Method for evaluating quality of intumescent fire retardant coating

Abstract

The invention discloses a quality evaluation method for an intumescent fireproof coating, and relates to the technical field of fireproof material detection. Comprising the following steps: based on a near infrared spectrogram of an intumescent fire retardant coating sample and a result spectrogram of a to-be-detected sample, obtaining a matching degree of the to-be-detected sample; detecting the element content of the to-be-detected sample to obtain the phosphorus content, the titanium content and the residual total inorganic content of the to-be-detected sample; normalizing the characteristic parameters of the sample to be detected and the plurality of standard substances qualified in the national standard, wherein the characteristic parameters comprise the matching degree, the phosphorus content, the titanium content and the residual total inorganic content; analyzing the normalized data through a PCA algorithm, and extracting principal components; according to the score of the principal component, the variance contribution rate and the load, identifying a mode and an abnormal value in the to-be-detected sample; and judging the quality of the sample to be detected according to the abnormal value. According to the invention, the quality of the intumescent fireproof coating can be rapidly and accurately detected on site, so that the acceptance work is simplified, and the fireproof safety of a structure is guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of fireproof material detection, and in particular to a method for evaluating the quality of an intumescent fireproof coating. Background Art

[0002] In order to prevent fire accidents and block the spread of fire and its chain reaction, it is necessary to lay a fireproof layer on equipment, components and structures with fireproof requirements. Among them, the intumescent fire-retardant coating uses the flame retardancy and fire resistance of the material or structure itself to control the fire within a small range, minimizing the loss and impact. Therefore, intumescent fire-retardant coatings are widely used in the fire protection of steel structure buildings, fire protection pipe networks, portal frames, cables, pipelines and bridges in substations (converter stations), providing important protection for the fire safety of power equipment and facilities in substations (converter stations).

[0003] However, the quality of intumescent fire retardant coatings on the market varies. In order to ensure that the intumescent fire retardant coatings used are of high quality, they need to be tested for quality. In the prior art, most of them can only be tested in laboratories. Although relatively accurate quality tests can be performed, the testing time is long and the cost is high. It is still difficult for users and supervisors to identify the quality of intumescent fire retardant coatings on site. Summary of the invention

[0004] In view of the technical problem that it is difficult to identify the quality of intumescent fire retardant coatings on site, the present invention provides a method for evaluating the quality of intumescent fire retardant coatings, which can quickly and accurately detect the quality of intumescent fire retardant coatings on site, so as to simplify the acceptance work of fire protection projects and ensure the fire safety of structures such as substations (converter stations).

[0005] The present invention is achieved through the following technical solutions:

[0006] The present invention provides a method for evaluating the quality of an intumescent fire retardant coating, comprising the following steps:

[0007] Based on the near infrared spectrum of the intumescent fire retardant coating sample and the result spectrum of the sample to be tested, the matching degree of the sample to be tested is obtained;

[0008] Detecting the element content of the sample to be tested to obtain the phosphorus content, titanium content and remaining total inorganic content of the sample to be tested;

[0009] Normalizing characteristic parameters of the sample to be tested and a plurality of national standard qualified standards, wherein the characteristic parameters include matching degree, phosphorus content, titanium content and remaining total inorganic content;

[0010] The normalized data were analyzed by PCA algorithm to extract the principal components;

[0011] Identify patterns and outliers in the sample based on the scores, variance contributions, and loadings of the principal components;

[0012] The quality of the sample to be tested is judged according to the abnormal value.

[0013] The method for evaluating the quality of an intumescent fire retardant coating provided by the present invention first obtains the matching degree of the sample to be tested based on the near-infrared spectrum of the intumescent fire retardant coating sample and the result spectrum of the sample to be tested, and the spectrum can be obtained on site, so that whether the consistency of the sample to be tested is qualified can be judged on site through the matching degree; then the element content of the sample to be tested is detected to obtain the phosphorus content, titanium content and remaining total inorganic content of the sample to be tested, and the element content of the sample to be tested can also be detected and obtained on site, so as to judge on site whether the content of each element of the sample to be tested is qualified through the content of each element, and the matching degree, phosphorus content, titanium content and remaining total inorganic content of the sample to be tested and a plurality of national standard qualified standard products are analyzed by the PCA algorithm on the normalized data, the principal component is extracted, and the mode and outlier in the sample to be tested are identified according to the score, variance contribution rate and load of the principal component, and finally the quality of the sample to be tested is judged according to the outlier, so as to comprehensively and accurately judge the quality of the sample to be tested by comprehensively judging the matching degree, phosphorus content, titanium content and remaining total inorganic content of the sample to be tested.

