China Nonferrous Metallurgy

Abstract:This study proposes a novel high-temperature sintering-hydrometallurgical leaching process for the separation and recovery of aluminum (Al) and palladium (Pd) from spent automotive catalysts. In this process, sodium carbonate (Na₂CO₃) and calcium carbonate (CaCO₃) are mixed with the raw material for sintering, and the crystalline phase formation mechanism during the sintering process is systematically investigated. Na₂CO₃ preferentially reacts with cordierite to form sodium aluminosilicates and other products; Na⁺ ions can inhibit the phase transformation of β-2CaO·SiO₂ to the γ phase and enhance its stability, while excessive Ca²⁺ ions tend to form refractory 3CaO·Al₂O₃ with low leachability. The optimal roasting parameters are determined as N/A 1.0-1.35 and C/S 2.0 ~ 2.16, and the main crystalline phases of the sintered clinker are γ-2CaO·SiO₂, β-2CaO·SiO₂, CaO·ZrO₂, Na₂O·Al₂O₃, and MgO. Alkali leaching at 80℃ with 10 g/L caustic lye for 30 minutes results in an Al leaching efficiency of over 95%. Subsequent acid leaching with 1 mol/L hydrochloric acid (HCl) at 50 ~ 70℃ for 40 minutes can effectively remove magnesium (Mg) with negligible Pd loss; the mass fraction of PdO in the acid leaching residue is upgraded from 0.043 4% to 0.314%, achieving nearly tenfold Pd enrichment. The proposed process not only enables efficient recovery of Al from spent catalysts but also effectively reduces the Mg content, thereby lowering the smelting temperature required for the subsequent pyrometallurgical recovery of Pd, which provides a new technical approach for the resource recycling and valorization of spent automotive catalysts.

Abstract:Spent automotive catalysts are an important secondary resource of platinum group metals (PGMs). Conventional iron-collection pyrometallurgical processes require the addition of extra iron collectors and large amounts of slag-forming fluxes, resulting in high processing costs. To reduce costs and promote the synergistic utilization of solid wastes, this study proposes a co-smelting process using iron tailings rich in SiO₂, CaO, and Fe₂O₃ together with spent automotive catalysts. In this process, iron tailings serve simultaneously as both a collector for pyrometallurgical recovery of PGMs and a fluxing agent. Thermodynamic analyses and single-factor experiments were conducted to investigate the effects of key process parameters on the recovery of PGMs, followed by validation experiments and characterization analyses. Thermodynamic calculations indicated that at 1 500℃, platinum group metal oxides could be reduced to their metallic states by carbon, iron, and iron oxides, with Gibbs free energy changes (ΔG) below zero. When the temperature exceeded 1 550℃, SiO₂ was readily reduced to elemental silicon, which subsequently reacted with iron to form Fe-Si alloys, thereby hindering the subsequent separation process. Furthermore, at 1 500℃, the SiO₂-CaO-Al₂O₃-MgO (5%) quaternary slag system exhibited a broad liquid-phase region, ensuring complete melting, favorable fluidity, and good stability. The experimental results showed that under the optimal conditions of a smelting temperature of 1 500℃, binary basicity (CaO/SiO₂) of 0.7, iron tailings addition of 30%, and reductant addition of 1.2%, the recovery rates of Pt, Pd, and Rh reached 98.45%, 98.23%, and 96.7%, respectively. XRD and SEM-EDS analyses revealed that PGMs were highly and uniformly dispersed within the Fe-based alloy phase, forming stable solid solutions with metallic iron and achieving efficient enrichment of PGMs. The final slag exhibited a glassy structure, enabling the stable immobilization of hazardous components. This technology provides an efficient, low-cost, and environmentally friendly solution for the synergistic resource utilization and harmless treatment of spent automotive catalysts and iron tailings. It also demonstrates significant industrial application potential for promoting the recycling of secondary platinum group metal resources and the sustainable management of metallurgical solid wastes.

