The Patent Landscape: Four Technical Clusters Driving Scrap Reduction

AI-driven process parameter optimization in injection molding is an active and rapidly expanding patent space, with more than 50 documents spanning jurisdictions including China, Japan, South Korea, the United States, Europe, and Brazil — of which approximately 25 are active or pending patents directly relevant to process optimization for scrap reduction. The dominant assignees with multiple filings include BASF SE, IMFLUX INC., Husky Injection Molding Systems, LS Mtron, and ITRI/Industrial Technology Research Institute. Chinese industrial assignees such as Kaos IoT Technology (Haier), Zhejiang University, and Beijing Baidu represent a growing cluster of data-driven and cloud-connected approaches.

The prevailing technical approaches cluster around four areas: (1) simulation-based closed-loop optimization, (2) reinforcement learning and iterative model training, (3) real-time melt and cavity pressure feedback control, and (4) multi-objective optimization using metaheuristic algorithms such as particle swarm optimization (PSO) and genetic algorithms. Each cluster addresses a distinct phase of the scrap-generation problem — from initial setup waste to in-cycle disturbance-driven defects to batch-change spikes.

Simulation-Coupled AI: How Digital Twins Eliminate Trial Scrap Before the First Shot

The most architecturally sophisticated approach to AI-driven scrap reduction involves coupling physics-based simulation with machine learning to generate predicted process parameters before any physical shot is made — eliminating the trial-and-error setup runs that account for a significant share of injection molding scrap. BASF SE’s 2023 patent defines a multi-step closed-loop in which an external processing unit first simulates the injection molding process using a simulation model combined with material-specific and machine parameters; an optimization algorithm then determines predicted process parameters that are fed directly to the physical machine.

The critical innovation in BASF SE’s architecture is the feedback loop: actual process data from the physical machine is used to recalibrate the simulation model itself, creating an adaptive digital twin that improves its predictions over time. The Chinese counterpart filing from BASF SE (2026) explicitly states that “process data can be used to optimize the modeling process using machine learning models” and that “cloud-based digital twins of materials and injection molding processes” enable batch-specific material information to be incorporated into the simulation — addressing the frequent source of scrap caused by material property variation between batches. According to WIPO, cloud-connected manufacturing systems of this type represent one of the fastest-growing patent categories in smart manufacturing globally.

A related simulation integration approach is described in INGLASS S.P.A.’s WO patent, where CAE simulation results are electronically processed to generate initial machine parameters, and actual machine execution results are stored in a database alongside simulation-derived parameters. This database allows subsequent CAE simulations to be corrected as a function of the divergence between simulated and real outcomes — progressively reducing the gap between virtual and physical process behavior that generates setup scrap. Space Solution Co.’s 2025 Korean patent extends the pressure-prediction dimension: an ANN-based (Artificial Neural Network) model is trained on CAE injection analysis outputs, extracting pressure distribution pattern vectors from elements and nodes of the analysis object to predict cavity pressure distributions under new conditions, enabling defect-prone pressure profiles to be identified and corrected before physical production begins.

“Cloud-based digital twins of materials and injection molding processes enable batch-specific material information to be incorporated into the simulation — directly addressing the scrap caused by material property variation between resin batches.”

Real-Time Melt and Cavity Pressure Control: Preventing Defects Cycle by Cycle

Real-time melt and cavity pressure feedback is the most mature scrap-reduction mechanism in the dataset, with multiple issued patents demonstrating cycle-by-cycle correction that prevents out-of-tolerance parts from being produced during process disturbances rather than detecting them after ejection. IMFLUX INC.’s 2021 patent establishes an optimal melt pressure curve from a baseline cycle and then adjusts injection pressure in subsequent cycles to force the monitored pressure to follow this reference trajectory. The filing explicitly describes the integration of machine learning algorithms to determine modifications to the optimal melt pressure curve, with training data drawn from multiple machines, molds, and melt materials to correlate part quality outcomes with pressure profiles.

The complementary autotuning architecture of IMFLUX INC.’s 2022 patent takes a model-based approach to determining initial control parameter values from a machine/mold/material model, then executes a run of injection cycles while measuring operational parameters. When any operational parameter exceeds a defined threshold, control parameters — including the melt pressure profile and PID gain values — are automatically adjusted for subsequent cycles. This threshold-triggered tuning prevents extended runs of defective parts while the process drifts, directly limiting cumulative scrap volumes without requiring operator intervention.

LS Mtron’s 2022 patent addresses the disturbance rejection problem: when a defective product is produced due to process disturbance, the system acquires current injection state data including viscosity profile and injection pressure values, inputs these into a trained “molding quality maintenance model,” and determines whether quality is being maintained. If the model indicates quality is at risk, the molding condition setting unit automatically generates corrected conditions — eliminating the sustained defect runs that characterize manual intervention-dependent production lines. This approach is particularly relevant given the findings documented by ISO on the impact of process variability in precision plastic part manufacturing.

ML Architectures for Parameter Recommendation and Multi-Objective Optimization

A distinct group of patents focuses on the specific ML model architectures and optimization algorithms that translate historical process data into actionable parameter recommendations — with scrap prevention embedded as an explicit objective function. ITRI’s 2020 patent couples an injection molding process simulator with a reinforcement-learning-style optimization loop: a molding condition optimizer proposes parameter sets, the simulator evaluates whether they produce acceptable outcomes, and parameters are iteratively refined over training rounds. The system explicitly tracks that, as training rounds increase, the number of parameter adjustments needed per round to achieve a simulated good part converges — meaning the model accumulates knowledge of the process space and navigates it more efficiently over time.