[0014] Therefore, the quality assessment method of the intumescent fire retardant coating provided by the present invention can quickly and accurately detect the quality of the intumescent fire retardant coating on site, so as to simplify the acceptance work of fire protection projects and ensure the fire safety of structures such as substations (converter stations).

[0015] In an optional embodiment of the present application, the step of obtaining the matching degree of the sample to be tested includes:

[0016] Collect near infrared spectra of qualified and unqualified intumescent fire retardant coating samples;

[0017] Using a plurality of data processing models to pre-process the corresponding near-infrared spectra respectively to obtain a plurality of standard sample spectra;

[0018] Using a characteristic wavelength selection model, feature extraction is performed on the spectrum of each standard sample to obtain a contribution band;

[0019] Based on the contribution band, determining a first threshold of the consistency spectrum;

[0020] A result spectrum of the sample to be tested is obtained, and a matching degree of the sample to be tested is calculated based on the first threshold and the absorbance of the second spectrum of the sample to be tested.

[0021] In an optional embodiment of the present application, a processing model for preprocessing the near-infrared spectrum includes standard normal variable transformation processing, SG convolution smoothing processing and vector normalization processing.

[0022] In an optional implementation of the present application, the standard normal variable transformation process includes the following steps:

[0023] Calculate the mean (μ) and standard deviation (σ) of the raw score (X) of the near-infrared spectra;

[0024] The original score X of the near infrared spectrum is converted into a standard normal variable Z based on a transformation model, wherein the transformation model is Z=(X-μ) / σ;

[0025] Among them, X is the original data point of the near-infrared spectrum, μ is the mean of the original data of the near-infrared spectrum, σ is the standard deviation of the original data of the near-infrared spectrum, Z is the transformed standard normal variable, and the Z score represents the distance between the original data point of the near-infrared spectrum and the mean, and is in units of standard deviation.

[0026] In an optional implementation manner of the present application, the SG convolution smoothing process includes the following steps:

[0027] An odd number of NIR spectrum data points are included as a window;

[0028] For each subset in the window, a polynomial is fitted using the least squares method;

[0029] Slide the window along the historical data set and repeat the above process until all near-infrared spectrum data points are covered.

[0030] In an optional implementation of the present application, the characteristic wavelength selection model is one of principal component analysis, partial least squares regression and genetic algorithm.

[0031] In an optional implementation manner of the present application, the step of obtaining the first threshold is:

[0032] Based on the standard sample in the contribution band, obtaining the absorbance of a first spectrum of the standard sample;

[0033] A first threshold value of the consistent spectrum is determined based on the maximum absorbance of the first spectrum among all standard samples.

[0034] In an optional embodiment of the present application, the calculation model of the absorbance of the first spectrum is:

[0035]

[0036] Where: S ijis the root mean square error of the absorbance between the i-th standard sample and the j-th standard sample, and is used as the absorbance of the first spectrum;

[0037] k is the kth wavelength;

[0038] n is the total number of wavelengths;

[0039] y ik is the absorbance value of the i-th standard sample at the k-th wavelength;

[0040] y jk is the absorbance value of the jth standard sample at the kth wavelength.

[0041] In an optional implementation of the present application, the calculation model of the matching degree is:

[0042]

[0043] Where: R is the matching degree;

[0044] S is the first threshold of the consistent spectrum;

[0045] S′ is the root mean square error of the absorbance between the result spectrum and the consistency spectrum of the intumescent fire retardant coating sample to be tested, and is used as the absorbance of the second spectrum;

[0046] k is the kth wavelength;

[0047] n is the total number of wavelengths;

[0048] y′ k is the absorbance value of the intumescent fire retardant coating sample to be tested under the kth wavelength condition;

[0049] y k is the absorbance value of the consistency spectrum at the kth wavelength.

[0050] In an optional implementation of the present application, when the normalized data is analyzed by the PCA algorithm, the maximum value of the characteristic parameter is used as a reference.