Abstract:The iron collector product of spent three-way catalysts (PGMs-iron alloy) is a typical intermediate product from precious metal smelting tailings and scrapped catalyst treatment. The large amount of iron present not only severely interferes with the purification accuracy of precious metals but also reduces the processing efficiency of subsequent refining processes while increasing energy consumption and costs. This study takes PGMs-iron alloy as the research object, focusing on the efficiency and controllability of electrolytic separation technology. Anodic dissolution rate (reflecting the stability and sustainability of alloy dissolution) and cathodic current efficiency (characterizing the selectivity of target ion reduction and energy utilization efficiency) were selected as key evaluation indicators. The synergistic effects of process parameters such as electrolysis temperature, electrolyte pH, Fe²⁺ concentration, and current density, as well as the composite electrolyte system composed of sodium citrate (15 g/L), ascorbic acid (2 g/L) and ammonium sulfate (10 g/L) on separation efficiency, were systematically investigated. Key operating conditions were optimized through single-factor variable experiments and mechanism analysis. Experimental results show that under the optimal conditions of temperature 50℃ , pH=5.0, Fe²⁺ concentration 120 g/L, current density 350 A/m², the enrichment factor of platinum, palladium and rhodium in anode slime is 20 times, and the recovery rate exceeds 99 %; The peak position of cathode iron is consistent with the standard card ( PDF 01-087-0721 ), and no other single phase is detected; The waste electrolyte produced in the process can be used to prepare ferrous sulfate by dissolution crystallization; Under the optimum operating conditions, there is no obvious passivation phenomenon in the anode, and there is no obvious side reaction of hydrogen evolution in the cathode. The purity and current efficiency of the product are at a high level. This study not only provides a solid theoretical basis and a direct reference technical scheme for the industrial efficient separation of PGMs-iron alloy, but also provides a new idea for the utilization of PGMs containing waste.

Abstract:Anode-free alkali metal batteries (AFAMBs) are regarded as a promising next-generation energy storage technology, featuring high energy density, low manufacturing cost and simple structure. However, the anode side is confronted with challenges such as unstable solid electrolyte interface (SEI), dendrite growth, irreversible deposition and severe volume changes, caused by the high reactivity of alkali metals. These issues severely limit the cycle life and commercialization of AFAMBs. In response to these issues, non-ferrous metals play key roles in the construction of the functionalized interfaces on the anode side of AFAMBs due to their excellent electrical conductivity, favorable mechanical properties and the ability to induce alkali metals. This article focuses on various strategies and mechanisms: (i) constructing alkaliphlic interface layers (e. g. , Au, In) to reduce the nucleation barrier and guide uniform deposition; (ii) building in-situ/artificial SEI with Ru, Zn, etc. , to enhance interface stability and suppress side reactions and dendrite growth; (iii) implementing three-dimensional porous scaffolds to alleviate volume strain and optimize ion/electron transport; (iv) developing advanced alloys and composite current collectors to regulate interface deposition behavior and improve cycle reversibility. This article further analyzes the research progress in the mechanism of non-ferrous metals in aspects such as lattice matching and electronic structure modulation. By introducing high affinity non-ferrous species or reconfiguring crystal planes, the energy barrier can be reduced and homogeneous nucleation can be guided. Certain non-ferrous metals with specific atomic radii facilitate in-situ alloying during deposition, forming stable alloy phases, ensuring reversible deposition. Moreover, by integrating non-ferrous nanoclusters, single-atom coordination bonds, or heterostructures with specific active sites, the local current density can be regulated, thereby accelerating ion transport kinetics, homogenizing the electric field and inhibiting dendrite growth. Future research should prioritize the development of cost-efficient modification materials; develop green, simple, and compatible functionalization techniques with existing production lines; and design advanced computational models to predict interface evolution, ultimately providing support for high-performance AFAMBs.

