In order to reduce the comprehensive power cost of the independent microgrid and to improve environmental protection and power supply reliability, a two-layer power capacity optimization model of a microgrid with electric vehicles (EVs) was established that considered uncertainty and demand response. Based on the load and energy storage characteristics of electric vehicles, the classification of electric vehicles was proposed, and their mathematical models were established. The idea of robust optimization was adopted to construct the uncertain scenario set. Considering the incentive demand response, a two-layer power capacity optimization model of a microgrid was constructed. The improved pelican optimization algorithm (IPOA) was proposed as the two-layer model. In view of the slow convergence rate of the pelican optimization algorithm (POA) and its tendency to fall into the local optimum, methods such as elite reverse learning were proposed to generate the initial population, set disturbance inhibitors, and introduce Lévy flight to improve the initial population of the algorithm and enhance its global search ability. Finally, an independent microgrid was used as an example to verify the effectiveness of the proposed model and the improved algorithm. Considering that the total power capacity optimization cost of the microgrid after addition of electric vehicles was reduced by CNY 139,600, the total power capacity optimization cost of the microgrid after IOPA optimization was reduced by CNY 49,600 compared with that after POA optimization.

Objectives Lymph node (LN) metastasis is a common cause of recurrence in oral cancer; however, the accuracy of distinguishing positive and negative LNs is not ideal. Here, we aimed to develop a deep learning model that can identify, locate, and distinguish LNs in contrast-enhanced CT (CECT) images with a higher accuracy. Methods The preoperative CECT images and corresponding postoperative pathological diagnoses of 1466 patients with oral cancer from our hospital were retrospectively collected. In stage I, full-layer images (five common anatomical structures) were labeled; in stage II, negative and positive LNs were separately labeled. The stage I model was innovatively employed for stage II training to improve accuracy with the idea of transfer learning (TL). The Mask R-CNN instance segmentation framework was selected for model construction and training. The accuracy of the model was compared with that of human observers. Results A total of 5412 images and 5601 images were labeled in stage I and II, respectively. The stage I model achieved an excellent segmentation effect in the test set (AP50-0.7249). The positive LN accuracy of the stage II TL model was similar to that of the radiologist and much higher than that of the surgeons and students (0.7042 vs. 0.7647 (p = 0.243), 0.4216 (p < 0.001), and 0.3629 (p < 0.001)). The clinical accuracy of the model was highest (0.8509 vs. 0.8000, 0.5500, 0.4500, and 0.6658 of the Radiology Department). Conclusions The model was constructed using a deep neural network and had high accuracy in LN localization and metastasis discrimination, which could contribute to accurate diagnosis and customized treatment planning. Key Points • Lymph node metastasis is not well recognized with modern medical imaging tools. • Transfer learning can improve the accuracy of deep learning model prediction. • Deep learning can aid the accurate identification of lymph node metastasis.

The Baeyer-Villiger Oxidation (BVO) of ketones and aldehydes produce lactones and formates, while aerobic carboxylation of aldehydes manufactures carboxylic acids, both having high added value. This work prepared a series of Al-containing silicates modified with organic ligands and SnO2 nanoparticles, which were then employed as catalyst in BVO and carboxylation. Characterizations revealed the morphology of the synthesized catalyst was changed from micron-sized thin sheets to smaller blocks, and then to uniform nanoparticles (size of 50 nm) having the doped SnO2 nanoparticles with a size of 29 nm. All catalysts showed high BET surface areas featuring silt-like mesopores. In determining the priority of BVO and carboxylation, an influence evaluation of the parameters showed the order to be substrate > oxidant > solvent > catalyst. Cyclic aliphatic ketones were suitable for BVO, but linear aliphatic and aromatic aldehydes for carboxylation. Coordination of (S)-binaphthol or doping of Sn into catalyst showed little influence on BVO under m-CPBA, but the Sn-doped catalyst largely increased BVO under (NH4)2S2O8 and H2O2. Calculations revealed that the catalyst containing both Al and Sn could give BVO intermediates lower energies than the Sn-beta zeolite model. The present system exhibited merits including wider substrate scope, innocuous catalytic metal, greener oxidant, as well as lower catalyst cost.

No abstract

Conventional optimization methods cannot fully satisfy the interests of multiparticipants and protect the privacy of participants in the integrated energy system that observe changes in the energy market structure. To allocate the benefits among the stakeholders in the integrated energy system and improve renewable energy accommodation, the manuscript proposes an optimal dispatching strategy for a park-level integrated energy system employing the Stackelberg game. Firstly, the benefits and cost models of each stakeholder of the integrated energy system are constructed by considering the integrated demand response and the uncertainty of renewable energy output. A master-slave game model that contains the energy system operator, energy producer, and energy users is then established, and the existence of the Stackelberg equilibrium is demonstrated. Furthermore, a distributed algorithm is proposed to resolve the game model by combining an improved coyote optimization algorithm with quadratic programming. Due to the shortcomings of the conventional coyote optimization algorithm, such as slow convergence rate and quickly falling into local optimum, a beetle antennae search is utilized to strengthen the optimal and the worst coyotes and to improve the convergence speed, global search ability, and optimization accuracy of the standard coyote algorithm. Finally, an industrial park in Northern China is adopted as an illustration to evaluate the effectiveness of the model and the improved algorithm.