[0051] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0052] The method for evaluating the quality of an intumescent fire retardant coating provided by the present invention first obtains the matching degree of the sample to be tested based on the near-infrared spectrum of the intumescent fire retardant coating sample and the result spectrum of the sample to be tested, and the spectrum can be obtained on-site, so that the consistency of the sample to be tested can be judged on-site through the matching degree; then the element content of the sample to be tested is detected to obtain the phosphorus content, titanium content and remaining total inorganic content of the sample to be tested, and the element content of the sample to be tested can also be detected and obtained on-site, so as to judge on-site whether the content of each element of the sample to be tested is qualified through the content of each element, and the matching degree, The phosphorus content, titanium content and remaining total inorganic content are analyzed by the PCA algorithm on the normalized data, and the principal components are extracted. According to the scores, variance contributions and loads of the principal components, the patterns and outliers in the samples to be tested are identified, and finally the quality of the samples to be tested is judged according to the outliers. The quality of the samples to be tested is judged comprehensively based on the matching degree, phosphorus content, titanium content and remaining total inorganic content of the samples to be tested, so as to comprehensively and accurately judge the quality of the samples to be tested. Therefore, the quality of the intumescent fire retardant coating can be quickly and accurately tested on site to simplify the acceptance of fire protection projects and ensure the fire safety of structures such as substations (converter stations). BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0054] in, Figure 1 A schematic flow chart of a method for evaluating the quality of an intumescent fire retardant coating provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0056] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0057] It should be noted that the embodiments and features in the embodiments in this application can be combined with each other without conflict. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not need to be adopted to implement the present invention. In other embodiments, in order to avoid confusing the present invention, known methods are not specifically described.

[0058] Example

[0059] Combination Figure 1 This embodiment provides a method for evaluating the quality of an intumescent fire retardant coating, comprising the following steps:

[0060] S100, obtaining a matching degree of the sample to be tested based on the near infrared spectrum of the intumescent fire retardant coating sample and the result spectrum of the sample to be tested;

[0061] It is understandable that near-infrared spectroscopy analysis technology has been widely used because of its rapid, non-destructive, high efficiency and low technical requirements for operation, and can measure the characteristics of solid, powder, slurry, liquid and other samples. In addition, near-infrared spectroscopy can reflect the structure and composition information of hydrogen-containing groups, and full-spectrum or multi-band spectral data can be combined with chemometric methods to establish models for quantitative or qualitative analysis.

[0062] Specifically, the steps of obtaining the matching degree of the sample to be tested include:

[0063] S110. Collect near-infrared spectra of qualified and unqualified intumescent fire retardant coating samples.

[0064] Specifically, after laboratory testing, the quality of the intumescent fire retardant coating can be clarified, that is, the intumescent fire retardant coating collected in this embodiment has clear results after laboratory testing, and is marked as qualified or unqualified, and the near-infrared spectra of historical intumescent fire retardant coating samples are collected.

[0065] In order to ensure that the collected intumescent fire retardant coating samples are representative, and the data collection process is standardized to reduce the impact of non-correlated differences between different samples. Therefore, in this embodiment, intumescent fire retardant coatings of different brands and models are collected, and different intumescent fire retardant coating types and possible variation factors are covered (solvent-based and water-based fire retardant coatings should be included, and the fire resistance limit should be as wide as possible) to make the samples diverse and representative, and each sample is determined to be a qualified product or a non-qualified product by standard experiments, and the near-infrared spectra of the above samples are collected respectively.

[0066] S120, using a plurality of data processing models to pre-process the corresponding near-infrared spectra respectively to obtain a plurality of standard sample spectra.

[0067] That is to say, different data preprocessing methods are used to preprocess the near-infrared spectrogram in turn to obtain the preprocessed standard sample spectrogram. In this embodiment, the data preprocessing method includes standard normal variable transformation processing, SG convolution smoothing processing and vector normalization processing. That is, the processing model for preprocessing the near-infrared spectrogram in this embodiment includes standard normal variable transformation processing, SG convolution smoothing processing and vector normalization processing. Thus, the quality of spectral data is improved through the above data preprocessing method. Among them, SG convolution smoothing processing is used for smoothing processing to reduce the influence of noise.

