18 results about "Multivariate calibration" patented technology

The invention relates to the technical field of water pollution detection, in particular to a water quality pollution multi-dimensional analysis system which comprises the following steps: synchronously measuring an ultraviolet-visible absorption spectrum of a water sample to be detected and scattered light intensity in an orthogonal direction, and deducting spectrum baseline interference caused by suspended particulate matters in the water sample through a correction algorithm; the interferents in the water sample are selectively converted and degraded by applying a programmed potential to the inflowing water sample to be detected; the reverse pulse is periodically applied, so that the online regeneration function of the electrode is realized; analyzing the spectral data by adopting a multivariate correction algorithm, and calculating the concentration of the target pollutant; the calculation residual error of the multivariate correction algorithm is analyzed, new interference and system performance drift of current measurement are diagnosed, and a corresponding system state code is generated; and responding to the system state code, automatically starting and carrying out state checking and multi-point calibration on the instrument, and using a calibration result to update the analysis model. Through multi-dimensional perception and closed-loop operation, accuracy and reliability of water quality monitoring are guaranteed.

The invention belongs to the technical field of wind tunnel tests, and discloses a self-adaptive multivariate calibration method for a wind tunnel balance. The calibration method is based on a multi-point force application body shafting balance calibration system, and loads are divided into a basic group load, an improved group load and a comprehensive group load. After basic group load loading is completed, a calibration formula with low precision is preliminarily fitted, then comprehensive group load is applied, a measured value is calculated according to the calibration formula, units with poor balance inspection precision are analyzed item by item, interference items with large errors are obtained, corresponding improved group load is re-compiled for the interference items, and the balance inspection precision is improved. The calibration value range of the improved group load is wider, the steps are denser, the group load loading is improved again, the calibration formula is continuously fitted, and the steps are repeated to carry out automatic iterative loading and continuously improve the accuracy of the calibration formula until the accuracy error of the calibration formula is smaller than a preset accuracy error value. The balance calibration method and the balance calibration system have practical engineering value.

This application relates to focused linear model correction and linear model correction of multivariate calibration model maintenance. A device can obtain master beta coefficients of a master calibration model associated with a master instrument. The master beta coefficients can be at a grid of a target instrument. The device can perform a constrained optimization of an objective function according to a set of constraints in order to determine a pair of transfer beta coefficients. The constrained optimization can be performed based on a pair of initial transfer beta coefficients, the master beta coefficients, and spectra associated with a scout set. The device can determine a transfer beta coefficient based on the pair of transfer beta coefficients. The device can determine a final transfer beta coefficient based on a set of transfer beta coefficients including the transfer beta coefficient. The final transfer beta coefficient can be associated with generating a transfer calibration model corresponding to the master calibration model for use by the target instrument.

The invention belongs to the technical field of testing materials by using optical means, and particularly relates to a method and a system for detecting impurities of wormwood essential oil by using a spectrum, aiming at solving the technical problems of single input information and low prediction precision of the existing model. The detection method comprises the following steps: S1, on a key scale, identifying candidate feature points meeting a cross-scale modulus maximum connection condition; s2, determining feature points, and forming a feature point set; s3, calculating an integral area of a preset impurity characteristic peak in the differential spectrum; and S4, taking the integral area, the average curvature and the average density as input variables, and substituting the input variables into an impurity content prediction model pre-established through a multivariate correction method so as to obtain an impurity content index. The stability and the anti-interference capability of the wormwood essential oil impurity content prediction model can be improved.

This invention discloses a static-weight displacement-type six-dimensional force sensor calibration device and method, belonging to the field of multi-dimensional force sensor calibration. It aims to solve the current challenges of high-precision, high-efficiency, and low-cost calibration of six-dimensional force sensors, and is of great significance for improving the accuracy and economy of multi-dimensional force sensor development. Compared with existing calibration equipment, this invention only includes a static-weight standard and a two-degree-of-freedom pitch and roll mechanism, resulting in a short error propagation chain and high loading repeatability and calibration uncertainty. By projecting the gravity vector constant of the static-weight standard onto the sensor in various displacement coordinate systems, the loading calibration of the sensor is achieved. A multivariate calibration method for the gravity vector constant is established, and the explicit formula of the sensor is obtained through multivariate fitting. The structure and calibration method described in this invention are highly conducive to automation, with high calibration efficiency and low development cost, making it particularly suitable for batch calibration of commercial six-dimensional force sensors and possessing considerable application prospects.

