Test Data Research Articles

A novel random-effect Birnbaum-Saunders distribution for reliability assessment considering accelerated mechanism equivalence

ABSTRACT Birnbaum-Saunders (B-S) distribution has been widely used in reliability assessment of products. However, the basic B-S distribution fails in characterizing the variation of damage caused by different units. In this paper, a novel random-effect B-S distribution which considers the unit heterogeneity is proposed. First, the novel distribution incorporates the inverse gamma distribution into the shape parameter of B-S distribution, and statistical inferences of the novel distribution are thoroughly investigated. Then, an extension of the proposed B-S distribution is developed to handle accelerated life test (ALT) data. In order to ensure the validity of ALTs, a coefficient of variation hypothesis test method employing the necessary condition for mechanism equivalence of the proposed distribution is executed. To estimate the parameters of the novel B-S distribution with explanatory variables, a two-step algorithm combining analogue least square and expectation maximization (ALS-EM) is proposed based on the data with constant coefficients of variation. A simulation is conducted to verify the effectiveness of the proposed ALS-EM method. Finally, two actual engineering cases are utilized to show the advantages of the random-effect B-S distribution in terms of goodness-of-fit and reliability assessment precision.

A deep learning GNSS spoofing and jamming detection method with dual-frequency C/N0 heatmap

Abstract The Global Navigation Satellite System (GNSS) is vulnerable to interference due to the open signal structure and low signal strength, posing a significant threat to the billions of terminals worldwide that rely on GNSS receivers for precise Positioning, Navigation, and Timing (PNT) services. In this paper, we propose a cloud-edge framwork for GNSS spoofing and jamming monitoring, comprising the data acquisition module, GNSS monitoring module, detecting and reporting module. In this framwork, we design a Deep Learning (DL) method for detecting GNSS interference through Dual-frequency Carrier-to-Noise density ratio (C/N0) heatmaps (DD-C/N0). This method involves extracting and correlating features from C/N0 heatmaps of visible navigation satellites operating in the GPS L1 and L2 frequency bands, allowing the identification of anomalous patterns. A U-BLOX receiver was utilized to capture the GNSS satellite signals, while commercial jammers and Software-Defined Radio (SDR) HackRF One kits were employed to simulate the interference sources. Experimental results demonstrate that the proposed method achieves significantly higher performance, with an accuracy of 99% and 98% on the public dataset and real-time testing data, compared to unsupervised, semi-supervised, and supervised detectors that rely solely on single-channel data (L1 frequency band). Integrated with the DD-C/N0 method, the online GNSS monitoring system will be improved and deployed to automate spoofing and jamming detection tasks in the next step.

Mixed training backdoor attack method improves generalization of backdoor attacks: An empirical study based on multiple models and datasets

In recent years, backdoor attacks have posed significant security threats to the training process of deep neural networks (DNNs). Attackers attempt to embed triggers in the training set, causing the victim model to behave normally when processing benign samples, but maliciously alter predictions when the hidden backdoor is activated by these triggers. However, our research found that when the trigger locations differ between the training and test data, the success rate of backdoor attacks decreases. This finding was consistent across different models and datasets through our experiments. To address the issue of decreased attack success rates caused by inconsistent trigger distributions between training and test data, we propose a mixed training backdoor attack method. This method enhances the generalization of attacks by constructing backdoor attack data with triggers in various positions during the training phase. When applying this mixed training backdoor attack approach to the BadNets and Blended attack methods, the ASR (Attack Success Rate) on test sets improved by up to 99.57% and 94.96%, respectively, across different models and datasets.

