Development of a softsensor for particle size monitoring

Villar, R.G. Del; Thibault, Jules; Villar, R. del

doi:10.1016/0892-6875(95)00131-x

Cited by 20 publications

(11 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because some variables cannot be measured online (due to the inexistence of reliable hardware sensors) or require offline analytical measurements, model-based sensors, usually indicated as soft sensors or estimators, represent an interesting alternative for their estimation (13,14). In order to evaluate the accuracy of the estimations, uncertainties have to be considered as well.…”

Section: Introductionmentioning

confidence: 99%

Monitoring ZnS Precipitation: Estimation, Error Analysis and Experiment Design

Sampaio

Keesman

Lens

2009

Separation Science and Technology

View full text Add to dashboard Cite

A mass balance model of total soluble sulfide and free zinc with a second-order reaction term is theoretically able to reconstruct the zinc effluent concentration and the kinetic parameter (k). However, under real conditions this model predicts the zinc effluent concentration four orders of magnitude higher than the measured ones. The applied error analysis, based on linearization of the model followed by first-order variance propagation, showed that the accuracy of several of the input variables (flows and influent concentrations) jeopardized the estimation of the Zn concentration in the effluent, which is a phenomenon expected for every fast reaction with low product concentration. In order to overcome the inaccuracy issue, an ''apparent solubility product'' as a function of pS (11-20) was calculated from the experimental data, allowing for the subsequent determination of an ''apparent kinetic parameter'' (k A ), that excluding parallel reactions was between 1.7 Â 10 23 -6.2 Â 10 24 L=(mol Á h). This allowed for further tuning of the model such that the estimates of the Zn effluent concentration became of the same order of magnitude as the measured ones (10 À7 M Zn).

show abstract

Section: Introductionmentioning

confidence: 99%

Monitoring ZnS Precipitation: Estimation, Error Analysis and Experiment Design

Sampaio

Keesman

Lens

2009

Separation Science and Technology

View full text Add to dashboard Cite

show abstract

“…Grinding is one of the most important operations in a mineral processing plant [16], and semi-autogenous grinding (SAG) has attained increasing interests due to its advantage over the conventional grinding in energy saving and operating expenses [17]. The key for improved efficiency and reduced cost for an industrial SAG mill is to maintain the system variables optimized and stable.…”

Section: Introductionmentioning

confidence: 99%

Time delay neural network modeling for particle size in SAG mills

Shang

2011

Powder Technology

View full text Add to dashboard Cite

“…To ensure that the operation of industrial grinding circuits is profitable, it largely depends on accurate and timely measurement of the product particle size (PPS) at the product output stream [2]- [8]. The PPS, which is usually defined as the fraction of particles in cyclone overflow passing a sieve of 200 mesh aperture (%, 200 mesh), is the most important quality index that directly affects the performance of the subsequent beneficiation process in terms of product concentrate grade and metal recovery rate [1].…”

Section: Introductionmentioning

confidence: 99%

“…However, it is very time-consuming (each round of manual sampling and analysis usually needs two hours), and it needs a large amount of samples of particles. Thus, sieving cannot be used for online measurement and control [7], [8]. There are also some hard instruments, such as the BT-9300S laser particle size analyzer, which can be used to measure the PPS online.…”

Section: Introductionmentioning

confidence: 99%

Data-Driven Soft-Sensor Modeling for Product Quality Estimation Using Case-Based Reasoning and Fuzzy-Similarity Rough Sets

Zhou

Chai

2014

IEEE Trans. Automat. Sci. Eng.

View full text Add to dashboard Cite

Efficient operation of the integrated optimization or automation system in an industrial plant depends mainly on good measurement of product quality. However, measuring or estimating the product quality online in many industrial plants is usually not feasible using the available techniques. In this paper, a data-driven soft-sensor using case-based reasoning (CBR) and fuzzy-similarity rough sets is proposed for product quality estimation. Owning to the sustained learning ability, the modeling of a CBR soft-sensor does not need any additional model correction which is otherwise required by the neural network based methods to overcome the slow time-varying nature of industrial processes. Because the conventional -nearest neighbor ( -NN) algorithm is strongly influenced by the value of , an improved -NN algorithm with dynamic adjustment of case similarity threshold is proposed to retrieve sufficient matching cases for making a correct estimation. Moreover, considering that the estimation accuracy of the CBR soft-sensor system is closely related to the weights of case feature, a feature weighting algorithm using fuzzy-similarity rough sets is proposed in this paper. This feature weighting method does not require any transcendental knowledge, and its computation complexity is only linear with respect to the number of cases and attributes. The developed soft-sensor system has been successfully applied in a large grinding plant in China. And the application results show that the system has achieved satisfactory estimation accuracy and adaptation ability.Note to Practitioners-In the process industry, the product quality reflects the operational and economic performance of a manufacturing process. However, online measurement and control of product quality is generally difficult. In this paper, a data-driven soft-sensor is proposed to online estimate the product quality for a typical grinding process with time-varying dynamics. The method combines an improved CBR soft-sensor algorithm and a feature weighting algorithm using fuzzy-similarity rough sets. First, the weighs of case features in CBR are determined using the fuzzy-similarity rough sets in offline mode, and then they are used in CBR to accurately estimate the PPS online. The CBR soft-sensor system has been applied in a large industrial grinding plant. The application results show that the soft-sensor system has achieved satisfied estimation precision, good adaptability and Manuscript robustness, and can be easily realized in practice with low cost and maintenance.Index Terms-Case feature weighting, case-based reasoning (CBR), data-driven, fuzzy-similarity rough sets (FSRS), product quality estimation, soft-sensor modeling.

show abstract

Development of a softsensor for particle size monitoring

Cited by 20 publications

References 2 publications

Monitoring ZnS Precipitation: Estimation, Error Analysis and Experiment Design

Monitoring ZnS Precipitation: Estimation, Error Analysis and Experiment Design

Time delay neural network modeling for particle size in SAG mills

Data-Driven Soft-Sensor Modeling for Product Quality Estimation Using Case-Based Reasoning and Fuzzy-Similarity Rough Sets

Contact Info

Product

Resources

About