When using HyperX with a FEM, element forces generated during the finite element analysis are post-processed into sets of Design-To Loads. This process of converting element forces to Design-To (i.e. stiffened panel) Loads is sometimes referred to as Load Extraction or "pulling the loads".
The challenge of pulling the loads is determining how to aggregate loads from multiple finite elements into a single set of Design-To Loads. For example, a typical smeared panel Zone is modeled using several finite elements. If the \(N_x\) element force varies among the elements, designing to the maximum element load for panel strength may be appropriate for material strength but could be too conservative for panel buckling since the peak element load may be localized. For this reason, strength loads are computed separately from buckling loads.
There are several methods available to pull FEA loads, which fall into three broad categories: Statistical, Element Based, and Filtered. The FEA load method is defined via Load Properties, found on the Property Tree, which can be assigned to various selections of Zones.
All the Load Property forms have dropdowns for Shell/Beam/Connecter load components. By default, all components are considered. If any of those load components are deactivated, then they are zeroed out in the final set of Design-To Loads.
As of HyperX version 2026.1.13, the Filtered Load Property features two methods of identifying potentially critical load cases: Peak Load and new, Solver-specific, Convex Hull. Each method results in its own list of critical load-case-and-element combinations. They can be used together or individually. If used together, the two critical case lists – processed separately - are combined into one master critical load case list for Analysis and Sizing.
Note
By default, all load components are considered by both processing techniques. If any of those load components for shell, beam, or connector are deactivated using the check-boxes, they are zeroed-out in the final list of Design-to Loads.
Peak Loads
The Peak Load (aka Peak-Element) method determines the critical element and load case for a series of metrics. It is based on peak loads in the elements but is far more efficient than the element-based approach.
This approach uses a set of metrics to identify the highest loads in the component. For each metric, there is a criterion that defines how an element-load case pair is selected for the metric. For example, the element and Design Load Case with the highest \(+N_x\) load is a strength Design-to Load - all other Design Load Cases are ignored for this metric.
There are metrics available for strength loads, combined loads, and buckling. A metric is only associated with a single Design Load Case, so the maximum number of Design-To Loads is always constant. There are 32 metrics/cases for Zones and there are 22 metrics/cases for Joints (as of version 2026.1.13 more can be added with the "Include Comprehensive Metrics" option for a total of 141 metrics). The comparatively low number of cases evaluated makes this the fastest method computationally.
There are three options available for users to control the speed and accuracy of the method:
-
Filter Design Loads - This is the recommended method. It uses a filtered list of critical Design Load Cases resulting in the fastest Analysis and Sizing.
-
Run All Design Loads - Still uses a subset of critical elements per Zone but uses ALL Design Load Cases resulting in longer runtimes. This option is only applicable to Zones.
-
Include Comprehensive Metrics - As of HyperX version 2026.1.13, you can enable this option to include additional load metrics when Sizing/Analyzing (up to 141 total metrics). While the above options are mutually exclusive, this option is compatible with either one. For more information on these metrics see:
The Statistical method averages the element forces and adds \(N\sigma\) number of standard deviations to increase the level of conservatism. This method captures the general load trend while accounting for outlier loads.
Note
Setting N=0 uses a direct average which is the same as using the Average method in Tension/Compression mode.
Each of the load components (\(N_{ij}\)) are calculated independently so the Design-to Loads are not representative of any one element. Tensile and compressive portions of the load are averaged separately. A set of tension and compression Design-to Loads is generated for each Design Load Case.
Panel Zones are processed on an element area-weighted average while beams are length-weighted. Edge Joints are weighted based on element edge length and Point Joints are processed using an unweighted average.
Example of a 2-Sigma Statistical Load Property
The Average method is a special case of the Statistical method when N=0. Element forces and moments are directly averaged based on the element areas.
For Panel Zones, the area-weighted average is calculated for the loads in the shell elements. For beams, the length-weighted average is computed. For Edge Joints, the element-edge-length weighted average is computed. For Point Joints, each load component in the connector finite elements is averaged with no weighting.
There are two modes:
-
Tension/Compression - Two averages are performed for each load component, one for elements with tensile load, and one for elements with compressive load. The two resulting load scenarios are considered during analysis.
-
For plate shear load components (\(N_{xy}\)/\(M_{xy}\)/\(Q_x\)/\(Q_y\)), HyperX takes the maximum value between the averaged tension and compression loads – e.g. if the average tension \(N_{xy}\) is 50 and the average compression \(N_{xy}\) is -100, the “tension” loads will be assigned an \(N_{xy}\) of -100.
-
If a component of load has either no tensile or no compressive loads, HyperX substitutes in the corresponding load from the opposite loading direction – e.g. if there are no elements with a positive \(N_x\) and the average compressive \(N_x\) is calculated to be -150, the “tension” loads will be assigned an \(N_x\) of -150.
-
-
Average - True average of each load component in all elements, regardless of whether they are tension or compression.
Note
Can result in some “cancelling out” of load if there is a reversal, so the Tension/Compression option is more conservative.
The Neighbor Average method is a special type of averaging scheme used explicitly on Edge Joint members. Essentially, a local 'neighborhood' is defined around each Joint member. The load components are then averaged among the elements in that neighborhood. The size of the neighborhood on each side of the Joint is defined by the user.
The Neighbor Average method is an "edge length" weighted average of the elements' loads where the "edge length" comes from the edge of the shell element running along the joint.
The Element-Based method performs strength analysis on each individual element instead of averaging. Since the Design-to Loads are generated per element, per Design Load Case, this method is the most comprehensive but also the most computationally expensive. It is primarily used for final margins of safety as opposed to Sizing.
Note
Each element is treated as a strength Design-to Load. Buckling analysis performed using average compressive loads.
The Element Based method is considered to be the most comprehensive load extraction method because all outliers of element loads in the Zone are considered in the analysis.
In the Load Property training, some of these methods are compared by Sizing the same Structure. The results are summarized below.
Peak-Element load processing leads to the same results as element-based sizing, in ~1/10th of the time.
Statistical processing metrics are quick, and lead to lighter solutions, but negative margins can still persist at the element-level.
Note
The time difference between statistical methods and peak element increases as the number of load cases increases