Inherent Randomness Research Articles

Short-Term Wind Speed Prediction via Sample Entropy: A Hybridisation Approach against Gradient Disappearance and Explosion

High-variant wind speeds cause aberrations in wind power systems and compromise the effective operation of wind farms. A single model cannot capture the inherent wind speed randomness and complexity. In the proposed hybrid strategy, wavelet transform (WT) is used for data decomposition, sample entropy (SampEn) for subseries complexity evaluation, neural network autoregression (NNAR) for deterministic subseries prediction, long short-term memory network (LSTM) for complex subseries prediction, and gradient boosting machine (GBM) for prediction reconciliation. The proposed WT-NNAR-LSTM-GBM approach predicts minutely averaged wind speed data collected at Southern African Universities Radiometric Network (SAURAN) stations: Council for Scientific and Industrial Research (CSIR), Richtersveld (RVD), Venda, and the Namibian University of Science and Technology (NUST). For comparison purposes, in WT-NNAR-LSTM-GBM, LSTM and NNAR are respectively replaced with a k-nearest neighbour (KNN) to form the corresponding hybrids: WT-NNAR-KNN-GBM and WT-KNN-LSTM-GBM. We assessed WT-NNAR-LSTM-GBM’s efficacy against NNAR, LSTM, WT-NNAR-KNN-GBM, and WT-KNN-LSTM-GBM as well as the naïve model. The comparative study found that the WT-NNAR-LSTM-GBM model was the most accurate, sharpest, and robust based on mean absolute error, median absolute deviation, and residual analysis. The study results suggest using short-term forecasts to optimise wind power production, enhance grid operations in real-time, and open the door to further algorithmic enhancements.

Exploring chaos and ergodic behavior of an inductorless circuit driven by stochastic parameters

There exist extensive studies on periodic and random perturbations of various smooth maps investigating their dynamics. Unlike smooth maps, non-smooth maps are yet to be studied extensively under a stochastic regime. This paper presents a stochastic piecewise-smooth map derived from a simple inductorless switching circuit. The stochasticity is introduced in parameter values. The distribution of the parameter values is bounded and randomly selected from uniform and triangular distributions and ranges between high and low bifurcation parameter values of the deterministic map. Due to this inherent stochasticity in parameter values, the time evolution of the state variable cannot be predicted at a specific time instant. We observe that the state variable exhibits completely ergodic behavior when the minimum value of the parameter is the same as the minimum bifurcation parameter of the deterministic system. However, the ensemble average of the state variable converges to a fixed value. The system demonstrates nonchaotic behavior for a particular range of parameter values but the deterministic map in that bifurcation range shows interplay between chaos and periodic orbits. The values of Lyapunov exponents decrease monotonically with increased asymmetry of the distribution from which the bifurcation parameter values are chosen. We determine the probability density function of the stochastic map and verify its invariance under initial conditions. The most noteworthy result is the disappearance of chaotic behavior when the lower range of the distribution is varied while maintaining a fixed upper threshold for a particular distribution, even though the deterministic map exhibits an array of periodic and chaotic behaviors within the range. As the period-incrementing cascade with chaotic inclusion only occurs in nonsmooth maps, this paper numerically shows the stochasticity of a piecewise-smooth map obtained from a practical system for the first time where randomness is introduced in the parameter space.

The randomness and determinacy of wall pressure fluctuations in incompressible flow