[0068] It should be understood that the standard normal variable transformation is a method of converting any normally distributed random variable into a standard normal distribution (also known as a Z distribution). The standard normal distribution is a normal distribution with a mean of 0 and a standard deviation of 1. In this embodiment, the standard normal variable transformation process includes the following steps:

[0069] S121a. Calculate the mean (μ) and standard deviation (σ) of the original score (X) of the near-infrared spectrum, that is, obtain the parameters of the original normal distribution.

[0070] S121b, converting the original score X of the near infrared spectrum into a standard normal variable Z based on a transformation model, wherein the transformation model is Z=(X-μ) / σ;

[0071] Among them, X is the original data point of the near-infrared spectrum, μ is the mean of the original data of the near-infrared spectrum; σ is the standard deviation of the original data of the near-infrared spectrum, Z is the transformed standard normal variable, and the Z score represents the distance between the original data point of the near-infrared spectrum and the mean, and is in units of standard deviation. For example, a Z score of 2 means that the original data point of the near-infrared spectrum is 2 standard deviations higher than the mean.

[0072] As for the SG convolution smoothing process, it is known that SG (Savitzky-Golay) convolution smoothing is a mathematical processing technique for smoothing data, which retains the high-frequency characteristics of the data while reducing noise by fitting a polynomial of a subset of data points.

[0073] In this embodiment, the SG convolution smoothing process includes the following steps:

[0074] S122a, taking an odd number of near-infrared spectrum data points as a window, that is, selecting a window containing an odd number of near-infrared spectrum data points. The window will slide in the historical data set, and the near-infrared spectrum data points in the window will be smoothed.

[0075] S122b, for each subset in the window, a polynomial is fitted using the least square method. The order of the polynomial can be selected as needed, but is usually second or third order. For each near infrared spectrum data point in the window, its smoothed value is the polynomial fitting value of the near infrared spectrum data point.

[0076] S122c, slide the window along the historical data set, and repeat the above process until all near-infrared spectrum data points are covered.

[0077] The vector normalization process includes the following steps:

[0078] S123a, calculating the length (or size) of the near infrared spectrum vector.

[0079] For a two-dimensional vector Its length Calculated by the following formula:

[0080]

[0081] For 3D vector The length is calculated as follows:

[0082]

[0083] S123b, divide each component of the near infrared spectrum vector by the length of the vector to obtain a unit vector For a two-dimensional vector:

[0084]

[0085] For a three-dimensional vector, it is:

[0086]

[0087] S130, using a characteristic wavelength selection model to perform feature extraction on the spectrum of each standard sample to obtain a contribution band.

[0088] Specifically, the characteristic wavelength selection model is one of principal component analysis (PCA), partial least squares regression (PLSR) and genetic algorithm (GA). And through the aforementioned characteristic wavelength selection model, the standard sample spectrum is feature extracted to screen out the wavelengths (i.e., contribution bands) that contribute significantly to the target attributes. In this embodiment, the contribution band is 1200nm to 2400nm, and this band range can better cover the characteristic spectral lines of the sample components.

[0089] S140. Determine a first threshold of a consistency spectrum based on the contribution band.

[0090] Specifically, the step of obtaining the first threshold is:

[0091] S141. Based on the standard sample in the contribution band, obtain the absorbance of a first spectrum of the standard sample.

[0092] Specifically, the calculation model of the absorbance of the first spectrum is:

[0093]

[0094] Where: S ij is the root mean square error of the absorbance between the i-th standard sample and the j-th standard sample, and is used as the absorbance of the first spectrum;

[0095] k is the kth wavelength;

[0096] n is the total number of wavelengths;

[0097] y ik is the absorbance value of the i-th standard sample at the k-th wavelength;

[0098] y jk is the absorbance value of the jth standard sample at the kth wavelength.

[0099] S142. Determine a first threshold value of the consistency spectrum based on the maximum absorbance of the first spectrum among all standard samples.

[0100] Specifically, according to the absorbance of the first spectrum, the largest absorbance of the first spectrum among all standard samples is found; based on the largest absorbance of the first spectrum, the first threshold of the consistency spectrum is determined. In this embodiment, the S of all standard samples are statistically analyzed. ij value, find the maximum absorbance root mean square error value S ij , determine the threshold S of the consistency spectrum based on the first threshold calculation model.