Steer Industrial’s 2024 patent trains an AI model on historical injection data that explicitly includes both yield parameters (good product rate) and adjustment records, setting the training iteration upper limit as a function of both metrics — ensuring the model is penalized for parameter choices that historically required multiple correction cycles. The system also determines the optimal timing for parameter re-optimization based on real-time machine data, avoiding wasted optimization cycles between legitimate re-optimization triggers. Kaos IoT Technology’s companion 2024 patent extends the multi-model pipeline further by incorporating a large language model (LLM) to parse natural-language user requests for parameter recommendations — an interface layer that lowers the barrier to AI-assisted process setup for non-expert operators on heterogeneous machine fleets.

The model-free approach patented by Zhejiang University in 2024 avoids surrogate model construction entirely. Instead, it calculates gradient directions iteratively using an iterative gradient estimation method and applies an adaptive moment estimation (Adam) algorithm to allocate adaptive step sizes to each parameter — enabling convergence to near-optimal parameters through online iteration without requiring extensive offline training data. This approach is particularly valuable during new mold qualification, where large historical datasets are not yet available, significantly reducing the number of physical test shots required. Research published by Nature on gradient-based optimization in manufacturing contexts has highlighted the practical advantage of avoiding offline model construction in low-data regimes. Beijing Baidu’s 2024 patent specifically trains a process parameter tuning model on injection pressure, injection volume, and injection speed per segment — parameters directly linked to short-shot and flash defects — as documented in manufacturing process standards referenced by SPE (Society of Plastics Engineers).

Key Players and the Shift Toward Multi-Objective Production Optimization

The injection molding AI optimization patent landscape is characterized by a clear stratification between established machinery OEMs, chemical materials companies, and a new wave of industrial IoT and cloud AI platform providers. Based on the frequency and technical depth of filings, five assignees represent the most active innovation centers in this space.

BASF SE — Closed-Loop Simulation Across Jurisdictions

BASF SE stands out for its closed-loop simulation architecture that integrates cloud-based digital twins with physical injection molding machines, with filings across JP, CN, BR, and KR jurisdictions. Their Brazilian filing extends the approach to a fully automated control system combining simulation, optimization, and real machine execution in a single programmatic framework — reflecting a strategy of protecting the architecture in every major manufacturing jurisdiction simultaneously.

IMFLUX INC. — Dominant in Real-Time Pressure-Profile Control

IMFLUX INC., a Procter & Gamble-affiliated company, is the dominant filer in real-time pressure-profile-based adaptive control, with multiple US, WO, and MX filings covering melt pressure tracking, cavity pressure control, PID autotuning, and compression rate optimization. Their 2022 WO patent on optimal compression rate process development demonstrates their systematic approach to characterizing the optimal compression rate window for a given material and then finding machine operating parameters that stay within that window — a methodology directly applicable to reducing flash and short-shot defects.

LS Mtron — AI Condition Generation and Disturbance Rejection

LS Mtron (Korea) has multiple JP-jurisdiction filings on AI-based molding condition generation, covering both rapid initial condition generation from mold specifications and real-time disturbance-triggered condition correction. This dual focus — on setup speed and on in-production stability — positions LS Mtron’s architecture as a comprehensive scrap reduction system across the full production lifecycle, from first article to sustained production runs.

Husky — Model Library Selection for Targeted Optimization

Husky Injection Molding Systems’ 2025 WO patent introduces a practical deployment architecture: the controller selects from a library of optimization models based on the operational parameter to be improved, enabling targeted optimization of whichever parameter is currently driving scrap. This operator-guided approach connects AI capability to specific production quality goals, making the system accessible to production engineers without requiring deep ML expertise. Husky’s approach reflects the broader industry shift documented by OECD toward human-in-the-loop AI systems in precision manufacturing.

Kaos IoT / Haier and Beijing Baidu — Cloud AI for Smart Factory Fleets

Kaos IoT/Haier and Beijing Baidu represent the Chinese industrial IoT and cloud AI trend, emphasizing multi-model pipelines, LLM-based parameter recommendation interfaces, and large-scale historical data utilization for injection parameter recommendation — approaches designed for deployment across heterogeneous fleets of injection machines in smart factory environments. Beijing Baidu’s 2024 patent specifically trains a process parameter tuning model on injection pressure, injection volume, and injection speed per segment, directly targeting the parameters most responsible for short-shot and flash defects at scale.

“The broader trend observable across the dataset is the migration from single-parameter closed-loop control toward multi-parameter, multi-objective AI frameworks that simultaneously optimize quality, cycle time, and energy consumption — a clear indicator that scrap rate minimization is now embedded as one objective within a broader production optimization agenda.”

The historical arc of the dataset makes this evolution concrete: the earliest relevant filing is a 1974 Pennwalt Corporation GB patent on viscosity-compensation control — a single-parameter, single-objective system. Fifty years later, the state of the art involves cloud-connected AI frameworks that simultaneously optimize dimensional tolerances, cycle time, energy consumption, and material utilization across fleets of machines, with LLM interfaces enabling non-expert operators to specify optimization goals in plain language. The trajectory from that 1974 filing to the 2024–2025 patents reviewed here maps precisely onto the broader digitalization of manufacturing intelligence tracked by organizations such as WIPO in their annual technology trend reports.

For R&D teams and process engineers seeking to understand where the competitive frontier lies, the PatSnap R&D Intelligence platform provides full access to the patent families, citation networks, and assignee landscapes described in this analysis. The PatSnap IP Intelligence suite enables freedom-to-operate analysis across the jurisdictions — US, CN, JP, KR, EP, BR — where the most active injection molding AI patents are concentrated.