Data-driven soft sensors targeting heat pump systems

Yang, Song; Rolando, Davide; Marchante-Avellaneda, Javier; Zucker, Gerhard; Madani, Hatef

doi:10.1016/j.enconman.2023.116769

Cited by 10 publications

(9 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…a ij (7) where b ij is the weight coefficient of unlabeled samples fused into labeled samples. By using the above method, integrate unlabeled samples into labeled samples to form a semisupervised dataset.…”

Section: The Semi-supervised Learning Methods Based On Time Correlationmentioning

confidence: 99%

“…In the equation, the variation factor ε is a random number that follows a uniform distribution on [0,1], while X t j and X t i represent pollen from different flowers of the same type of plant. (4) Evaluate the new solution using the fitness function in equation (7). If the new solution is superior to the current solution, the new solution enters the next generation population; Otherwise, the current solution will enter the next generation population, and the current optimal solution g * will be found and saved.…”

Section: Model Parameter Optimization Based On Fpamentioning

confidence: 99%

“…Data-driven modeling refers to relying on data mining methods to find the mapping relationship between variables and outputs in data, without relying on expert experience and defining assumptions, without analyzing internal mechanisms, which belongs to black box modeling and strong flexibility, suitable for handling nonlinear process. Therefore, artificial neural network (ANN) [7], convolutional neural network (CNN) [8], extreme learning machines (ELMs) [9], least squares support vector regression (LSSVR) [10], and longshort term memory network (LSTM) [11] are widely used in chemical production process [12]. Especially, LSTM is widely used in modeling methods due to its advantages in processing long-time series and long-term memory information.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Research on soft sensor modeling method for complex chemical processes based on local semi-supervised selective ensemble learning

Liu,

Li,

Zhang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

To address the adverse effects of nonlinearity and dynamic time-varying in complex chemical processes on the accuracy of the soft sensor model, a local-semi-supervised ensemble learning for soft sensor modeling (LS-SEL-LSTM) method is proposed in this article. Firstly, a hierarchical clustering method incorporating spatiotemporal criteria is proposed to reduce the influence of nonlinearity in global model prediction accuracy. The method considers the dynamic time-varying characteristics of temporal data and generates multiple local datasets. Then, to address the issue of multi-rate between auxiliary variables and dominant variables, a semi-supervised weight fusion mechanism based on temporal correlation is proposed, which effectively utilizes auxiliary variables to reconstruct local semi-supervised datasets and establishes local soft sensing models using LSTM. Concurrently, the parameters of the established model were optimized using the Flower Pollination Algorithm (FPA). Subsequently, a selective ensemble learning method based on sub-model prediction accuracy and an adaptive combination weight calculation method for sub-models were proposed to improve the prediction accuracy. Finally, the effectiveness of the proposed method was verified through the actual dataset of the sulfur recovery process. The results indicate that LS-SEL-LSTM performs well in handling complex chemical processes with nonlinear and dynamic time-varying characteristics.

show abstract

Section: The Semi-supervised Learning Methods Based On Time Correlationmentioning

confidence: 99%

Section: Model Parameter Optimization Based On Fpamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Research on soft sensor modeling method for complex chemical processes based on local semi-supervised selective ensemble learning

Liu,

Li,

Zhang

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…( 6), the inputs and output of the model are originally selected as shown in Table 1. In previous research [10], a sensitivity analysis is conducted and the predominated inputs are figured out. The result shows that with only water inlet and outlet temperatures as inputs, the model can estimate the condensation temperature accurately.…”

Section: Condenser Multivariate Polynomial Regression Modelmentioning

confidence: 99%

Development and validation of data-driven soft sensors for heat pumps

Song,

Rolando,

Avellaneda

et al. 2024

Preprint

View full text Add to dashboard Cite

Modern heat pump systems often come equipped with sensors, enabling the collection of substantial operational data. However, many residential heat pumps installed in preceding decades lack pressure sensors, energy meters, or mass flow meters, primarily due to financial limitations. As a result of these incomplete measurements, the direct analysis of the heat pump system's performance or the leveraging of the amassed data for inventive applications like prognosticating energy consumption, detecting and diagnosing faults, and implementing intelligent control becomes challenging.In existing literature, the focus of soft sensors in heat pump systems has been on estimating a single parameter. This approach, however, overlooks the reality that multiple parameters are often missing due to the lack of all-encompassing physical meters and sensors. Furthermore, current soft sensor models are typically developed using inputs such as compressor power consumption, pressures, evaporation, and condensation temperatures. These inputs, unfortunately, tend to be inaccessible within existing heat pump monitoring installations.In practice, it is a challenge to compensate for several critical measurements, encompassing mass flow rate, pressures, power consumption, and heating capacity, by using only commonly available sensors such as secondary loop temperatures and compressor frequency are available. Currently, there is a notable gap in research concerning this practical issue.To address the problems associated with inadequate measurements, this study presents the development and validation of soft sensors based on a data-driven approach, which can compensate for the parameters often unavailable with data collected from a limited number of commonly used sensors. Each component model employs a multivariate polynomial regression that calculates the evaporation temperature, condensation temperature, mass flow rate, and compressor power consumption, respectively. Subsequently, we present an integrated heat pump model that combines these component models into a comprehensive heat pump model.Finally, we validate the data-driven model against field test installations, demonstrating its accuracy with a relative root mean squared error (RRMSE) ranging from 10% to 20%.

show abstract

“…Due to space constraints, harsh inspection environments, and the high cost of complex industrial processes, soft measurements are often used to increase the point coverage of a measurement system [1,2]. Soft measurement modeling can be generally classified into mechanism-based and datadriven approaches [3,4]. Although data-driven soft measurement modeling has become a research hotspot in recent years, thanks to its simplicity and efficiency [5], the problem of missing data has become common during the collection of sensor data, and is affected by internal sensor failures, the different sampling rates of the sensors, and communication limitations.…”

Section: Introductionmentioning

confidence: 99%

Missing data filling in soft sensing using denoising diffusion probability model

Jiang,

Wang,

Shen

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

With the aim of addressing the problem of degradation in soft measurement accuracy due to missing data in industrial processes, a filling method based on the denoising diffusion probability model (DDPM) is proposed here to improve the accuracy of soft measurement modeling. First, missing regions are detected with the help of an improved Isolation Forest algorithm to obtain information such as the locations and numbers of missing data regions. Next, a data generation model is constructed based on DDPM and new samples are obtained. By adjusting the threshold for normal operation of the system and the weight sampler, filler samples that are similar to the distribution of the original data can be filtered from the new samples to form a complete dataset. The feasibility of the proposed missing data filling method is explored through numerical simulations, and its superiority in terms of improving the prediction accuracy of soft measurements is verified in regard to the nickel flash smelting process.

show abstract

Data-driven soft sensors targeting heat pump systems

Cited by 10 publications

References 29 publications

Research on soft sensor modeling method for complex chemical processes based on local semi-supervised selective ensemble learning

Research on soft sensor modeling method for complex chemical processes based on local semi-supervised selective ensemble learning

Development and validation of data-driven soft sensors for heat pumps

Missing data filling in soft sensing using denoising diffusion probability model

Contact Info

Product

Resources

About