Wall pressure fluctuations caused by turbulent boundary layers have a significant impact on aircraft structural vibration and cabin noise. This study aims to investigate the mechanism of turbulence-induced pressure fluctuations by focusing on the randomness of wall pressure fluctuations, analyzed in both the time–frequency and spatial-wavenumber domains using measured data obtained from a phase array in a wind tunnel. Three roughness elements were designed and installed upstream of the plate to manipulate the turbulent boundary layer at a specific Mach number. The results of the investigation demonstrate that the disturbance strength induced by the roughness element influences the randomness of wall pressure fluctuations, in addition to the parameters utilized for data analysis. Generally, stronger turbulence fluctuations tend to decrease the randomness of pressure fluctuations. Moreover, wall pressure fluctuations also exhibit certain statistical principles that cannot be precisely calculated using mathematical expressions, highlighting their inherent randomness. Further investigation into randomness in the spatial-wavenumber domain revealed the hydrodynamic modes of turbulence fluctuations with varying convection velocity analyzed through wavenumber maps computed using the beamforming algorithm. These modes with variable convective speed significantly contribute to the generation of randomness in wall pressure fluctuations. Both the time–frequency domain and the spatial-wavenumber domain affect the randomness characteristics of wall pressure fluctuations. However, such effects are not easily discernible through a rudimentary analysis of the space–time correlation of turbulence fluctuations.

Channel assisted noise propagation in a two-step cascade.

Signal propagation in biochemical networks is characterized by the inherent randomness in gene expression and fluctuations of the environmental components, commonly known as intrinsic and extrinsic noise, respectively. We present a theoretical framework for noise propagation in a generic two-step cascade (S→X→Y) regarding intrinsic and extrinsic noise. We identify different channels of noise transmission that regulate the individual and the overall noise properties of each component. Our analysis shows that the intrinsic noise of S alleviates the general noise and information transmission capacity along the cascade. On the other hand, the intrinsic noise of X and Y acts as a bottleneck of information transmission. We also show a hierarchical relationship among the intrinsic noise levels of S, X, and Y, with S exhibiting the highest level of intrinsic noise, followed by X and then Y. This hierarchy is preserved within the two-step cascade, facilitating the highest information transmission from S to Y via X.

A novel RUL prediction method for rolling bearings based on dynamic control chart and adaptive incremental filtering

Abstract Degradation of rolling bearings typically consists of two stages: a stable stage (Stage I) characterized by stable fluctuations in the health indicator (HI), and a degradation stage (Stage II) where early damage leads to HI degradation, eventually reaching the failure threshold. Therefore, to achieve online RUL prediction for bearings, three aspects should be studied: 1) Degradation modeling; 2) Inter stage change point identification; 3) Online degradation state updating. Firstly, a two-stage degradation model is constructed by simultaneously considering inherent randomness, individual differences, and measurement errors. Then, a dynamic statistical process control (SPC) method is proposed to identify the change point from Stage I to Stage II. The SPC is designed to dynamically control limits based on the bearing's condition monitoring (CM) data to prevent false alarms. An adaptive incremental filtering (AIF) is proposed to update the degradation states by simultaneously considering the state increment and the dynamics of the system noise and measurement noise. The effectiveness of the proposed method is validated on 16004 bearing test data and XJTU-SY bearing data. Results show that the proposed method can accuracy identify the change point and improve the accuracy of the prediction result during stage II.

Procrystalline Self-Assembly of Desymmetrized Pentaphenylcyclopentadiene.

The interplay between the molecular shape and the intermolecular interaction plays a decisive role in self-assembled structures. Recently, inherent randomness of low ordered assemblies, resulting from lack of short- and long-range periodicities, has attracted significant attention due to the unique structural, electronic, and mechanical properties. Here, we present procrystalline self-assemblies of pentaphenyl cyclopentadienyl derivatives on Ag(111) and Au(111) with scanning tunneling microscopy, operating at 4.3 K under ultrahigh vacuum conditions. Two examples, using 5-fold symmetric molecules substituted with methyl or fluorine groups, show that weak interactions, such as π-π stacking, CH-π interactions, and CH···F hydrogen bonding, play a pivotal role in formation of the procrystalline assembly. Our results may give insights into the intricate relationship between the molecular shape and the intermolecular interaction in the formation of non-crystalline assemblies.