The method for detecting the aromatic hydrocarbon concentration of transformer oil based on fluorescence spectrum correction belongs to the technical field of transformer fault diagnosis, solves the problem of how to correct the measured aromatic hydrocarbon concentration of transformer oil under the current condition, and improves the detection accuracy of the aromatic hydrocarbon concentration of transformer oil.The method excites the fluorescence spectrum of transformer oil through monochromatic excitation light of the optimal excitation wavelength, establishes the mapping relationship between the fluorescence spectrum of transformer oil and the aromatic hydrocarbon compound concentration by adopting the multivariate regression correction method, designs a multivariate correction filter to calculate the aromatic hydrocarbon concentration of transformer oil under the current condition, corrects the measured aromatic hydrocarbon concentration of transformer oil under the current condition by adopting the calibration coefficient, and thus the real aromatic hydrocarbon concentration of transformer oil is obtained, the influence of external factors under the current condition on the detection accuracy of the aromatic hydrocarbon concentration of transformer oil is eliminated, and the detection accuracy of the aromatic hydrocarbon concentration of transformer oil is improved.

The application discloses a fault diagnosis device and method based on transformer oil fluorescence multivariate correction analysis, belongs to the technical field of transformer fault diagnosis, and solves the problems of large device volume and weight, high cost and inconvenience in use caused by directly collecting the fluorescence spectrum of transformer oil by using a fluorescence spectrometer in the prior art; the optimal excitation wavelength monochromatic excitation light generated by a fluorescence excitation source is used to excite the transformer oil in a fluorescence excitation detection device to generate fluorescence; the fluorescence excitation detection device generates fluorescence according to the input monochromatic excitation light and inputs the fluorescence into a fluorescence signal collection and analysis device; the fluorescence collection and analysis device adopts a multivariate correction filter group to collect and analyze the fluorescence signal emitted by the transformer oil, replaces the emission monochromator component of the fluorescence spectrometer, reduces the equipment cost and the equipment volume, the data processing is fast, the performance-price ratio of the transformer oil fault detection is improved, and the engineering application of the fluorescence monitoring technology in the transformer fault online diagnosis is realized.

The present application relates to the technical field of laser-induced breakdown spectroscopy detection, and specifically discloses a uranium content detection method based on uranium molecular spectrum formation mechanism, which comprises the following steps: step 1: using a laser-induced breakdown spectroscopy system to collect the spectrum of a uranium concentrate standard sample, and obtaining the laser-induced breakdown characteristic spectrum of the uranium concentrate standard sample with known uranium content; step 2: extracting the uranium molecular spectrum peak intensity, the oxygen atom characteristic spectrum line intensity and the uranium atom characteristic spectrum line intensity from the laser-induced breakdown characteristic spectrum of the uranium concentrate standard sample, and establishing a multivariate calibration model of the laser-induced breakdown spectroscopy signal intensity and the uranium content in the uranium concentrate standard sample; and step 3: detecting the uranium element content in the uranium concentrate sample. The method of the present application can significantly improve the detection accuracy of the laser-induced breakdown spectroscopy in detecting the uranium content in uranium concentrate products.