Abstract:China generates over 10 million tonnes of electroplating sludge annually, which contains high concentrations of heavy metals such as Cr, Cu, and Ni. Conventional hydrometallurgical techniques (e. g. , acid, alkaline, and ammonia leaching) face challenges including secondary pollution and equipment corrosion. Amino acids exhibit high affinity for chalcophile metals (e. g. , Cu, Ag, Au) and form stable complexes, making them a research focus in hydrometallurgy. This study proposes a copper leaching process from electroplating sludge using amino acid complexation. The effects of different amino acids on copper leaching efficiency were systematically investigated. Lysine was identified as the most effective leaching agent, and optimal leaching conditions were determined through single-factor experiments and comprehensive tests. The resulting products were characterized, leading to the following key conclusions: Among arginine, citrulline, glycine, glutamine, theanine, and lysine, lysine achieved near-complete copper leaching (>99%) within 6 h at a relatively low concentration of 0.2 mol/L, demonstrating strong industrial application potential; other amino acids did not meet production requirements. The optimal parameters for copper leaching with lysine were: lysine concentration 0.75 mol/L, liquid-solid ratio 50:1, temperature 328 K, and leaching time 5 h, under which the copper leaching rate reached 93%. Meanwhile, other valuable metal components were scarcely leached, indicating that lysine selectively extracts copper from electroplating sludge. The mechanism involves complexation of Cu2?with -COO? and -NH? groups in lysine, with -NH? showing stronger coordination ability than -COO?. Kinetic analysis revealed that the leaching process follows a mixed diffusion-controlled mechanism, with an apparent activation energy of 36.5 kJ/mol. Compared with existing glycine-based systems, lysine offers significantly faster leaching rates and superior selectivity for copper. These findings provide new insights into the eco-friendly disposal of heavy metal contaminants and green recovery of valuable metals.

Abstract:Given the difficulty in completely removing cyanide from cyanide tailings and the associated environmental risks during high-temperature smelting, this study investigates cyanide tailings derived from a gold concentrate in Shandong Province and systematically examines cyanide removal behavior as well as the migration and transformation of sulfide minerals under oxidative roasting conditions. By systematically regulating roasting temperature, residence time, oxygen concentration, and gas flow rate, the optimal operating conditions for cyanide removal were identified, and the sulfur fixation behavior and its influence on the cyanide removal process were analyzed. The results indicate that at a roasting temperature of 500°C for 30 min under an air atmosphere with a gas flow rate of 400 mL·min^-1, the total cyanide removal efficiency reached 97. 36%, and the cyanide concentration in the leachate of the roasted product decreased to 0. 187 mg·L^-1, satisfying national discharge standards. The study demonstrates that CaO and MgO generated from the decomposition of carbonate minerals, together with iron oxides formed by sulfide oxidation, can capture SO₂ in situ during roasting, thereby partially fixing sulfur in the form of sulfates, with the sulfur fixation efficiency remaining stably above 60%. Meanwhile, SO₂ released from sulfide oxidation can synergistically enhance the oxidation and decomposition of cyanide in the presence of O₂ under appropriate conditions. Combined XPS characterization and thermodynamic analysis elucidated the mechanism underlying the breakdown of Fe-CN complexes and the complete elimination of cyanide-containing species. Overall, the results confirm that oxidative roasting enables efficient cyanide removal from tailings while significantly reducing sulfur-bearing gaseous emissions, thereby providing both theoretical insight and technical support for the safe utilization and pollution control of cyanide tailings in high-temperature co-smelting systems.