Text classification for distribution substation inspection based on BERT-TextRCNN model

With the advancement of source-load interaction in the new power systems, data-driven approaches have provided a foundational support for aggregating and interacting between sources and loads. However, with the widespread integration of distributed energy resources, fine-grained perception of intelligent sensing devices, and the inherent stochasticity of source-load dynamics, a massive amount of raw data is being recorded and accumulated in the data center. Valuable information is often dispersed across different paragraphs of the raw data, making it challenging to extract effectively. Distribution substation inspection plays a crucial role in ensuring the safe operation of the power system. Traditional methods for inspection report text classification typically rely on manual judgment and accumulated experience, resulting in low efficiency and a significant misjudgment rate. Therefore, this paper proposes a text classification method for inspection reports based on the pre-trained BERT-TextRCNN model. By utilizing the dense connection between the BERT embedding layer and the neural network, the proposed method improves the accuracy of matching long texts. This article collected 2,831 maintenance data for the first quarter of 2023 from the distribution room, including approximately 58 environmental testing data, 738 environmental box testing data, approximately 672 distribution room testing data, and approximately 1,363 box type substation testing data. A text corpus was constructed for experiments. Experimental results demonstrate that the proposed model automatically classifies a large volume of manually recorded inspection report data based on time, location, and faults, achieving a classification accuracy of 94.7%, precision of 92%, recall of 92%, and F1 score of 90.3%.

Interpretable feature-temporal transformer for short-term wind power forecasting with multivariate time series

The inherent randomness and volatility of wind power generation present significant challenges to the reliable and secure operation of the power system. Therefore, it is crucial to have interpretable wind power forecasting (WPF) to ensure seamless grid integration and effective risk assessment. Existing forecasting models often focus on improving WPF performance and ignore the interpretability of the model, resulting in ambiguous forecasting. In this paper, the interpretable feature-temporal transformer (IFTT) for short-term wind power forecasting with multivariate time series is presented. The model uses an encoder-decoder architecture to effectively integrate historical information and future prior information from multiple variables. The designed decoupled feature-temporal self-attention (DFTA) module and variable attention network (VAN) effectively realize the interpretability of temporal information and multi-variable inputs while extracting important features. The Auxiliary Forecasting Network (AFN) plays a key role in providing pseudo-future wind speed predictions, which serve as an essential input for the model's decoder, and enhancing forecasting accuracy through multi-task learning. Experimental results on multiple datasets in different geographical locations show that the proposed algorithm is superior to various advanced methods. Besides, the interpretability of the IFTT model offers valuable insights for ensuring the safety of wind power utilization and supporting informed risk decision-making.

Geological Modeling of Shale Oil in Member 7 of the Yanchang Formation, Heshui South Area, Ordos Basin

In recent years, the Chang 7 member of the Mesozoic Triassic Yanchang Formation in the Ordos Basin has emerged as a significant repository of abundant and distinctive unconventional oil resources. The Heshui area boasts substantial shale oil reserves, with reported third-level reserves surpassing 600 million tons. However, the region in the southern part of Heshui is marked by pronounced variability in reservoir quality, intricate oil–water dynamics, low formation energy, and suboptimal fluid properties, leading to divergent development outcomes for horizontal wells. There is an imperative need to devise and refine new geological models to underpin the efficient exploitation of shale oil in the southern Heshui area. This study focuses on the shale oil reservoir of the Chang 7 member in the southern Heshui area of the Ordos Basin, conducting detailed stratigraphic correlation and establishing a refined isochronous stratigraphic framework. Utilizing PetrelTM modeling software (version 2018), we integrate deterministic and stochastic modeling approaches, adhering to the principles of isochronous and phased modeling. By assessing the thickness of sand and mudstone layers and the overall stratigraphic sequence, we derive a geological probability surface. Subsequently, this surface is harnessed to constrain the lithofacies, yielding a constrained lithofacies model. Employing sequential indicator simulation and sequential Gaussian stochastic simulation, we develop a reservoir attribute model that is anchored in the lithofacies model and its controls, culminating in a robust and dependable static model. Employing the geological probability surface constraint method, we meticulously construct the reservoir matrix model, amalgamating individual well data with the inherent certainty and randomness of reservoir plane thickness. This approach further enhances the model’s accuracy and mitigates the uncertainty and randomness associated with inter-well interpolation to a significant degree.

