagrams. The occurrences of different multistability (bistability/tristability) phenomena, stochastic switching, and critical transitions are explored in detail. Further, we calculate the possibility of achieving each alternative state using the basin stability metric to characterize multistability. In addition, critical transitions from free flow to congestion are identified at different magnitudes of stochastic fluctuations. The applicability of critical slowing down based generic indicators, e.g., variance, lag-1 autocorrelation, skewness, kurtosis, and conditional heteroskedasticity are investigated to forewarn the critical transition from free flow to traffic congestion. It is demonstrated through the use of simulated data that not all of the measures exhibit sensitivity to rapid phase transitions in traffic flow. Our study reveals that traffic congestion emerges because of either bifurcation or stochasticity. The result provided in this study may serve as a paradigm to understand the qualitative behavior of traffic jams and to explore the tipping mechanisms occurring in transport phenomena.

A Bionic Localization Memristive Circuit Based on Spatial Cognitive Mechanisms of Hippocampus and Entorhinal Cortex.

In this article, a bionic localization memristive circuit is proposed, which mainly consists of head direction cell module, grid cell module, place cell module and decoding module. This work modifies the two-dimensional Continuous Attractor Network (CAN) model of grid cells into two one-dimensional models in X and Y directions. The head direction cell module utilizes memristors to integrate angular velocity and represents the real orientation of an agent. The grid cell module uses memristors to sense linear velocity and orientation signals, which are both self-motion cues, and encodes the position in space by firing in a periodic mode. The place cell module receives the grid cell module's output and fires in a specific position. The decoding module decodes the angle or place information and transfers the neuron state to a 'one-hot' code. This proposed circuit completes the localizing task in space and realizes in-memory computing due to the use of memristors, which can shorten the execution time. The functions mentioned above are implemented in LTSPICE. The simulation results show that the proposed circuit can realize path integration and localization. Moreover, it is shown that the proposed circuit has good robustness and low area overhead. This work provides a possible application idea in a prospective robot platform to help the robot localize and build maps.

Intelligent speech elderly rehabilitation learning assistance system based on deep learning and sensor networks

Recently, deep learning has been proved to significantly improve the quality of speech recognition. Convolutional neural networks are often used in speech recognition tasks because of their special network structure and powerful learning function. In order to solve the problem that the traditional convolutional neural network can not reflect the one-dimensional basic attributes of speech signal, this paper proposes to set the number of frames for one dimension of convolution kernel, and use one-dimensional model and two-dimensional convolutional network model for speech recognition. By moving the convolution kernel of time axis and frequency band, it can adapt to the time change of speech signal and maintain the correlation between frequency bands to a great extent. At the same time, this paper also discusses the speech signal preprocessing, feature parameter extraction and regularization algorithm. Due to the lack of hospital resources, the lag of information technology, the poor ability of accompanying and other reasons, the current accompanying service for elderly rehabilitation can not meet the needs of elderly patients. The rapid development and wide use of information technology provide opportunities for the optimization of health services for the elderly. In view of the difficulty of word memory and interpretation in current rehabilitation courses, a VR intelligent teaching system based on intelligent voice technology is developed. The application results show that the system can effectively improve the ability of language expression and word writing. At present, the system of intelligent speech function has not been completed, and lacks speech synthesis function. The next research will focus on the use of speech synthesis technology, in order to realize the man-machine dialogue between people and the system, and show a more real training situation.

A novel deep learning method based on 2-D CNNs and GRUs for permeability prediction of tight sandstone

The accurate calculation of tight sandstone reservoir permeability is crucial for optimizing production and maximizing natural gas recovery. Traditional physical model-based permeability calculation methods heavily rely on core experimental parameters. However, petrophysical experiments are costly, prompting the development of fast and reliable reservoir permeability prediction methods. In this paper, a novel deep learning model that combines two-dimensional convolutional neural network (2-D CNN) model and gated recurrent neural network is proposed to predict tight sandstone reservoir permeability. The 1-D conventional well logs data is converted into the 2-D geological feature map to fit the model's input dimension. The TimeDistributed layer wrapping technology is used to couple the 2-D CNN and GRU to extract comprehensive features of geological feature maps. By training the model, a depth nonlinear mapping between the spatial and temporal features in the well logs data and permeability is established. The model is divided into two subnetworks, 2-D CNNs and GRUs, which are used to compare and analyze the functions of different modules in the proposed model. Additionally, two traditional machine learning algorithms, SVM and XGBoost, are also introduced as comparison models. Four regression evaluation metrics are used to quantify the predictive performance of different models. The test results in a blind well demonstrate that the proposed method outperforms the comparison models in predicting permeability, with a coefficient of determination (R2) of 0.9305, a root mean square error (RMSE) of 0.0895, a mean squared error (MSE) of 0.0080, and a mean absolute error (MAE) of 0.0725. This model can provide insight into the petrophysical characteristics of reaction reservoirs based on conventional well logs data and can accurately estimate reservoir permeability under limited core data, which has great application potential in improving the exploration accuracy of unconventional oil and gas resources.