Abstract:Transitional laterite nickel ore is a sluggish and difficult-to-utilize resource. The existing acid leaching process tends to introduce impurity elements such as iron, magnesium, and aluminum into the leachate, making effective separation difficult. The reduction roasting-ammoniacal leaching process tends to generate Fe(OH)₃ colloid, resulting in low recovery rates of nickel and cobalt and a lengthy process. This study addresses this issue by focusing on a certain mine's transitional laterite nickel ore as the research subject. It systematically investigates its process mineralogy characteristics, proposes a process for recovering nickel and iron through carbothermic reduction-magnetic separation, examines the impact of temperature on nickel and iron recovery, and analyzes the structural characteristics of the products. The main conclusions are as follows. The chemical composition and mineral phase analysis of the ore reveal that the nickel and iron grades in the sample are 1.23% and 28.14%, respectively, with SiO₂/MgO ratio of 1.53, indicating a low-iron, high-magnesium transitional laterite nickel ore. The microscopic analysis of minerals indicates that nickel and iron elements occur in multiple carriers, with nickel magnetite and goethite being the main occurrence phases; nickel is distributed in olivine, limonite, and hypersthene in an isomorphism manner. By adding 25% coal powder and 5% CaO to the raw material and reducing it through roasting at 1 340°C for 2 hours, the phase composition of the product consists of Fe-Ni, Mg₂SiO₄, and Al₂SiO₅. The grades of Ni and Fe in the magnetic separation product reach 4.04% and 83.67%, respectively, with corresponding recovery rates of 92.26% and 79.87%. After magnetic separation, the Fe-Ni alloy is enriched in the magnetic separation product, which exhibits saturation magnetization with a saturation magnetization intensity of 145.51 A·m²/kg, allowing for efficient enrichment through low-intensity magnetic separation. Impurities such as Si and Mg are mainly discharged with the magnetic separation tailings in the form of Mg₂SiO₄.

Abstract:The similar physicochemical properties of indium and cadmium often result in an excessively high cadmium content in crude indium produced by current smelting processes, making it difficult to meet market purity requirements. The glycerol-potassium iodide method exploits the difference in affinity between indium and impurity elements toward iodine, selectively converting cadmium into iodides, thereby enabling targeted cadmium removal. In this study, crude indium was refined using the glycerol-potassium iodide process to deeply eliminate impurities. Single-factor experiments were conducted to determine the optimal refining conditions, followed by validation tests. The main conclusions are as follows. The optimal conditions obtained are a refining temperature of 190℃, a refining time of 10 min, a liquid-to-solid ratio of 8:1, and an I₂-KI excess coefficient of 3. Under these conditions, the cadmium content in the refined indium drops to 250 ppm, the indium purity increases to 99.93%, and the direct yield reaches 98.60%. The results demonstrate that this process efficiently removes cadmium from crude indium while also synergistically eliminating impurities such as arsenic, silicon, and thallium. Consequently, indium purity is significantly enhanced without sacrificing a high direct yield. These findings provide technical support for efficient preliminary refining of crude indium.

Abstract:Taking the northern by-product scandium sandstone uranium ore as the research object, this study addresses the problem of low uranium leaching rates caused by pore blockage in high-acid leaching systems. An innovative method is proposed using the organic chelating agent nitrilotriacetic acid (NTA) under low acid concentration to enhance the co-leaching of uranium and scandium, aiming to improve the co-leaching of uranium and scandium. Research on the NTA-enhanced co-leaching of uranium and scandium was conducted using northern by-product scandium sandstone uranium ore as the subject. The effects of single factors on leaching were examined, including sulfuric acid concentration, nitrilotriacetic acid concentration, fluorinated surfactant concentration, and leaching time. Based on this, a four-factor two-level orthogonal design experiment was carried out. Through linear regression analysis and standardized effect Pareto chart analysis, the factors significantly affecting the uranium and scandium leaching were identified, and a linear regression equation model correlating leaching with each factor was constructed. The results indicate that, taking P<0. 000 1 as the criterion, sulfuric acid concentration has a significantly greater effect on uranium leaching than leaching time and NTA concentration; the important factors affecting scandium leaching are sulfuric acid concentration and leaching time. Based on the identified significant factors (P<0. 000 1) and significant quadratic factors (0. 01

Abstract:This paper employs the fault tree analysis (FTA) method, taking the aluminum leakage incident from electrolytic cells as the top event, to construct a qualitative analysis model comprising 12 basic events. Analysis of minimal cut sets and minimal path sets identifies 11 distinct paths leading to aluminum leakage. Specifically, bottom leakage can be triggered by any single factor among X₁-X₉, whereas side leakage requires the combined effect of (X₁₀∨X₁₁) and X₁₂. The structural importance of events was calculated using the minimal cut set approximation method. The results indicate that the factors related to the cell bottom, such as poor lining material quality and poor furnace construction quality (I_φ=1), together with the burning through of the side steel shell (I_φ=1), have the greatest impact on the accident, followed by side corrosion and excessively high bath temperature (I_φ=0.5). Quantitative evaluation shows that the annual probability of the top event is approximately 4.40%, which closely matches actual statistical data, validating the model's effectiveness. Sensitivity analysis indicates that prioritizing control of high-probability, high-importance events (e. g. , X₁, X₂) can significantly reduce overall risk.