The invention discloses a hydroxylammonium nitrate-based propellant component content testing method and system, and the method comprises the steps: (1) obtaining a standard sample set, the standard sample set comprises a plurality of standard samples, each standard sample comprises a plurality of components of a hydroxylammonium nitrate-based propellant, and the content of each component is known; (2) collecting the near infrared spectrum of the standard sample set to form a standard spectrum data set; (3) on the basis of the standard spectrum data set and the known content of each component, constructing a quantitative analysis model for predicting the content of each component through a multivariate correction algorithm; and (4) collecting the near infrared spectrum of a to-be-detected sample, and simultaneously predicting the contents of various components in the to-be-detected sample by using the quantitative analysis model. The system comprises a near-infrared spectrometer, a temperature control device and a calculation processing unit. The testing method provided by the invention can realize one-step determination of the content of all components, and is simple and convenient to operate, rapid, non-destructive, accurate and reliable, so that an efficient quality assurance means is provided.

The application belongs to the technical field of pharmaceutical process quality measurement and control, and provides an AI-based pharmaceutical process density cross-modal measurement method and a measurement and control robot system; comprising the following steps: S1, real-time collection of video data on the surface of material in a plurality of pharmaceutical processes; S2, frame processing of the video to obtain a plurality of image sets; S3, preprocessing of each frame of image in the image set to extract cross-modal image features of density and obtain a feature vector; S4, sorting and splicing of the feature vector to obtain a global feature matrix of the plurality of videos and generate a density label vector; S5, standardization processing of the global feature matrix to obtain a standardized feature matrix; S6, taking the standardized feature matrix and the corresponding density label vector as a data set, fitting a nonlinear relationship between image features and density measurement values, and constructing a human-like visual depth multivariate correction model to cross-modally measure the density of the pharmaceutical process. The application can realize non-contact dynamic monitoring of the density of the pharmaceutical process.

The invention provides a method for rapidly determining the content of wax in crude oil based on infrared spectroscopy, and belongs to the field of petroleum exploration and development. The method comprises the following steps: forming a sample library by using various types of crude oil samples; collecting original infrared spectrum data of all crude oil samples in the sample library, and carrying out normalization pretreatment on the original infrared spectrum data to obtain an infrared spectrum; establishing a multivariate calibration model of the characteristic spectrum region of the infrared spectrum and the wax content in the crude oil sample; and collecting original infrared spectrum data of a crude oil sample to be detected, carrying out normalization pretreatment on the original infrared spectrum data, and then calculating the wax content in the crude oil sample to be detected according to the multivariate correction model. A crude oil sample is detected through an infrared spectroscopy method, a multiple regression model of crude oil physical property parameters is established by utilizing chemometrics software, and accurate prediction of the wax content of the crude oil is realized. Therefore, the detection period of the wax content in the crude oil is shortened.

The application discloses a model training method, a prediction method and equipment for particle size prediction. The method comprises the following steps: obtaining a sample set, wherein the sample set comprises a plurality of spectra corresponding to loose particle accumulations with different center particle diameters; setting the center particle diameter of the loose particle accumulation corresponding to the spectrum as a sample label, and randomly dividing the sample label into a plurality of sets for model training according to the sample label; performing data preprocessing on the spectrum in the plurality of sets, including overlapping averaging processing on the spectrum, and obtaining a plurality of averaged average spectra; and performing model training based on the average spectrum and a preset machine learning model to obtain a particle size prediction model. The application can obtain a multivariate calibration model with high confidence through preprocessing of samples and machine learning. The particle size prediction model can improve the accuracy of particle size prediction of loose particle accumulations.

The invention discloses a method and system for testing equivalent water content in a nitro oxidant, and the method comprises the following steps: (1) collecting near infrared spectrums of a plurality of nitro oxidant standard samples to form a near infrared spectrum set of the standard samples; (2) constructing a quantitative detection model of equivalent water content through a multivariate correction algorithm based on the near infrared spectrum set and the equivalent water content value of the nitro oxidant standard sample; (3) collecting the near infrared spectrum of a to-be-detected sample, and determining the equivalent water content by using the quantitative detection model; the nitro oxidant is selected from at least one of nitrogen tetroxide, nitrogen tetroxide containing nitric oxide and modified products thereof. The system comprises a near-infrared spectrometer, a temperature control device and a calculation processing unit. The test method provided by the invention has the advantages of stronger adaptability, wider applicability and higher test precision, and can realize rapid and accurate quantification of equivalent water content of various nitro oxidants under various environmental conditions and different sample pretreatment modes.