Nanoscale mapping of relativistic photocarrier transports in epitaxial graphene surface and edge states

We report the direct mapping of photo-transport properties of relativistic carriers generated on terrace-surface and terrace-edge structures of epitaxial graphene grown on silicon carbide. In this method, a conductive probe made a direct contact with epitaxial graphene and scanned the surface, allowing us to map the topography, current, and noise of graphene with nanoscale resolutions via a scanning noise microscopy. The maps were further analyzed by a two-dimensional network model to obtain the sheet-conductance (S□) and noise-source density (neff) distributions of epitaxial graphene. The topography image showed the formation of terraces, which could be attributed to the different decomposition rate of silicon atoms during the graphene growth. The terraces of graphene exhibited uniform S□, being separated by one-dimensional terrace-edges. Interestingly, terrace-edges exhibited a rather large neff with a power-law relationship of S□ ∝ neff−0.5, indicating a typical hopping conduction. However, in terrace-surfaces, sheet-conductance was nearly independent on neff, which was attributed to the relativistic nature of charge carriers unaffected by charge-trapping potential. Notably, under illumination, we observed the relativistic behavior of photocarriers in overall graphene terraces, presumably because photocarriers were generated through charge-transfer complexes formation at epitaxial graphene/substrate interface without the participation of noise-sources. Interestingly, we observed nearly homogeneous photoconductance and neff, which could be possibly due to the merger of one-dimensional edge-states into two-dimensional surface-states. The mapping of photoconductance and photo-traps provides valuable insights into the generation of relativistic photocarriers in epitaxial graphene.

Diagnostic performance of a deep-learning model using 18F-FDG PET/CT for evaluating recurrence after radiation therapy in patients with lung cancer.

We developed a deep learning model for distinguishing radiation therapy (RT)-related changes and tumour recurrence in patients with lung cancer who underwent RT, and evaluated its performance. We retrospectively recruited 308 patients with lung cancer with RT-related changes observed on 18F-fluorodeoxyglucose positron emission tomography-computed tomography (18F-FDG PET/CT) performed after RT. Patients were labelled as positive or negative for tumour recurrence through histologic diagnosis or clinical follow-up after 18F-FDG PET/CT. A two-dimensional (2D) slice-based convolutional neural network (CNN) model was created with a total of 3329 slices as input, and performance was evaluated with five independent test sets. For the five independent test sets, the area under the curve (AUC) of the receiver operating characteristic curve, sensitivity, and specificity were in the range of 0.98-0.99, 95-98%, and 87-95%, respectively. The region determined by the model was confirmed as an actual recurred tumour through the explainable artificial intelligence (AI) using gradient-weighted class activation mapping (Grad-CAM). The 2D slice-based CNN model using 18F-FDG PET imaging was able to distinguish well between RT-related changes and tumour recurrence in patients with lung cancer.

Study of the Flow Field and Permeability Characteristics in the Goaf using the OpenPNM Package

The roof-caving step scale goaf behind the working face is sensitive to the region’s spontaneous combustion and gas concentration distribution, including many rock block cracks and holes. A severe deviation from the dynamics of fluids in porous media by representative element volume (REV), leading to the results of Computational Fluid Dynamics (CFD) simulation, has a significant error. A heterogeneous two-dimensional pore network model was established to simulate the goaf flow accurately. The network was first created using the simple cubic lattice in the OpenPNM package, and the spatial distribution of the “O-ring” bulking factor was mapped to the network. The bulking factor and Weibull distribution were combined to produce the size distribution of the pore and throat in the network. The constructed pore network model was performed with single-phase flow simulations. The study determined the pore structure parameters of the pore network through the goaf’s risked falling characteristics and described the flow field’s distribution characteristics in the goaf. The permeability coefficient increases as pore diameter, throat diameter, pore volume and throat volume increase and increases as throat length decreases. The correlation between throat volume and permeability coefficient is the highest, which indicates that the whole throat is the main control factor governing the air transport capacity in the goaf. These results may provide some guidelines for controlling thermodynamic disasters in the goaf.