Abstract:Low-radioactivity alumina is highly demanded in high-end applications. However, the emerging synthetic route using 4 N high-purity aluminum (with extremely low radionuclide content) as feedstock leads to a remarkable increase in production cost. In this work, low-α alumina was prepared from ordinary industrial aluminum hydroxide via a combined process: hydrothermal synthesis of boehmite, acid leaching for removal of radioactive elements, and high-temperature calcination. The process was optimized, and the products were systematically characterized. The maximum specific surface area and pore volume of boehmite were achieved at a hydrothermal temperature of 200℃ for 2 h. Grinding pretreatment destroyed the agglomerate structure of aluminum hydroxide, exposing the encapsulated radioactive elements, which facilitated the efficient removal of Th, U and other impurities in the subsequent hydrothermal process. Acid neutralization converted the entrapped U and Th in the boehmite precursor into soluble ions. After acid leaching with 10% HCl for 2 h and calcination at 1 250℃ for 2 h, the contents of Th and U in the product were both below 10 ppb(10^-9), the Na₂O content was only 0.006%, and the α-phase conversion ratio exceeded 90%. Using industrial aluminum hydroxide as raw material, high-performance low- α alumina satisfying high-end application requirements was obtained via the grinding-hydrothermal-acid washing-calcination process. This strategy breaks the dependence on high-purity feedstocks for low-α alumina production, and significantly reduces the manufacturing cost.

Abstract:The α-Fe?O? powder prepared by traditional methods exhibits poor dispersibility, irregular morphology, and imposes high demands on equipment and operational requirements. The precursor of iron oxide was prepared by coprecipitation method using FeCl? and NaOH as raw materials, and spherical α-Fe?O? powder with a particle size of 1 ~ 4 μm was prepared by spray drying and further calcination. The samples were characterized by TG, XRD, SEM and Nitrogen adsorption-desorption. The effects of calcination temperatures, types of precipitants and valence state of iron ions in the raw material solution on the phase, morphology and specific surface area of the products were analyzed in systematically. The results showed that the precursor of iron oxide was amorphous after spray drying, and the total weight loss of the precursor after further calcination was about 15.59%, accompanied by the dehydration of the precursor of iron oxide precursor to produce iron oxide. As the calcination temperature increased, the intensity of the diffraction peak of iron oxide increased, the specific surface area decreased, and the surface of the spherical product changed from many small particles to rod-shaped and then to sheet-like particles, indicating sintering between particles. The reddish brown product changed from light to dark. When the precipitant was ammonium bicarbonate or ammonia water, the product was a mixture of α-Fe?O? and γ-Fe?O?. When the valence state of iron ions in the raw material solution changed from +3 to +2, the prepared product transformed from a single α-Fe?O? phase to a mixture of α-Fe?O? and γ-Fe?O?.