Smart Industrial Internet of Things Framework for Composites Manufacturing.

Composite materials are increasingly important in making high-performance products. However, contemporary composites manufacturing processes still encounter significant challenges that range from inherent material stochasticity to manufacturing process variabilities. This paper proposes a novel smart Industrial Internet of Things framework, which is also referred to as an Artificial Intelligence of Things (AIoT) framework for composites manufacturing. This framework improves production performance through real-time process monitoring and AI-based forecasting. It comprises three main components: (i) an array of temperature, heat flux, dielectric, and flow sensors for data acquisition from production machines and products being made, (ii) an IoT-based platform for instantaneous sensor data integration and visualisation, and (iii) an AI-based model for production process forecasting. Via these components, the framework performs real-time production process monitoring, visualisation, and prediction of future process states. This paper also presents a proof-of-concept implementation of the framework and a real-world composites manufacturing case study that showcases its benefits.

Fast and Accurate Recognition of Perovskite Fluorescent Anti-counterfeiting Labels Based on Lightweight Convolutional Neural Networks.

Anti-counterfeiting technology has always been a key issue in the field of information security. Physical Unclonable Function (PUF) labels, which are random patterns produced by a stochastic process, emerge as an effective anti-counterfeiting strategy due to the inherent randomness of their physical patterns. In this study, we developed a high-throughput droplet array generation technique based on surface tension confinement to prepare perovskite crystal films with controllable shapes and sizes. We utilized the random distribution of perovskite nanocrystal particles to construct the PUF textures of the labels. Compared to other anti-counterfeiting labels, our labels not only possess fluorescent properties but also feature microscale dimensions (less than 5.3 × 10-2mm2), low cost (less than 3 × 10-4 USD), and high encoding capacity (1.7 × 101956), providing support for multilevel anti-counterfeiting protection. Additionally, we introduce an innovative PUF recognition method based on a Partial Convolutional Network (PaCoNet), effectively addressing the limitations of previous methods, in terms of recognition accuracy and speed. Experimental validation on a data set of perovskite nanocrystal films with up to 60 different macroscopic shapes and unique microscopic textures demonstrates that our method achieves a recognition accuracy of up to 99.65% and significantly reduces the recognition time per image to just 0.177 s, highlighting the potential application of these labels in the field of anti-counterfeiting.

Novel Frequency Regulation Scenarios Generation Method Serving for Battery Energy Storage System Participating in PJM Market

As one of the largest frequency regulation markets, the Pennsylvania-New Jersey-Maryland Interconnection (PJM) market allows extensive access of Battery Energy Storage Systems (BESSs). The designed signal regulation D (RegD) is friendly for use with BESSs with a fast ramp rate but limited energy. Designing operating strategies and optimizing the sizing of BESSs in this market are significantly influenced by the regulation signal. To represent the inherent randomness of the RegD signal and reduce the computational burden, typical frequency regulation scenarios with lower resolution are often generated. However, due to the rapid changes and energy neutrality of the RegD signal, generating accurate and representative scenarios presents challenges for the methods based on shape similarity. This paper proposes a novel probability-based method for generating typical regulation scenarios. The method relies on the joint probability distribution of two features with a 15-min resolution, extracted from the RegD signal with a 2 s resolution. The two features can effectively portray the characteristic of RegD signal and its influence on BESS operation. Multiple regulation scenarios are generated based on the joint probability distributions of these features at first, with the final typical scenarios chosen based on their probability distribution similarity to the actual distribution. Utilizing regulation data from the PJM market in 2020, this paper validates and analyzes the performance of the generated typical scenarios in comparison to existing methods, specifically K-means clustering and the forward scenarios reduction method.