Axial compressive performance of spirally reinforced high-strength CFST column

The concrete-filled steel tubular (CFST) member using high-strength materials could have less ductility than its counterpart using normal-strength materials. In this study, spiral reinforcement is used to enhance the performance of CFST columns using high-strength materials. Experiments were conducted for spirally reinforced high-strength CFST stub columns. The highest yield strengths of spiral reinforcement and steel tube were 1597.6 MPa and 771.7 MPa, respectively, and the highest compressive strength of concrete was 103.1 MPa. Results showed that the presence of high-strength spiral reinforcement effectively improved both the strength and ductility of composite columns, while the fracture of spiral reinforcement was observed in tests. A numerical study revealed that spiral reinforcement provided effective confinement to the high-strength core concrete. The steel tube and spiral reinforcement reached respective yield strength corresponding to peak load. The spirally reinforced CFST column exhibited higher resistance than the equivalent normal CFST column using a thick tube. A design-oriented strength prediction method was proposed and was validated against the test data. The average ratio of formula-predicted to experimental resistances was 0.934, with a coefficient of variation being 0.079.

Digitizing DNA Sequences Using Multiset-Based Nucleotide Frequencies for Machine Learning-Based Mutation Detection

Investigating algebraic structures in a non-conventional framework supplements mathematics for hard-nosed practical applications to the fields of theoretical biology and computer science. One such algebraic structure is multigroup whose underlying set is a multiset. The genome is the entire set of DNA instructions found within a cell which contains all the information needed for an individual to develop and function. DNA and RNA are the hereditary materials that play a vital role in the metabolism process of living things, especially protein synthesis. In genomic database DNA sequences are stored in the form of string or text data types. The only data that works with machine learning algorithms is numerical. Thus, it is necessary to transform DNA sequence strings to numerical values. This article is organized in the following manner. Firstly, we prove that standard genetic code is a multigroup and genome architecture of the whole population can be interpreted as the sum of multisets. Next, it is described how a numerical representation of DNA bases relates to its algebraic representation. We further employed Gated Recurrent Unit, Long Short-Term Memory, and Bidirectional Long Short-Term Memory to identify changes between the DNA sequences. Experimental results show that GRU with multiset-based numerical values for DNA bases offers 98% accuracy on testing data. This novel technique will aid in the detection of mutations in complex diseases.

Research and Application of Sealed coring Technology in In-situ Coal Seam of Directional Long borehole in Coal mine

In order to accurately obtain the gas content of in-situ coal seams in coal mines, a sealed coring technology for in-situ coal seams in coal mines has been proposed. By utilizing the pressure difference generated by high-pressure water at both ends of the piston, the piston is driven to cut off the positioning pin, which in turn drives the ball valve in the coring device to rotate, achieving the goal of cutting off and sealing the in-situ coal core. Performance tests were conducted on the sealing pressure of the coring device by opening the amount of water holes on the piston and using suspension pins of different materials, verifying the working parameters of the piston opening amount and suspension pins of different materials, providing basic data for subsequent industrial underground tests. Finally, during the industrial test underground, it was found that the gas content in the coal seam measured by closed sampling was 1.9-2.5 times higher than that of the coal seam sampled by the hole, which verified the successful design of the closed sampling device.

Thickness-extensible higher order plate theory with enforced C1 continuity for the analysis of PEEK medical implants

Plate-like structures had been thoroughly studied in literature over years to reduce the computational space from 3D to 2D. Many of these theories suffer either from satisfying the free traction condition or thickness extensibility in addition to the consistency of transverse shear strain energy. This work presents a higher order shear deformation thickness-extensible plate theory (eHSDT) for the analysis of plates. The proposed eHSDT satisfies the condition of free traction as other theories do but it also satisfies the condition of consistency of transverse shear strain energy which is neglected by many theories in the area of plates and shells. The implementation of the proposed theory in displacement-based finite element procedure requires continuity of derivatives across elements. This necessary condition was achieved using the penalty enforcement method for derivative-based nodal degrees of freedom across the standard 9-nodes Lagrange element. The theory was tested for elastic bending deformation of Polyether-ether-ketone (PEEK) which is one of the basic materials for medical implants. The theory showed good accuracy compared to experimental data of the three-points bending test. The present eHSDT was also tested for different conditions with a wide range of aspects ratios (thin to thick plates) and different boundary conditions. The accuracy of the proposed eHSDT was verified against exact solutions for these conditions which showed the advantage over other approaches and commercial finite element packages.