[0101] Among them, the calculation model of the first threshold is:

[0102] S=2Max(S ij )

[0103] Where S is the first threshold; Max() is the maximum function; S ij is the absorbance of the first spectrum.

[0104] S150, obtaining a result spectrum of the sample to be tested, and calculating a matching degree of the sample to be tested based on a first threshold and the absorbance of the second spectrum of the sample to be tested.

[0105] The specific steps are:

[0106] S151, obtaining a result spectrum of the sample to be tested. In this step, the sample to be tested is coated on a cable.

[0107] S152. Preprocess the result spectrum of the intumescent fire retardant coating sample to be tested by using standard normal variable transformation processing, SG convolution smoothing processing and vector normalization processing to obtain the preprocessed spectrum to be tested; the standard normal variable transformation processing, SG convolution smoothing processing and vector normalization processing are the same as the specific algorithms in step S120.

[0108] S153, determining the absorbance S′ of the second spectrum of the sample to be tested according to the spectrum to be tested;

[0109] S154, and based on the first threshold value and the absorbance S′ of the second spectrum, calculate the matching degree of the intumescent fire retardant coating sample to be tested;

[0110] In this embodiment, the calculation model of the matching degree is:

[0111]

[0112] Where: R is the matching degree;

[0113] S is the first threshold of the consistent spectrum;

[0114] S′ is the root mean square error of the absorbance between the result spectrum and the consistency spectrum of the intumescent fire retardant coating sample to be tested, and is used as the absorbance of the second spectrum;

[0115] k is the kth wavelength;

[0116] n is the total number of wavelengths;

[0117] y′ k is the absorbance value of the intumescent fire retardant coating sample to be tested under the kth wavelength condition;

[0118] y k is the absorbance value of the consistency spectrum at the kth wavelength.

[0119] It is understandable that samples whose matching degree reaches a set threshold can be judged as qualified in sample consistency. In this embodiment, the samples to be tested and sampled are characterized, and a band of 1200nm to 2400nm is selected for comparison. When the matching degree R is 95%, the sample consistency can be judged on-site as qualified.

[0120] Therefore, this embodiment selects different brands and models of intumescent fire retardant coatings, and combines absorbance and matching degree to achieve on-site determination of whether the sample to be tested is qualified.

[0121] It is understandable that the existing intumescent fire retardant coatings can foam and expand under high temperature or fire to form a foam carbonized layer that is dozens or even hundreds of times thicker than the original coating, thereby achieving excellent fireproofing and heat insulation effects, protecting the substrate and preventing the substrate under the carbonized layer from direct contact with the flame, ignition, and further thermal oxidation degradation.

[0122] The formula of intumescent fire retardant coatings must include a foaming agent (gas source), a carbonizing agent (carbon source) and a carbonizing catalyst (acid source). The synergistic effect of the three is indispensable at high temperatures. It is usually a ternary flame retardant system of ammonium polyphosphate-pentaerythritol-melamine. Titanium dioxide is an important skeleton support structure of intumescent fire retardant coatings, and currently no other reagent can replace it. At the same time, inorganic minerals such as magnesium hydroxide, aluminum hydroxide, expanded vermiculite, and wollastonite contain a large amount of silicon, aluminum, magnesium, calcium and other elements respectively. A small amount of inorganic minerals can be combined with titanium dioxide as a skeleton support for intumescent fire retardant coatings.

[0123] Therefore, if the matching degree does not meet the requirements, it is a defective product, while if the matching degree meets the requirements, further judgment is required.

[0124] S200, detecting the element content of the sample to be tested to obtain the phosphorus content, titanium content and remaining total inorganic content of the sample to be tested.

[0125] Specifically, X-ray fluorescence spectrometry (XRF) can be used to characterize the elemental composition and content of each sample surface. The analysis mode is a geochemical mode, which can detect Mg and all elements after Mg in the periodic table.

[0126] In this embodiment, the specific detection steps are:

[0127] S210, apply an appropriate amount of intumescent fire retardant coating on the flat surface of glass / filter paper / wooden board / steel. Usually filter paper is used. The coating thickness is greater than 1mm and the area is greater than 4cm. 2 , let it stand at room temperature until its surface is dry.

[0128] S220. X-ray fluorescence spectrometry (XRF) is used to characterize the surface element composition and content of the coating layer after surface drying. The analysis mode is geochemical mode (or all elements after Mg can be tested), and the content of elements such as P, Ti, Mg, Si, Al, and Ca is recorded. The remaining total inorganic content is the total content of Mg, Si, Al, and Ca.