The application provides a gibberellic acid derivative synthesis path monitoring method and system, specifically comprising the following steps: in the reaction process, adopting an immersion near-infrared spectrum probe, collecting full-waveband near-infrared spectrum of a reaction mixture, and constructing a time series spectrum matrix; according to the vibration characteristics of the functional groups of the gibberellic acid derivative molecules, dividing the full-waveband into at least two characteristic regions, and carrying out multi-region combined pretreatment combined with the noise characteristics of each region; carrying out two-stage characteristic wavelength screening: in the first stage, constructing a composite index related to the time series stability and response intensity of the spectrum, and preliminarily screening to obtain a wavelength candidate set; in the second stage, using a chemometrics optimization algorithm for secondary optimization to determine an optimal characteristic wavelength combination; through timed sampling and offline chemical analysis to determine the actual concentration of the target product, establishing a multivariate calibration model; analyzing the time series data of the concentration, and when the absolute values of the first derivative and the second derivative are lower than the corresponding threshold values in a continuous preset period, it is determined that the reaction has reached the end point.

This invention, a transformer fault type discrimination method based on fluorescence spectroscopy correction, belongs to the field of transformer fault diagnosis technology. It addresses the problem of designing a simple, computationally efficient method for rapidly identifying transformer fault types, particularly thermal aging and discharge breakdown faults. The invention establishes a mapping relationship between the fluorescence spectrum of transformer oil and the concentration of aromatic hydrocarbons using a multivariate regression correction method. A multivariate correction filter is designed based on transmittance to calculate the aromatic hydrocarbon concentration in the transformer oil, replacing the fluorescence spectrometer. This not only reduces equipment costs and size but also improves the cost-effectiveness of transformer oil fault detection. Furthermore, based on the time-periodic characteristics of thermal aging and discharge breakdown faults, the rate of change in aromatic hydrocarbon concentration is used to distinguish the transformer fault type, greatly simplifying the calculation process and reducing computational load. This is of significant importance for extending the service life of transformers.

The invention relates to a detection method, in particular to a method for detecting residual protein in auxiliary material lactose based on a machine learning assisted visual chip. The visual chip detection method based on the smart phone is designed and developed on the basis of a coomassie brilliant blue protein color development principle, and the detection sensitivity and flux are improved by optimizing reaction conditions, chip design and image analysis parameters; meanwhile, a visual chip method assisted by machine learning is innovatively established, the dependence of a traditional method on experiment condition optimization is broken through, and quantitative analysis without instrument assistance is achieved. According to the invention, chromogenic reaction, intelligent image analysis and machine learning algorithms are further integrated, a dual-mode detection system with both chromaticity parameter standard curve calibration and machine learning multivariate correction is constructed, and a new path is provided for low-cost, portable and rapid detection of residual total protein in lactose.

The invention provides a biological signal real-time calibration system and a biological signal real-time calibration method for wearable equipment. The system comprises a flow control assembly used for collecting body fluid of a user of the wearable device in real time; a biological affinity sensor, a temperature sensor, a pH value sensor and an ionic strength sensor are at least integrated in the sensing cavity assembly in a mutually isolated manner; the driving acquisition assembly is used for periodically driving each sensor and synchronously acquiring the electric signals acquired by each sensor; and the micro-control assembly is used for acquiring the signals acquired by the sensors and transmitted by the driving acquisition assembly, firstly, aiming at the signals of the temperature sensor, the pH value sensor and the ion strength sensor, calculating real-time temperature, pH value and ion strength based on a preset calibration curve, and adopting a preset multivariable calibration model to calibrate the temperature, the pH value and the ion strength in real time. And correcting the original voltage signal representing the biological affinity by taking the calculated real-time temperature, the calculated pH value, the calculated ionic strength and the original voltage signal representing the biological affinity as input to obtain the calibrated concentration of the biomarker.