Precise single step and multistep short-term photovoltaic parameters forecasting based on reduced deep convolutional stack autoencoder and minimum variance multikernel random vector functional network

Accurate prediction of photovoltaic (PV) power can significantly alleviate energy crises. However, the inherent randomness and intermittency of PV power pose challenges to the stability and safety of PV-penetrated grid systems. To address this, we have developed a novel hybrid model: a reduced deep convolutional stack autoencoder with a minimum variance multikernel random vector functional link network (RDCSAE-MVMRVFLN). This model enhances grid efficacy and safety. We extract the most informative band-limited intrinsic mode functions (BLIMFs) of highly nonlinear and nonstationary solar energy parameters using an entropy, kurtosis, and correlation coefficient-based information-oriented variational mode decomposition (IOVMD). These efficient BLIMFs are concatenated and input into the RDCSAE for rich, abstract, and discriminative representation computation. A less computationally complex MVMKRVFLN regression method, incorporating these refined representations, is proposed for superior prediction accuracy and reduced computational complexity. Our method shows exceptional performance in predicting solar temperature, irradiation, and power for multi-horizon forecasts with minimal error metrics (correlation coefficients of 0.999±0.001, 0.992±0.001, 0.986±0.02 and 0.978±0.02, and RMSE of 0.016±0.001, 0.024±0.001, 0.034±0.001 and 0.045±0.001 for the interval of 10 minutes, 30 minutes, 1 hour and 3 hours respectively) in both single-step and multi-step forecasting compared to conventional methods. The RDCSAE-MVMKRVFLN model is implemented on a high-speed Xilinx Virtex-5 FPGA embedded processor to validate its simplicity, robustness, and practicability. Additionally, we examine the prediction performance using real-time data from a 1 MW solar farm in Odisha, India, demonstrating the model’s effectiveness and superiority.

Using Crested Porcupine Optimizer Algorithm and CNN-LSTM-Attention Model Combined with Deep Learning Methods to Enhance Short-Term Power Forecasting in PV Generation

Due to the inherent intermittency, variability, and randomness, photovoltaic (PV) power generation faces significant challenges in energy grid integration. To address these challenges, current research mainly focuses on developing more efficient energy management systems and prediction technologies. Through optimizing scheduling and integration in PV power generation, the stability and reliability of the power grid can be further improved. In this study, a new prediction model is introduced that combines the strengths of convolutional neural networks (CNNs), long short-term memory (LSTM) networks, and attention mechanisms, so we call this algorithm CNN-LSTM-Attention (CLA). In addition, the Crested Porcupine Optimizer (CPO) algorithm is utilized to solve the short-term prediction problem in photovoltaic power generation. This model is abbreviated as CPO-CLA. This is the first time that the CPO algorithm has been introduced into the LSTM algorithm for parameter optimization. To effectively capture univariate and multivariate time series patterns, multiple relevant and target variables prediction patterns (MRTPPs) are employed in the CPO-CLA model. The results show that the CPO-CLA model is superior to traditional methods and recent popular models in terms of prediction accuracy and stability, especially in the 13 h timestep. The integration of attention mechanisms enables the model to adaptively focus on the most relevant historical data for future power prediction. The CPO algorithm further optimizes the LSTM network parameters, which ensures the robust generalization ability of the model. The research results are of great significance for energy generation scheduling and establishing trust in the energy market. Ultimately, it will help integrate renewable energy into the grid more reliably and efficiently.

Robust optimization dispatch for PV rich power systems considering demand response and energy storage systems

In recent years, the ever-rising penetration of distributed photovoltaics (PV) power has presented substantial challenges in power system dispatch due to its inherent randomness and unpredictability. To bridge this gap, this paper proposes a two-stage robust optimization method for power system security dispatch considering traditional generators as well as flexible resources, such as load demand response and energy storage systems. Specifically, a price-based demand response model is established to optimize the system’s load curve during a day. On this basis, a two-stage optimization problem for day-ahead and intra-day power system dispatch model is proposed. The dispatch objective is to minimize the overall cost in worst-case scenarios through properly scheduling unit commitment (UC) as well as flexible resources in each dispatch interval. Column and constraint generation (C&CG) algorithm is adopted for problem solving. The effectiveness of the proposed method is validated by case studies based on a modified 6-node system and a 24-node system. Simulation results indicate that through appropriately scheduling the energy storage system and load demand response, the proposed dispatch method can significantly reduce the total operation cost of a PV rich power system, which in turn facilitates the integration of PV power.