[0129] S230: Extract characteristic parameters and mark them. The specific results are shown in Table 1.

[0130]

[0131] Table 1

[0132] In other words, if one of the following conditions is met, the quality can be judged as unqualified:

[0133] 1. P content: 0-4.5%;

[0134] 2. Ti content: 0-4.5%;

[0135] 3. Si+Al+Mg+Ca content: more than 15%;

[0136] 4. Near infrared spectrum matching degree is less than 50%.

[0137] For qualified products, the above conditions must be met at the same time.

[0138] S300, normalizing the characteristic parameters of the sample to be tested and the standard products that meet the national standards (GB 28374-2012 "Cable Fire Retardant Coating" and GB 14907-2018 "Steel Structure Fire Retardant Coating"), wherein the characteristic parameters include matching degree, phosphorus content, titanium content and remaining total inorganic content.

[0139] That is to say, there are four characteristic parameters in this embodiment, specifically, parameter 1: near infrared matching degree (specific value of more than 50%); parameter 2: P content of more than 4.5% specific value; parameter 3: Ti content of more than 4.5% specific value; parameter 4: Si+Al+Mg+Ca total content of 0-15% specific value. As for the number of standard products that meet the national standard, there are usually dozens of them to ensure the accuracy of the evaluation. In this embodiment, 60 standard products that meet the national standard are collected.

[0140] S400, analyzing the normalized data by using a PCA algorithm to extract principal components.

[0141] In an embodiment, when the normalized data is analyzed by the PCA algorithm, the maximum value of the characteristic parameter is used as a benchmark. The characteristic parameter principal component score data is thus decomposed, and each sample corresponds to a score value for each principal component. In this embodiment, all modeling samples are plotted in three-dimensional space with the score values ​​of the first principal component, the second principal component, and the third principal component, the X-axis represents the first principal component score (PC1) of the sample, the Y-axis represents the second principal component score (PC2) of the sample, and the Z-axis represents the third principal component score (PC3) of the sample, and a cluster analysis diagram is obtained.

[0142] S500, identifying patterns and outliers in the sample to be tested according to the scores, variance contributions and loads of the principal components;

[0143] S600, judging the quality of the sample to be tested according to the abnormal value, that is, if the matching degree R is greater than or equal to 95%, the sample consistency can be judged to be qualified on site.

[0144] In summary, the intumescent fire retardant coating quality assessment method provided in this embodiment first obtains the matching degree of the sample to be tested based on the near-infrared spectrum of the intumescent fire retardant coating sample and the result spectrum of the sample to be tested, and the spectrum can be obtained on-site, so that the consistency of the sample to be tested can be judged on-site by the matching degree. Then, the element content of the sample to be tested is detected to obtain the phosphorus content, titanium content and remaining total inorganic content of the sample to be tested, and the element content of the sample to be tested can also be detected and obtained on-site, so as to judge on-site whether the content of each element of the sample to be tested is qualified through the content of each element.

[0145] On this basis, the normalized data of the matching degree, phosphorus content, titanium content and remaining total inorganic content of the samples to be tested and multiple national standard qualified standards are analyzed by the PCA algorithm, and the principal components are extracted. According to the scores, variance contributions and loadings of the principal components, the patterns and outliers in the samples to be tested are identified, and finally the quality of the samples to be tested is judged according to the outliers. The quality of the samples to be tested is judged comprehensively based on the matching degree, phosphorus content, titanium content and remaining total inorganic content of the samples to be tested, so as to comprehensively and accurately judge the quality of the samples to be tested.

[0146] Therefore, the quality assessment method of the intumescent fire retardant coating provided in this embodiment can quickly and accurately detect the quality of the intumescent fire retardant coating on site, so as to simplify the acceptance work of fire protection projects and ensure the fire safety of structures such as substations (converter stations).