Abstract:To solve the problems of poor structural stability and cycling performance degradation faced by traditional nickel-based cathode materials in high-energy-density applications, this research has been conducted on improving their comprehensive electrochemical performance through rare earth element composite modification. Rare earth composite nickel-based thin film materials were prepared using high oxygen pressure assisted melting reaction method, and Pt electrodes were constructed using magnetron sputtering technology. The material structure and properties were characterized by X-ray diffraction, electrochemical workstation, and other methods. System regulation of rare earth species (Nd, Sm, Eu), substrate orientation, and hydrogenation process parameters, with a focus on studying the phase structure, resistance characteristics, and electrochemical response laws. The results showed that after hydrogenation treatment, the basic resistance change ratio of NdNiO₃/LaAlO₃ films with (110) orientation at room temperature reached 15, and the resistance significantly decreased in the low temperature range; The Pt electrode forms a low contact resistivity (2.081 × 10^-5 Ω·cm²) with NdNiO₃/LaAlO₃, effectively ensuring the accuracy of electrical performance testing. Research has confirmed that rare earth elements can significantly enhance the structural stability and electrochemical cycling stability of nickel-based positive electrode materials by regulating the integrity of the material's crystal lattice, optimizing lithium ion transmission channels, and improving the ion diffusion kinetics performance. The research results have clarified the logical process of process optimization and the mechanism of performance improvement for rare earth-modified nickel-based positive electrode materials, providing important experimental support and technical references for the research and industrial application of high-performance and long-life lithium battery positive electrode materials.

Abstract:Aluminium slag is a typical industrial solid waste generated in the production process of electrolytic aluminium and recycled aluminium. In this study, by regulating the phase composition of aluminium slag, magnesium aluminate spinel (MA) materials were prepared using the solid-state sintering method, and the effects of typical impurities in aluminium slag on the bulk density of the sintered body were systematically investigated. The results show that the four impurities SiO₂, CaO, Fe₂O₃ and TiO₂ can all increase the bulk density of aluminium slag-based MA sintered materials. Under the process conditions of sintering at 1 400 °C for 3 h, when 5%SiO₂, 5%CaO, 4.5%Fe₂O₃ and 6%TiO₂ were respectively added to the aluminium slag to sinter MA materials, the bulk density of the materials changed from 2.02 g·cm^-3 to 2.03, 2.05, 2.21 and 2.04 g·cm^-3, respectively. Among them, Fe₂O₃ promotes densification better than TiO₂ and CaO, while the densification effect of SiO₂ is relatively weak. The mechanisms by which different impurities enhance the bulk density of MA materials differ: SiO₂ and CaO can form low-melting-point liquid phases during sintering, strengthening the mass transfer process through liquid-phase sintering effects and thereby promoting material densification; Fe₂O₃ reacts with components in the system to form a (Mg, Fe)₂SiO₄ phase that is eutectic with the MA phase, causing the microstructure to develop multi-closed pore features, thus increasing bulk density; TiO₂ forms fine Mg_xAl_yTi_zO_(2x+3y+4z)/2 particles during sintering, which accumulate at the grain boundaries of MA coarse grains, providing grain boundary bonding to achieve material densification. In addition, the diffusion behaviour of impurity ions during sintering is also an important reason for the change in bulk density of MA materials.

Abstract:The zinc content in a certain type of chemical fiber sludge reaches 5.98%, indicating a high potential for comprehensive recovery. In this paper, a sulfuric acid leaching method was employed to recover zinc from the sludge. The effects of key factors, including sulfuric acid concentration, leaching time, temperature, and liquid-to-solid ratio, on the zinc leaching efficiency were systematically investigated. Based on the single-factor experimental data, a PSO-BP model was constructed to globally optimize the process parameters, and the optimization results were subsequently validated. The main conclusions are as follows. The optimal process parameters determined by single-factor experiments were: sulfuric acid concentration of 2 mol·L^-1, leaching time of 60 min, temperature of 30°C, and liquid-to-solid ratio of 4:1 mL·g^-1, under which the average zinc leaching efficiency was 97.91%. The optimal parameters obtained from the PSO-BP model were: sulfuric acid concentration of 2.16 mol·L^-1, leaching time of 64.1 min, temperature of 33.58°C, and liquid-to-solid ratio of 4.18:1 mL·g^-1. Under these conditions, the model-predicted zinc leaching efficiency was 98.96%, and the average experimental zinc leaching efficiency was 98.91%, with a relative error of 0.05%, fully validating the accuracy of the optimization model and the reliability of the obtained process parameters. Compared with the single-factor results, the optimal zinc leaching efficiency after model optimization was improved by one percentage point, yielding significant economic and environmental benefits. This study not only provides process parameters and decision support for the recovery of zinc from chemical fiber sludge but also offers a new approach for the intelligent and precise development of solid waste resource utilization processes.