Dynamic production capacity assessment of aircraft overhaul shop based on digital twin

Aircraft overhaul is of the highest level in aircraft MRO (Maintenance, Repair, Overhaul), which is an important and necessary activity for ensuring safety and reliability of aircraft throughout its whole lifecycle. Production capacity of aircraft overhaul shop has huge strategic value and determines the cycle time and throughput of the overhauled aircraft. However, due to the inherent complexity and stochasticity, aircraft overhaul tasks are generally executed manually by skilled workers, whose stochastic task execution time (TET) makes it challenging to assess the production capacity of aircraft overhaul shop. Digital twin provides an up-to-date technical measure to simulate, evaluate, and predict the whole lifecycle process of its physical entity, which makes it possible to tackle the complexity and stochasticity of aircraft overhaul process. This paper proposes to apply digital twin to assess the production capacity of aircraft overhaul shop dynamically. Firstly, the aircraft overhaul process and its dynamic production capacity assessment are analyzed. A modeling framework of aircraft overhaul shop digital twin is proposed with a novel twin fusion mechanism. Then, an improved kernel density estimation (KDE) method and a quota hour density proportion (QHDP) method are proposed and verified to process stochastic TET for production capacity assessment. Finally, the digital twin of a typical aircraft avionics repair shop is constructed to assess its production capacity. Four assessment results are compared with real observations from the aircraft avionics repair shop, which verifies the feasibility and advantage of the proposed methodology. A bottleneck analysis is also conducted to optimize the production capacity, which provides managerial insights.

Operational reliability evaluation integrating prediction models for enhanced situational awareness in wind-integrated system

The extensive implementation of wind power generation (WPG) challenges power systems' flexibility and reliability during operations. Fluctuating wind speeds lead to variability in generation within wind energy systems, directly impacting operational reliability. Therefore, evaluating wind energy systems' situational awareness (SA) is imperative to forecast two-day-ahead operational reliability, facilitating effective power system planning. The foundation of SA for wind energy systems lies in monitoring the wind profile. This involves gaining insights into the system's state through perceiving wind speed and WPG, comprehending the dynamics of its state, and subsequently predicting operational reliability. To achieve this, a Simulink model of the wind system has been developed and integrated with dSPACE hardware through a real-time interface for real-time validation and implementation of a cloud-based IoT platform. This IoT platform aids in acquiring real-time data on wind profiles, enhancing the assessment of SA. The representation of the wind energy system utilizes a multi-state wind speed model, incorporating the inherent randomness of wind energy generation. A time-series-based non-linear autoregressive with exogenous input (NARX) artificial neural network (ANN) model is employed to predict wind speed and WPG for SA-based operational reliability. This model proves instrumental in forecasting the dynamic nature of wind speed and WPG. The proposed approach for operational planning is validated by implementing the wind energy system in real-time on the dSPACE platform and integrating it with IoT. This approach considers the uncertainties associated with wind generation, providing a comprehensive framework for assessing SA and optimizing power system operations.