[0147] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating the quality of an intumescent fire retardant coating, characterized in that: The following steps are involved: Based on the near infrared spectrum of the intumescent fire retardant coating sample and the result spectrum of the sample to be tested, the matching degree of the sample to be tested is obtained; Detecting the element content of the sample to be tested to obtain the phosphorus content, titanium content and remaining total inorganic content of the sample to be tested; Normalizing characteristic parameters of the sample to be tested and a plurality of national standard qualified standards, wherein the characteristic parameters include matching degree, phosphorus content, titanium content and remaining total inorganic content; The normalized data were analyzed by PCA algorithm to extract the principal components; Identify patterns and outliers in the sample based on the scores, variance contributions, and loadings of the principal components; The quality of the sample to be tested is judged according to the abnormal value.

2. The method for evaluating the quality of an intumescent fire retardant coating according to claim 1, characterized in that: The steps of obtaining the matching degree of the sample to be tested include: Collect near infrared spectra of qualified and unqualified intumescent fire retardant coating samples; Using a plurality of data processing models to pre-process the corresponding near-infrared spectra respectively to obtain a plurality of standard sample spectra; Using a characteristic wavelength selection model, feature extraction is performed on the spectrum of each standard sample to obtain a contribution band; Based on the contributing wavelength band, determining a first threshold of a consistent spectrum; A result spectrum of the sample to be tested is obtained, and a matching degree of the sample to be tested is calculated based on the first threshold and the absorbance of the second spectrum of the sample to be tested.

3. The method for evaluating the quality of an intumescent fire retardant coating according to claim 2, characterized in that: The processing model for preprocessing the near-infrared spectrum includes standard normal variable transformation processing, SG convolution smoothing processing and vector normalization processing.

4. The method for evaluating the quality of an intumescent fire retardant coating according to claim 2, characterized in that: The standard normal variable transformation process comprises the following steps: Calculate the mean (μ) and standard deviation (σ) of the raw score (X) of the near-infrared spectra; The original score X of the near infrared spectrum is converted into a standard normal variable Z based on a transformation model, wherein the transformation model is Z=(X-μ) / σ; Among them, X is the original data point of the near-infrared spectrum, μ is the mean of the original data of the near-infrared spectrum, σ is the standard deviation of the original data of the near-infrared spectrum, Z is the transformed standard normal variable, and the Z score represents the distance between the original data point of the near-infrared spectrum and the mean, and is in units of standard deviation.

5. The method for evaluating the quality of an intumescent fire retardant coating according to claim 2, characterized in that: The SG convolution smoothing process comprises the following steps: An odd number of NIR spectrum data points are included as a window; For each subset in the window, a polynomial is fitted using the least squares method; Slide the window along the historical data set and repeat the above process until all near-infrared spectrum data points are covered.

6. The method for evaluating the quality of an intumescent fire retardant coating according to claim 2, characterized in that: The characteristic wavelength selection model is one of principal component analysis, partial least squares regression and genetic algorithm.

7. The method for evaluating the quality of an intumescent fire retardant coating according to claim 2, characterized in that: The steps of obtaining the first threshold are: Based on the standard sample in the contribution band, obtaining the absorbance of a first spectrum of the standard sample; A first threshold value of the consistent spectrum is determined based on the maximum absorbance of the first spectrum among all standard samples.

8. The method for evaluating the quality of an intumescent fire retardant coating according to claim 7, characterized in that: The calculation model of the absorbance of the first spectrum is: Where: S ij is the root mean square error of the absorbance between the i-th standard sample and the j-th standard sample, and is used as the absorbance of the first spectrum; k is the kth wavelength; n is the total number of wavelengths; y ik is the absorbance value of the i-th standard sample at the k-th wavelength; y jk is the absorbance value of the jth standard sample at the kth wavelength.

9. The method for evaluating the quality of an intumescent fire retardant coating according to any one of claims 2 to 8, characterized in that: The calculation model of the matching degree is: Where: R is the matching degree; S is the first threshold of the consistent spectrum; S′ is the root mean square error of the absorbance between the result spectrum and the consistency spectrum of the intumescent fire retardant coating sample to be tested, and is used as the absorbance of the second spectrum; k is the kth wavelength; n is the total number of wavelengths; y ′ k is the absorbance value of the intumescent fire retardant coating sample to be tested under the kth wavelength condition; y k is the absorbance value of the consistency spectrum at the kth wavelength.

10. The method for evaluating the quality of an intumescent fire retardant coating according to claim 1, characterized in that: When the normalized data is analyzed by the PCA algorithm, the maximum value of the characteristic parameter is used as the benchmark.

Images