COVID-19 Policy Decisions in Manitoba and the Experiences of the Red River Métis: A partnership-based, whole-population linked administrative data study

Systematically marginalized populations, like Red River Métis, have been greatly affected by COVID-19. Manitoba’s Indigenous COVID-19 vaccine policy initially delayed prioritization of Métis. Our research team, which included Métis partners, examined the consequences of these decisions (COVID-19 infections, health service use, vaccine uptake) among Métis, and how earlier prioritization could have improved outcomes. This retrospective cohort study linked data from the Métis Population Database to whole-population COVID testing and vaccination data, and administrative data on health service use. Restricted mean survival time models tested whether vaccination uptake differed between Métis and all other Manitobans (AOM), adjusting for sociodemographic characteristics and comorbidities. A Bayesian model will simulate how prioritization of Métis for vaccination two weeks earlier could have impacted infections. Cumulative prevalence of COVID-19 infection rates were similar among Métis and AOM until May 2021 when rates became higher among Métis. Between May and December 2021, rates of first vaccination were lower among Metis than AOM, as were second vaccination rates between July and November 2021. Métis were more likely than AOM to be hospitalized due to Covid-19 between March and August 2021 and visit physicians for COVID related reasons from October 2020 to Feb 2021 and November 2021 to March 2022. Our analyses simulated what would have occurred had Métis been prioritized for vaccination two weeks before AOM, alongside other Indigenous peoples. Understanding the experiences of Métis relative to AOM is critical to identifying public health strategies which close gaps in vaccine uptake and infections.

Peridynamics modelling of projectile penetration into concrete targets

A non-ordinary state-based peridynamics (NOSB-PD) model is proposed to simulate the projectile penetration into concrete targets. In this model, the Kong-Fang concrete material model recently proposed is firstly implemented into the NOSB-PD framework to describe the complex dynamic behavior and failures in concrete material subjected to penetration loading, and then an improved point-to-volume discrete frictional contact model is proposed to simulate the physical interaction between projectile and target. After the mesh-free discretization and explicit time integration, the proposed NOSB-PD model is used to numerically predict two sets of projectile penetration experiments into low-strength and high-strength concrete targets. And numerical predictions are found to be in good agreements with corresponding test data including penetration depth, projectile deceleration, deformation of projectile and failures in concrete targets.

Investigation of a Tube-Launched Unmanned Aerial Vehicle with a Variable-Sweep Wing

Foldable wings are designed for tube-launched unmanned aerial vehicles (UAVs), aiming to improve portability and meet launch platform requirements. However, conventional tube-launched UAVs cannot operate across the wide speed ranges required for the performance of multiple missions, due to the fixed configuration of their wings after launch. This study therefore proposes a tube-launched UAV which can change wing-sweep angle to expand the flight speed range and enhance the UAV’s agility. A computational aerodynamics method is employed to assess the transient aerodynamic performance of the UAV during the sweep morphing process. The simulation results indicate that the transient aerodynamic forces generate a dynamic hysteresis loop around the quasi-steady data. The lift and drag coefficients exhibit maximum relative deviations of 18.5% and 12.7% from the quasi-steady data for the sweep morphing period of 0.5 s. The hysteresis effect of the flow structure, rather than the additional velocity resulting from wing-sweep morphing, is the major contributor to the aerodynamic hysteresis loop. Compared to the conventional tube-launched UAVs, the proposed tube-launched UAV with a variable-sweep wing shows a wider flight speed range, from 22.59 to 90.12 m/s, and achieves an 82.84% increase in loitering speed. To verify the effectiveness of the wing-sweeping concept, a prototype was developed, and a flight test was carried out. The test data obtained from flight control system agree well with the simulation data, which demonstrates the feasibility and effectiveness of the variable-sweep wing in widening the speed range for tube-launched UAVs. This work can provide a reference for the design of tube-launched UAVs for wide speed range flight.