Abstract:Viscosity is a crucial physicochemical parameter that regulates mass and heat transfer kinetics at the slag-metal interface in pyrometallurgical reactors. In this study, a quaternary CaO-MgO-Al₂O₃-SiO₂ reference slag system was examined within the framework of the Ion-Molecule Coexistence Theory. Using MATLAB based algorithms, the equilibrium activities of all likely structural entities were carefully calculated; then, an Arrhenius-type regression was applied to develop a semi-empirical viscosity model. Experimental validation over different compositional ranges showed that the predicted viscosities closely matched the measured data, with deviations in activation energy for viscous flow limited to 0.08, 0.61, 0.10, and 0.28 kJ·mol^-1, respectively. Spectroscopic and phase-equilibrium analyses confirm that the melt consists of simple ionic species (Ca²⁺, Mg²⁺, O^2-) coexisting with [Si_xO_y]^z- and [AlO_m]^n- polymeric units. Increasing concentrations of Ca²⁺ and Mg²⁺ ions have a network-modifying effect, depolymerizing the aluminosilicate structure and causing a structural shift from molecular to ionic dominance. During controlled cooling, the primary solid phases identified were 2CaO·SiO₂, MgO·Al₂O₃, CaO·Al₂O₃·SiO₂, and CaO·MgO·SiO₂. This model provides the metallurgical community with a fast, composition-dependent viscosity predictor for CaO-MgO-Al₂O₃-SiO₂ slags and explains the direct link between melt speciation and overall transport properties.

Abstract:To address the issues of insufficient prediction accuracy and slow convergence in sinter mixture moisture content forecasting, this study proposes a hybrid prediction model based on an Improved Northern Goshawk Optimization algorithm (UNGO) integrated with a CNN-BiLSTM-Self Attention architecture. The model first employs a Convolutional Neural Network (CNN) to extract spatial correlation features from the input data, then utilizes a Bidirectional Long Short-Term Memory network (BiLSTM) to capture the temporal dependencies of the sinter mixture moisture content. Meanwhile, a Self-Attention mechanism is incorporated to strengthen key feature weighting and long-range dependency modeling, thereby improving prediction stability and accuracy. To reduce model complexity and enhance computational efficiency, the Spearman correlation coefficient is applied to screen input features and remove variables with low relevance. To overcome the tendency of traditional optimization algorithms to fall into local optima, the proposed UNGO introduces chaotic initialization and a nonlinear sine-cosine decay strategy, significantly enhancing both global exploration and local exploitation capabilities. This enables adaptive optimization of key hyperparameters, including the learning rate, number of BiLSTM neurons, regularization coefficient, and attention key dimension. Experimental results based on production data from a large steel plant in Hebei Province demonstrate that, under identical data conditions, the CNN-BiLSTM-Self Attention model achieves MSE = 0. 057 5, RMSE = 0. 239 8, MAE = 0. 192 3, and R² = 0. 790 8, outperforming CNN, LSTM, CNN-LSTM, and CNN-BiLSTM models. This confirms the model's superior capability in feature representation and temporal dependency learning. After optimization with the UNGO algorithm, the performance metrics are further improved to MSE = 0. 022 4, RMSE = 0. 149 8, MAE = 0. 118 5, and R² = 0. 918 4, showing faster convergence and higher prediction accuracy than models optimized by NGO, PSO, and WOA. Furthermore, compared with other mainstream prediction models such as UNGO-Transformer, UNGO-BP, and UNGO-XGBoost, the proposed model exhibits superior prediction accuracy, generalization, and robustness. In summary, the UNGO-CNN-BiLSTM-Self Attention model demonstrates excellent overall performance in both structural design and parameter optimization, enabling high-precision prediction of sinter mixture moisture content. It provides a reliable technical foundation for intelligent control and energy efficiency optimization in the sintering process.