A Two-Dimensional Planar Fracture Network Model for Broken Rock Mass Based on Packer Test and Fractal Dimension

Broken rock masses with the complexity and concealment widely exist in nature such as underground mine, collapse column, and zone. It is extremely difficult to model fracture networks and to simulate water diffusion for broken rock masses. To explore a reasonable fracture network model for broken rock masses, a new method for modeling a two-dimensional planar fracture network model is proposed in this paper. It includes packer test, empirical relationship, fractal width description, and symmetric expansion modeling. Then, the fluid-solid coupling is used to simulate the diffusion properties of water in the two-dimensional planar fracture network model. It is found that the diffusion velocities vmax and vmin do not appear in the fracture widths λmax and λmin. It indicates that the fracture widths λmax and λmin in the fracture network model for broken rock mass have little impact on the diffusion velocity. Furthermore, the fracture distribution pattern in the fracture network model is an important factor affecting the diffusion velocities vmax and vmin. The simulation results of water diffusion in the currently proposed model are almost consistent with the actual process of the packer test. Also, the validity of the two-dimensional planar fracture network model is verified by comparing the simulation results with the existing research.

A novel approach for one-step defect detection and depth estimation using sequenced thermal signal encoding

ABSTRACT Pulsed thermography is a technique of significant interest in non-destructive testing, particularly in defect detection and depth characterisation of composite materials. This study presents an innovative methodology for simultaneously detecting defects and estimating depth using a combination of sequenced thermal signal encoding and a two-dimensional convolution neural network (CNN) model. We compare the results of the proposed method with those obtained from the feed-forward neural network (FFNN), a one-dimensional CNN, and a long short-term memory recurrent neural network (LSTM-RNN). The findings demonstrate that the proposed approach exhibits superior accuracy and robustness compared to the benchmarks.

An AI-powered Acoustic Detection System Based on YAMNet for UAVs in Search and Rescue Operations

An emerging application of unmanned aerial vehicles (UAVs) is related to search and rescue operations, where drones are used to detect emergency events, such as people screaming, explosions, or gunshots. As a result, this kind of service can provide critical support and enable a faster response from the first responders to an area of interest. UAVs equipped with acoustic sensors constitute a beneficial surveillance system as opposed to visual sensors that face limitations due to lighting conditions or obstacles to the field of view. In this paper, an acoustic detection system (ADS) capable of identifying such scenarios is presented using a ReSpeaker 4 Microphone Array and a Raspberry Pi 4B. For this purpose, a custom two-dimensional convolutional neural network (CNN) model based on YAMNet architecture has been implemented. The model was trained on public audio datasets on targets related to search and rescue operations. Furthermore, an added class with mixup augmentation for overlapping events was created. The final system was adjusted to run in real-time conditions achieving an overall accuracy of 70.13% in indoor and 70.52% in outdoor environments, respectively, revealing the potential of the solution.

Estimating stellar parameters and identifying very metal-poor stars for low-resolution spectra (R∼ 200)

AbstractVery metal-poor (VMP, [Fe/H]<-2.0) stars serve as invaluable repositories of insights into the nature and evolution of the first-generation stars formed in the early galaxy. The upcoming China Space Station Telescope (CSST) will provide us with a large amount of spectral data that may contain plenty of VMP stars, and thus it is crucial to determine the stellar atmospheric parameters ($T_{\textrm{eff}}$,$\log$g, and [Fe/H]) for low-resolution spectra similar to the CSST spectra ($R\sim 200$). This study introduces a novel two-dimensional Convolutional Neural Network (CNN) model, comprised of three convolutional layers and two fully connected layers. The model’s proficiency is assessed in estimating stellar parameters, particularly metallicity, from low-resolution spectra ($R \sim 200$), with a specific focus on enhancing the search for VMP stars within the CSST spectral data. We mainly use 10 008 spectra of VMP stars from LAMOST DR3, and 16 638 spectra of non-VMP stars ([Fe/H]>-2.0) from LAMOST DR8 for the experiments and apply random forest and support vector machine methods to make comparisons. The resolution of all spectra is reduced to$R\sim200$to match the resolution of the CSST, followed by pre-processing and transformation into two-dimensional spectra for input into the CNN model. The validation and practicality of this model are also tested on the MARCS synthetic spectra. The results show that using the CNN model constructed in this paper, we obtain Mean Absolute Error (MAE) values of 99.40 K for$T_{\textrm{eff}}$, 0.22 dex for$\log$g, 0.14 dex for [Fe/H], and 0.26 dex for [C/Fe] on the test set. Besides, the CNN model can efficiently identify VMP stars with a precision rate of 94.77%, a recall rate of 93.73%, and an accuracy of 95.70%. This paper powerfully demonstrates the effectiveness of the proposed CNN model in estimating stellar parameters for low-resolution spectra ($R\sim200$) and recognizing VMP stars that are of interest for stellar population and galactic evolution work.