Personal characteristics and transmission dynamics associated with SARS-CoV-2 semi-quantitative PCR test results: an observational study from Belgium, 2021-2022.

Following harmonization efforts by the Belgian National Reference Center for SARS-CoV-2, semi-quantitative PCR test (SQ-PCR) results, used as a proxy for viral load, were routinely collected after performing RT-qPCR tests. We investigated both the personal characteristics associated with SQ-PCR results and the transmission dynamics involving these results. We used person-level laboratory test data and contact tracing data collected in Belgium from March 2021 to February 2022. Personal characteristics (age, sex, vaccination, and laboratory-confirmed prior infection) and disease stage by date of symptom onset were analyzed in relation to SQ-PCR results using logistic regression. Vaccine effectiveness (VE) against a high viral load (≥107 copies/mL) was estimated from the adjusted probabilities. Contact tracing involves the mandatory testing of high-risk exposure contacts (HREC) after contact with an index case. Odds ratios for test positivity and high viral load in HREC were calculated based on the SQ-PCR result of the index case using logistic regression models adjusted for age, sex, immunity status (vaccination, laboratory-confirmed prior infection), variant (Alpha, Delta, Omicron), calendar time, and contact tracing covariates. We included 909,157 SQ-PCR results of COVID-19 cases, 379,640 PCR results from index cases, and 72,052 SQ-PCR results of HREC. High viral load was observed more frequently among recent cases, symptomatic cases, cases over 25 years of age, and those not recently vaccinated (>90 days). The vaccine effectiveness (VE) of the primary schedule in the first 30 days after vaccination was estimated at 47.3% (95%CI 40.8-53.2) during the Delta variant period. A high viral load in index cases was associated with an increased test positivity in HREC (OR 2.7, 95%CI 2.62-2.79) and, among those testing positive, an increased likelihood of a high viral load (OR 2.84, 95%CI 2.53-3.19).

Method for determining reagent optimum volume for acid treatments of carbonate reservoirs based on a pressure build-up curve method

Relevance. The need to increase the efficiency of acid treatments of carbonate reservoirs. Often, the design of treatments does not take into account the actual operating conditions of production wells, which leads to an overestimation or underestimation of the required volume of the reagent. Aim. To propose a method for determining the optimal amount of reagent for carrying out acid treatments of carbonate reservoirs in order to restore the permeability of the rock in the near-well zone, taking into account the actual operating conditions of the well, which allows rationalizing the use of subsoil user resources. Objects. Producing oil wells exploiting oil deposits in carbonate deposits of the Perm region, as well as the results of hydrodynamic studies carried out on these wells. Methods. Theoretical substantiation of the possibility of determining the zone of deteriorated permeability around the well using the pressure recovery curve graph, graphical processing of diagnostic graphs of the pressure recovery curve, mathematical calculation of technological indicators based on hydrodynamic testing data of wells, integration of inflow volumes, analysis of technological data and conditions of applicability of the methodology. Results. The authors have developed the methodology for determining the optimal volume of reagent for acid treatments of carbonate reservoirs based on the pressure recovery curve method. The optimal volume of reagent for 21 production wells was calculated. Using the example of 13 wells, the efficiency of the injected volume of reagent was determined and an overestimated volume of reagent injected into the formation was demonstrated. The authors assessed the subsoil user's costs taking into account the excess consumption of the reagent. The factors influencing the success of the method application depending on the operating conditions of the wells were identified. Criteria inappropriateness of acid treatment were identified.

