In this paper, a design method for the yarn tension sensor based on surface acoustic wave (SAW) is presented. When the yarn tension is applied to the sensor's doubly-clamped beam (i.e., piezoelectric substrate), it is linearly related to output frequency shift, which achieves the purpose of measuring yarn tension. This paper deduces the functional relationship between the oscillation frequency shift of the sensor and the yarn tension. We put forward the choice of piezoelectric substrate material, the choice of the piezoelectric die size and the placement of interdigital transducers (IDTs) as the three key issues to be determined. By using the ANSYS software, we determined that the optimal piezoelectric substrate is quartz, and that the optimal size of the die, and the maximum strain area of the die. When IDTs are placed in the maximum strain area of the piezoelectric substrate, the sensor has the best sensitivity. We have designed and produced the sensor, and also have tested the sensor. Experimental results: the range is 0~2 N, the linearity is 1.0112%, the hysteresis is 0.6153%, the repeatability is 1.1067%, the accuracy is 1.6205%, the maximum overload capacity is 200% (i.e., 4 N), and the fracture strength of the substrate is 5 N.

Nowadays, the commonly used yarn tension sensors cannot meet the requirement well in the textile manufacturing process. So surface acoustic wave (SAW) yarn tension sensor is a good selection to adapt to today's textile manufacturing and improve the production efficiency. First, two key problems of SAW yarn tension sensor in practical application are proposed in this article, one is the construction of the sensor measurement system, the other is the temperature compensation. Then, the integrated design of measurement circuit and measurement structure is implemented and introduced in this article, which improves the accuracy, reliability, and sensitivity of SAW yarn tension sensor. Finally, binary regression analysis method and least squares support vector machine (LSSVM) algorithm optimized by particle swarm optimization (PSO-LSSVM) method are applied to the temperature compensation of SAW yarn tension sensor, which both achieve satisfactory compensation effect. By comparison, PSO-LSSVM method is more convenient and flexible, and still has a large space for improvement. Through those research contents, it can be concluded that the design and implementation of the measurement system lay a foundation for the practical application of SAW yarn tension sensor, the research of temperature compensation is of great significance to the wide application of the sensor.

This paper presents a 60 MHz surface acoustic wave (SAW) yarn tension sensor incorporating a novel SAW oscillator with high-frequency stability. A SAW delay line was fabricated on ST-X quartz substrate using the unbalanced-split electrode and bi-directional engraving slots. The dual differential channel delay linear acoustic surface wave oscillator is designed and implemented to test yarn tension, which can effectively remove the interference of temperature, humidity, and other peripheral factors through differential design. The yarn tension sensor using the surface acoustic wave has high-precision characteristics, and the SAW delay line oscillator is designed to ensure the test system’s stable operation. The effect of time and tension on oscillator frequency stability is studied in detail, and the single oscillator and the dual differential channel system were tested, respectively. After using the dual differential channel system, the short-term frequency stability from is reduced from $1.0163 \mathrm{ppm}$ to $0.17726 \mathrm{ppm}$, the frequency accuracy of the tension sensor is improved from 134 Hz to 27 Hz, and the max frequency jump steady is reduced from $2.2395 \mathrm{ppm}$ to $0.45123 \mathrm{ppm}$.

With the continuous increase in the proportion of distributed photovoltaic power stations, the demand for photovoltaic power grid connection is becoming more and more urgent, and the requirements for the accuracy of regional distributed photovoltaic power forecasting are also increasing. A distributed regional photovoltaic power prediction model based on a stacked ensemble algorithm is proposed here. This model first uses a graph attention network (GAT) to learn the structural features and relationships between sub-area photovoltaic power stations, dynamically calculating the attention weights of the photovoltaic power stations to capture the global relationships and importance between stations, and selects representative stations for each sub-area. Subsequently, the CNN-LSTM-multi-head attention parallel multi-channel (CNN-LSTM-MHA (PC)) model is used as the basic model to predict representative stations for sub-areas by integrating the advantages of both the CNN and LSTM models. The predicted results are then used as new features for the input data of the meta-model, which finally predicts the photovoltaic power of the large area. Through comparative experiments at different seasons and time scales, this distributed regional approach reduced the MAE metric by a total of 22.85 kW in spring, 17 kW in summer, 30.26 kW in autumn, and 50.62 kW in winter compared with other models.

In this paper, a SAW winding tension sensor is designed and data fusion technology is used to improve its measurement accuracy. To design a high-measurement precision SAW winding tension sensor, the unbalanced split-electrode interdigital transducers (IDTs) were used to design the input IDTs and output IDTs, and the electrode-overlap envelope was adopted to design the input IDT. To improve the measurement accuracy of the sensor, the particle swarm optimization-least squares support vector machine (PSO-LSSVM) algorithm was used to compensate for the temperature error. After temperature compensation, the sensitivity temperature coefficient α_s of the SAW winding tension sensor was decreased by an order of magnitude, thus significantly improving its measurement accuracy. Finally, the error with actually applied tension was calculated, the same in the LSSVM and PSO-LSSVM. By multiple comparisons of the same sample data set overall, as well as the local accuracy of the forecasted results, which is $5.95\mathrm{\%}$, it is easy to confirm that the output error predicted by the PSO-LSSVM model is $0.50\mathrm{\%}$, much smaller relative to the LSSVM’s $1.42\mathrm{\%}$. As a result, a new way for performing data analysis of the SAW winding tension sensor is provided.