Engineering microcracks in MWCNT/elastomer bilayers for high-performance stretchable sensor development

Stretchable strain sensors in motion detection, health monitoring, and human-machine interfaces are limited by device sensitivity, linearity, hysteresis, stability, and reproducibility in addition to stretchability. Engineering defect structures in sensing material is an effective approach in modulating the material's physical properties, particularly those associated with mechanical responses. Here, we demonstrate that bilayers of carbon nanotubes deposited on an elastomer substrate are mechanically coupled. The microcrack size, density, and distribution in the nanotube thin film can be engineered through uniaxial tensile training to exhibit highly tunable and stable piezoresistive responses with sensitivity, linearity, range, and reproducibility. These responses far exceeding those in uniform metallic films, patterned structures, and composites. In addition, numerical analyses performed on a two-dimensional network model of the cracked nanotube film provide quantitative explanations of how crack configuration, and evolvement under strain, lead to the significant enhancements in stretchable sensor performance using current bilayer structures.

DeepSG2PPI: A Protein-Protein Interaction Prediction Method Based on Deep Learning.

Protein-protein interaction (PPI) plays an important role in almost all life activities. Many protein interaction sites have been confirmed by biological experiments, but these PPI site identification methods are time-consuming and expensive. In this study, a deep learning-based PPI prediction method, named DeepSG2PPI, is developed. Firstly, the protein sequence information is retrieved and the local context information of each amino acid residue is calculated. A two-dimensional convolutional neural network (2D-CNN) model is employed to extract features from a two-channel coding structure, in which an attention mechanism is embedded to assign higher weights to key features. Secondly, the global statistical information of each amino acid residue and the relationship graph between the protein and GO (Gene Ontology) function annotation are built, and the graph embedding vector is constructed to represent the biological features of the protein. Finally, a 2D-CNN model and two 1D-CNN models are combined for PPI prediction. The comparison analysis with existing algorithms shows that the DeepSG2PPI method has better performance. It provides more accurate and effective PPI site prediction, which will be helpful in reducing the cost and failure rate of biological experiments.

A light-weight neural network for marine acoustic signal recognition suitable for fiber-optic hydrophones

Fiber-optic hydrophones based on optical fiber sensing and communication technology have developed rapidly in recent years. With the advantages of high sensitivity, small size, and large dynamic range, they are easy to detect underwater acoustic signals, showing great potential in the field of marine ecological protection. Researchers can monitor the marine ecological environment by recognizing underwater optical fiber sensing signals collected by fiber-optic hydrophones in real-time. Deep learning models can achieve high recognition accuracy, but they usually have large amounts of parameters that makes them difficult to be deployed in mobile terminals of marine survey ships. To solve the key problem, this paper discussed the design of the Deep Light-Weight Attention Residual Network (DLA-ResNet). It introduces the depth-wise separable convolution (DSC) and the convolutional block attention module (CBAM) into ResNet18 to make its performance better. ResNet18 is a two-dimensional convolutional network (CNN) model applying residual blocks. Experiments verify that the accuracy of DLA-ResNet improved from ResNet18 is increased by about 2.04% and its size is reduced by 85.67%. The proposed scheme in this paper realizes an average recognition rate of 96.50% for eight types of underwater optical fiber sensing signals in the application of marine ecological environment monitoring. Moreover, its model size is only 6.12 MB and it only takes 27.90 ms, which is available to meet the needs of reliability and effectiveness.

Development and Evaluation of Deep Learning Models for Automated Estimation of Myelin Maturation Using Pediatric Brain MRI Scans.