Are we focusing on the right parameters? Insights from Global Sensitivity Analysis of a Functional-Structural Plant Model

Abstract Performing global sensitivity analysis on functional-structural plant models (FSP models) can greatly benefit both model development and analysis by identifying the relevance of parameters for specific model outputs. Setting unimportant parameters to a fixed value decreases dimensionality of the typically large model parameter space. Efforts can then be concentrated on accurately estimating the most important input parameters. In this work, we apply the Elementary Effects method for dimensional models with arbitrary input types, adapting the method to models with inherent randomness. Our FSP model simulated a maize stand for 160 days of growth, considering three outputs, namely yield, peak biomass and peak leaf area index (LAI). Of 52 input parameters, 12 were identified as important for yield and peak biomass and 14 for LAI. Over 70 % of parameters were deemed unimportant for the outputs under consideration, including most parameters relating to crop architecture. Parameters governing shade avoidance response and leaf appearance rate (phyllochron) were also unimportant; variations in these physiological and developmental parameters do lead to visible changes in plant architecture but not to significant changes in yield, biomass or LAI. Some inputs identified as unimportant due to their low sensitivity index have a relatively high standard deviation of effects, with high fluctuations around a low mean, which could indicate non-linearity or interaction effects. Consequently, parameters with low sensitivity index but high standard deviation should be investigated further. Our study demonstrates that global sensitivity analysis can reveal which parameter values have the most influence on key outputs, predicting specific parameter estimates that need to be carefully characterized.

Microgrid energy management strategy based on LSTM-DDQN

Abstract This paper introduces a novel algorithm, LSTM-DDQN, designed to address the stochastic economic dispatch problem in microgrids. The inherent variability and randomness of photovoltaic (PV) energy pose challenges for accurate output prediction. Traditional approaches, including physical and statistical methods, often fall short in terms of accuracy. For microgrids with a high proportion of PV generation, we propose a new algorithm that combines Long Short-Term Memory (LSTM) neural networks with the Double Deep Q-Network (DDQN) algorithm. The algorithm leverages the LSTM model to capture the uncertainty of the learning environment and optimizes the Q-value iteration rules in the DDQN algorithm, thereby enhancing the training speed of the neural network. Experimental results demonstrate the algorithm’s remarkable dispatch capabilities in managing the interplay between PV generation, battery capacity, and overall load demand. This algorithm provides robust support for the stable operation of microgrids.

Progressive failure and quantitative stability analysis of rock slope with cross-type discrete fracture network using spring-based smoothed particle hydrodynamics method

The propagation and coalescence of discontinuous cross-type joints are often the primary drivers of rock slope failures. Due to the inherent randomness of these joints, identifying suitable methods to accurately characterize progressive failure and quantitatively assess the stability of slopes with cross-type joints is particularly challenging. To address this, a Discrete Fracture Network-Spring Based-Smoothed Particle Hydrodynamics (DFN-SB-SPH) method is proposed, which considers cross-type DFN generation, rock matrix heterogeneity, cracking, block contact, and deposition within a unified SPH framework. In this approach, damage and contact of blocks are achieved by introducing a fracture sign to enhance the kernel function and applying Newton's law to the damaged particles to realize the point to point contact. A uniaxial compression test on rock samples with double fissures is carried out to verify the accuracy of the proposed method. The influence of joint density, geometric parameters (i.e., trace length, dominate dip angle) of cross joints, and rock mass heterogeneity on the failure characteristics of RSCD are investigated. The results revealed that the DFN-SB-SPH model possesses robust capabilities for crack tracking, mass movement, and contact slipping along fracture surfaces. Moreover, DFN generation, heterogeneous characterization, shear and tensile fractures within a single SPH framework without mesh distortion are realized. Five typical failure modes calculated by DFN-SB-SPH are primarily determined by the combination of joint dip angles and trace lengths. Furthermore, the dip angle, trace length, joint density, and rock mass heterogeneity significantly alter the stability of RSCD with sensitivity coefficients for factor safety of 27.7 %, 44.6 %, 72.9 % and 35.0 %, respectively. Higher dominant dip angles (above 60°) lead to significantly longer run-out distances for RSCD compared to lower angles (below 30°) once instability occurs. The sudden rise in major principal stress at the slope's top and toe, along with the shift of tensile stress concentration towards the middle, predicts the stages of crack propagation and sliding in slope failure.