Advancing Model Generalization in Continuous Cyclic Test-Time Adaptation with Matrix Perturbation Noise

Test-time adaptation (TTA) aims to optimize source-pretrained model parameters to target domains using only unlabeled test data. However, traditional TTA methods often risk overfitting to the specific, localized test domains, leading to compromised generalization. Moreover, these methods generally presume static target domains, neglecting the dynamic and cyclic nature of real-world settings. To alleviate this limitation, this paper explores the continuous cyclic test-time adaptation (CycleTTA) setting. Our unique approach within this setting employs matrix-wise perturbation noise in batch-normalization statistics to enhance the adaptability of source-pretrained models to dynamically changing target domains, without the need for additional parameters. We demonstrated the effectiveness of our method through extensive experiments, where our approach reduced the average error by 39.8% on the CIFAR10-C dataset using the WideResNet-28-10 model, by 38.8% using the WideResNet-40-2 model, and by 33.8% using the PreActResNet-18 model. Additionally, on the CIFAR100-C dataset with the WideResNet-40-2 model, our method reduced the average error by 5.3%, showcasing significant improvements in model generalization in continuous cyclic testing scenarios.

The Diagnostic Ability of GPT-3.5 and GPT-4.0 in Surgery: Comparative Analysis.

ChatGPT (OpenAI) has shown great potential in clinical diagnosis and could become an excellent auxiliary tool in clinical practice. This study investigates and evaluates ChatGPT in diagnostic capabilities by comparing the performance of GPT-3.5 and GPT-4.0 across model iterations. This study aims to evaluate the precise diagnostic ability of GPT-3.5 and GPT-4.0 for colon cancer and its potential as an auxiliary diagnostic tool for surgeons and compare the diagnostic accuracy rates between GTP-3.5 and GPT-4.0. We precisely assess the accuracy of primary and secondary diagnoses and analyze the causes of misdiagnoses in GPT-3.5 and GPT-4.0 according to 7 categories: patient histories, symptoms, physical signs, laboratory examinations, imaging examinations, pathological examinations, and intraoperative findings. We retrieved 316 case reports for intestinal cancer from the Chinese Medical Association Publishing House database, of which 286 cases were deemed valid after data cleansing. The cases were translated from Mandarin to English and then input into GPT-3.5 and GPT-4.0 using a simple, direct prompt to elicit primary and secondary diagnoses. We conducted a comparative study to evaluate the diagnostic accuracy of GPT-4.0 and GPT-3.5. Three senior surgeons from the General Surgery Department, specializing in Colorectal Surgery, assessed the diagnostic information at the Chinese PLA (People's Liberation Army) General Hospital. The accuracy of primary and secondary diagnoses was scored based on predefined criteria. Additionally, we analyzed and compared the causes of misdiagnoses in both models according to 7 categories: patient histories, symptoms, physical signs, laboratory examinations, imaging examinations, pathological examinations, and intraoperative findings. Out of 286 cases, GPT-4.0 and GPT-3.5 both demonstrated high diagnostic accuracy for primary diagnoses, but the accuracy rates of GPT-4.0 were significantly higher than GPT-3.5 (mean 0.972, SD 0.137 vs mean 0.855, SD 0.335; t285=5.753; P<.001). For secondary diagnoses, the accuracy rates of GPT-4.0 were also significantly higher than GPT-3.5 (mean 0.908, SD 0.159 vs mean 0.617, SD 0.349; t285=-7.727; P<.001). GPT-3.5 showed limitations in processing patient history, symptom presentation, laboratory tests, and imaging data. While GPT-4.0 improved upon GPT-3.5, it still has limitations in identifying symptoms and laboratory test data. For both primary and secondary diagnoses, there was no significant difference in accuracy related to age, gender, or system group between GPT-4.0 and GPT-3.5. This study demonstrates that ChatGPT, particularly GPT-4.0, possesses significant diagnostic potential, with GPT-4.0 exhibiting higher accuracy than GPT-3.5. However, GPT-4.0 still has limitations, particularly in recognizing patient symptoms and laboratory data, indicating a need for more research in real-world clinical settings to enhance its diagnostic capabilities.