To predict the corresponding age of myelin maturation from brain MRI scans in infants and young children by using a deep learning algorithm and to build upon previously published models. Brain MRI scans acquired between January 1, 2011, and March 17, 2021, in our institution in patients aged 0-3 years were retrospectively retrieved from the archive. An ensemble of two-dimensional (2D) and three-dimensional (3D) convolutional neural network models was trained and internally validated in 710 patients to predict myelin maturation age on the basis of radiologist-generated labels. The model ensemble was tested on an internal dataset of 123 patients and two external datasets of 226 (0-25 months of age) and 383 (0-2 months of age) healthy children and infants, respectively. Mean absolute error (MAE) and Pearson correlation coefficients were used to assess model performance. The 2D, 3D, and 2D-plus-3D ensemble models showed MAE values of 1.43, 2.55, and 1.77 months, respectively, on the internal test set, values of 2.26, 2.27, and 1.22 months on the first external test set, and values of 0.44, 0.27, and 0.31 months on the second external test set. The ensemble model outperformed the previous state-of-the-art model on the same external test set (MAE = 1.22 vs 2.09 months). The proposed deep learning model accurately predicted myelin maturation age using pediatric brain MRI scans and may help reduce the time needed to complete this task, as well as interobserver variability in radiologist predictions.Keywords: Pediatrics, MR Imaging, CNS, Brain/Brain Stem, Convolutional Neural Network (CNN), Artificial Intelligence, Pediatric Imaging, Myelin Maturation, Brain MRI, Neuroradiology Supplemental material is available for this article. © RSNA, 2023.

Gearbox Fault Diagnosis Method Based on Multidomain Information Fusion

Traditional methods of gearbox fault diagnosis rely heavily on manual experience. To address this problem, our study proposes a gearbox fault diagnosis method based on multidomain information fusion. An experimental platform consisting of a JZQ250 fixed-axis gearbox was built. An acceleration sensor was used to obtain the vibration signal of the gearbox. Singular value decomposition (SVD) was used to preprocess the signal in order to reduce noise, and the processed vibration signal was subjected to short-time Fourier transform to obtain a two-dimensional time–frequency map. A multidomain information fusion convolutional neural network (CNN) model was constructed. Channel 1 was a one-dimensional convolutional neural network (1DCNN) model that input a one-dimensional vibration signal, and channel 2 was a two-dimensional convolutional neural network (2DCNN) model that input short-time Fourier transform (STFT) time–frequency images. The feature vectors extracted using the two channels were then fused into feature vectors for input into the classification model. Finally, support vector machines (SVM) were used to identify and classify the fault types. The model training performance used multiple methods: training set, verification set, loss curve, accuracy curve and t-SNE visualization (t-SNE). Through experimental verification, the method proposed in this paper was compared with FFT-2DCNN, 1DCNN-SVM and 2DCNN-SVM in terms of gearbox fault recognition performance. The model proposed in this paper had the highest fault recognition accuracy (98.08%).

Crop classification methods and influencing factors of reusing historical samples based on 2D-CNN

ABSTRACT Crop classification is a crucial task in agricultural remote sensing, with the accuracy of such classification heavily relies on field sampling. Reusing historical samples can minimize the reliance on annual field sampling for crop classification. Although previous research primarily focused on the classification accuracy based on the reused historical samples, the underlying factors that influence the accuracy have not been adequately investigated. In this study, we employed a two-dimensional convolutional neural network (2D-CNN) model for crop classification reusing historical samples and investigated the factors influencing the classification accuracy. First, we calculated a normalized difference vegetation index (NDVI) time series from historical data to characterize crop growth patterns. Secondly, we assessed three different time-series construction methods. Subsequently, we used the 2D-CNN model to automatically extract abstract features of crop growth patterns. Finally, we designed various strategies for reusing historical samples to explore the influencing factors. Experiments were conducted in Kuitun, Xinjiang Uygur Autonomous Region, China, from 2016 to 2020, employing a long time series of Sentinel-2 images as remote sensing data. Our results indicated that optimal 2D-CNN models using an irregular satellite image time series (irSITS) outperformed random forest models in inter-annual classifications (the overall accuracies for 2016–2020 were 0.78, 0.61, 0.89, 0.89, and 0.70, respectively). In addition, we identified that the primary factors affecting classification accuracy were 1) the time-series construction method used; 2) the crop growth patterns; and 3) the sample diversity. By reusing historical samples and considering the factors influencing classification accuracy, this study provides valuable insights into high-quality crop classification mapping under limited field sample conditions.

Deep Learning Models for Efficient Detection of Electricity Fraud in Smart Grids

—Electricity theft is one of the biggest problems facing energy companies. These companies eventually claimed that the conventional methods for preventing electricity fraud were insufficient, which led to the creation of systems based on artificial intelligence to identify thieves among electricity consumers. In this paper, we propose a system based on deep learning to identify customers who have engaged in fraudulent activity on smart grids. We selected one-dimensional (1D) and twodimensional (2D) convolutional neural network models from deep learning models to achieve our objective. Also, we proposed a new method to fill in missing values in the data set. Our findings show that our models enhance the performance of systems that identify electricity thieves.