Prediction of H3K27M alteration in midline gliomas of brain using radiomics- A multi-institute study

Abstract Background Non-invasive prediction of H3K27M-altered Diffuse midline gliomas is important because of the involvement of deep locations and proximity to eloquent structures. We aim to predict H3K27M alteration in midline gliomas using radiomics features of T2W images. Methods Radiomics features extracted from 124 subjects (69 H3K27M-altered/55 H3K27M-wildtype). T2W-images were resampled to 1x1x1mm voxel-size, pre-processed, and normalized for artefact correction, intensity variations. Feature-set was normalized, subjected to reduction by variance thresholding, correlation coefficient thresholding, and sequential feature selector. Adaptive synthesis oversampling technique was used to over-sample the training data. Random forest classifier (RFC), Decision tree classifier (DTC) and K-nearest neighbors classifier (KNN) were trained over the training dataset and the performance was assessed over the internal test dataset and external test data set (52 subjects: 33 H3K27M-altered/19-H3K27M-wildtype). Results DTC achieved validation score of 77.33% (5-fold cross-validation) and accuracy of 80.64%, 75% on internal and external test dataset. RFC achieved validation score of 80.7% (5-fold cross-validation) and accuracy of 80.6%, 73% on internal and external test dataset. DTC achieved validation score of 78.67% (5-fold cross-validation) and accuracy of 80.64%, 61.53% on internal and external test dataset. Accuracy score of DTC, RFC, and KNN on internal test dataset was approximately 80% while on the external test dataset, DTC achieved 75% accuracy, RFC achieved 73% accuracy and KNN achieved 65.1% accuracy. Conclusion H3K27M alteration is a potential immunotherapeutic marker and is associated with poor prognosis and radiomics features extracted from conventional T2W-images can help in identifying H3K27M-altered cases non-invasively with high precision.

Optimization of Daily Warehouse Storage PT. XYZ with Shared Storage Method

PT XYZ is a company engaged in ship repair and maintenance services and port management services. The problems faced by PT XYZ in carrying out its business lines are related to the effective and efficient operational managerial activities of raw material warehousing. Problem solving steps are carried out by testing data processing and analysis using the shared storage method as a solution step related to the problem of warehousing layout and optimization. Problem solving with the shared storage method is a method of reconditioning the warehousing layout with the principle of First in First out or fast moving where the layout of goods is classified based on the order of the flow of goods movement according to the operational needs of PT XYZ activities. Based on the calculation results, if the company applies the proposed layout, there is an average distance savings of 4 m2. The layout of the placement of raw materials based on the proposed improvements has a more organized condition and facilitates the flow of goods movement as well as inputting the flow of goods in and out which can be more efficient if the goods are organized and inputted administratively, this also affects the quality of goods.

A Mixed Convection Model for Estimating the Critical Velocity to Prevent Smoke Backlayering in Tunnels

AbstractA novel mathematical model for the critical ventilation velocity to prevent smoke backlayering in tunnels is presented, addressing limitations of prior approaches. The basis of the model is a rigorous characterisation of the physical processes by the characteristic quantities. Empirical parameters within the new model are determined, to align with results from both full-size and small-scale tunnel experiments. Data from numerical simulations (CFD, Computational Fluid Dynamics), validated by known test data, are then used to estimate the effects of tunnel slope and other parameters on the critical velocity. The model is seen to approximate the critical velocity well, following all trends identified by test data and CFD parameter studies. The empirically calibrated equation permits prediction of the critical velocity beyond the narrow range of tunnel geometries where known results already give an answer. The resulting equation has practical application for tunnel design.

Efficient compressed storage and fast reconstruction of large binary images using chain codes

AbstractLarge binary images are used in many modern applications of image processing. For instance, they serve as inputs or target masks for training machine learning (ML) models in computer vision and image segmentation. Storing large binary images in limited memory and loading them repeatedly on demand, which is common in ML, calls for efficient image encoding and decoding mechanisms. In the paper, we propose an encoding scheme for efficient compressed storage of large binary images based on chain codes, and introduce a new single-pass algorithm for fast parallel reconstruction of raster images from the encoded representation. We use three large real-life binary masks to test the efficiency of the proposed method, which were derived from vector layers of single-class objects – a building cadaster, a woody vegetation landscape feature map, and a road network map. We show that the masks encoded by the proposed method require significantly less storage space than standard lossless compression formats. We further compared the proposed method for mask reconstruction from chain codes with a recent state-of-the-art algorithm, and achieved between $$12\%$$ 12 % and $$33\%$$ 33 % faster reconstruction on test data.

SP5.13 - Assessing the use of electronic QLIK app in Mortality and Morbidity meetings compared with traditional manual hand collected data in colorectal surgery

Abstract Aims To validate and compare the effectiveness of the QLIK app as an electronic method of data collection compared to traditional manual methods for mortality and morbidity meetings (M&amp;Ms). Methods All data from the theatre operating management system was linked in the Trust Data Warehouse to clinical test and vital statistics data and displayed using a QLIK App. A retrospective study was conducted comparing the manual collection of data of elective colorectal M&amp;M with the QLIK app data for the years 2021 and 2022. Data collected included: total operations, number of deaths, readmissions, reoperations and post-operative complications. The combined data was then analysed in three different rounds of validation. Results There was a high level of concordance between the manual data and the QLIK app, with both correctly identifying the same number of deaths. In 2021 the app recorded 569 total operations compared to 584 in the manual data, and 127 post operative complications to 126 respectively. In 2022, the app recorded 607 operations to 489 in the manual data, and 120 complications compared to 149. Of the complications recorded in the manual data but not the app, 96.8% of these were Clavien-Dindo Grade 1 and 2 complications. Manual data collection of the M&amp;M data was estimated to take approximately 250 hours per year from collection to analysis, whilst the app displayed the data in near real time. Conclusions Use of data visualization tools and coded electronic hospital data for M&amp;Ms can minimise errors and allow for greater efficiency. This approach gives real time surgical performance data.

Experimental correlations to predict the cooling performance and compressor energy consumption of domestic refrigerator with phase change material

ABSTRACT The phase change material technology is continuously evolving in commercial market. In order to develop/design PCM modules of different capacity and different PCM material, it is necessary to develop the equation which will provide required information for development of PCM module. This paper establishes correlations to predict backup time and equivalent energy consumption of domestic refrigerator based on experimental data of testing of PCMs, potassium chloride (KCl), sodium chloride (NaCl), and sodium fluoride (NaF) dissolved in water for PCM masses varying between 0 and 1.75 kg in both closed door and door opening conditions. This paper investigates the impact of incorporating a phase-change material (PCM) on the cooling performance and energy consumption of a domestic refrigerator. The PCM module, filled with potassium chloride (KCl), sodium chloride (NaCl), and sodium fluoride (NaF) dissolved in water, is strategically placed on the contact surface of the refrigerator’s evaporator coil. The utilization of PCM aims to mitigate cabin temperature fluctuations and reduce overall energy consumption. Through experimental evaluation, the study assesses the effectiveness of different phase-change materials in enhancing the cooling performance and energy efficiency of the domestic refrigerator. The findings reveal that the incorporation of phase-change material significantly improves the cooling performance, as evidenced by the extended backup time during power cut-off and a reduction in compressor energy consumption. The tested phase-change materials, including KCl, NaCl, and NaF, demonstrate varying impacts on performance parameters. The backup time for closed-door tests is 1.83 – 1.03 times longer than that for door opening tests, respectively. The backup time of door opening tests is shortened due to heat entered to the cabin for all masses, between 45.45% and 3.23%. The backup time is increased by the masses of PCM KCl between 14% and 155%, the masses of PCM NaCl between 18% and 159%, and the masses of PCM NaF between 27% and 182%. The equivalent energy consumption for door opening tests is 1.12–1.02 times larger than that for the closed-door tests. The correlation equation established based on the existing experimental data can be used to accurately predict the backup time and equivalent energy consumption of refrigerator for tested PCM of any mass within the interval of 0 Kg to 1.75 Kg with ± 10% accuracy.