KIDD_Executive_Summary.pdf

For Official Use Only

(U) KIDD Version 7.0.0 Executive Summary

29 May 2014

DISTRIBUTION STATEMENT C: Distribution authorized to U.S. Government agencies and their contractors for administrative and operational use, 2 November 2011. Refer all other requests for this document to: MDA/DESL, 5222 Martin Road, Redstone Arsenal, AL 35898.

WARNING: This document contains technical data whose export is restricted by the Arms Export Control Act (Title 22, U.S.C., Sec 2751 et seq) or Export Administration Act of 1979, as amended, Title 50, U.S.C., app 2401 et seq.

Violation of these export laws are subject to severe criminal penalties. Disseminate in accordance with provisions of DoD Directive 5230.25.

DESTRUCTION NOTICE: Destroy by any method that will prevent disclosure of contents or reconstruction of document.

This document is unclassified.

For Official Use Only KIDD Version 7.0

For Official Use Only ii

(U) TABLE OF CONTENTS

Page SECTION 1. (U) INTRODUCTION ....................................................................... 1-1

SECTION 2. (U) KIDD OVERVIEW ..................................................................... 2-1

2.1 (U) General Information .................................................................... 2-1

2.2 (U) Data Sources ................................................................................ 2-4

2.2.1 (U) Gun Tests ......................................................................... 2-4

2.2.2 (U) Sled Tests ........................................................................ 2-5

2.2.3 (U) Flight Tests ...................................................................... 2-5

2.3 (U) Threat Types Supported .............................................................. 2-5

2.4 (U) Applications ................................................................................. 2-6

SECTION 3. (U) HISTORICAL OVERVIEW ....................................................... 3-1

3.1 (U) Prior Versions .............................................................................. 3-1

3.1.1 (U) Version 1 ......................................................................... 3-1

3.1.2 (U) Version 2 ......................................................................... 3-1

3.1.3 (U) Version 3 ......................................................................... 3-1

3.1.4 (U) Version 4 ......................................................................... 3-3

3.1.5 (U) Version 5 ......................................................................... 3-5

3.1.6 (U) Version 6 ......................................................................... 3-7

3.2 (U) Current Version ......................................................................... 3-10

3.2.1 (U) KIDD Kernel ................................................................. 3-10

3.2.2 (U) iKIDD ............................................................................ 3-10

3.3 (U) Future Versions ......................................................................... 3-10

(U) LIST OF FIGURES

Figure No. Title Page

2-1 (U) Debris Observables from Flight Test ................................................ 2-2 2-2 (U) Debris Recovered from Sled Test ..................................................... 2-2 2-3 (U) KIDD Kernel Models and Interaction .............................................. 2-4

For Official Use Only 1-1

SECTION 1. (U) INTRODUCTION

(U) Hit-to-kill (HTK) kinetic energy weapons (KEWs), such as the Exoatmospheric Kill Vehicle (EKV), Terminal High Altitude Area Defense (THAAD) interceptor, Standard Missile-3 (SM-3), and the Patriot Anti-missile Capability (PAC-3) interceptor, utilize interceptors or kill vehicles (KVs) which intercept ballistic missile targets at high closing velocities. These high-energy collisions result in large numbers of debris fragments from both the target and interceptor or KV. The target may be a separating reentry vehicle (RV) or a non-separating ballistic missile in which the RV remains attached to all or part of the booster. Intercepts of the latter may result in debris from the warhead or booster section or both depending on the engagement geometry.

(U) Debris fragments can affect the performance and survivability of other interceptors and interoperability with other missile defense elements by serving as additional objects requiring observation and discrimination or by physically impacting interceptors attempting to perform subsequent engagements. There are also important kill assessment, ground effects, space debris effects, and range safety implications associated with the debris environments generated by KEW intercepts.

(U) To address the evaluation, mitigation, and exploitation of impact debris environments, the MDA has an ongoing task to operate, maintain, sustain, and develop the Kinetic Impact Debris Distribution (KIDD) computer model. KIDD is an MDA Core Lethality Model (CLM) intended to generate the complete initial debris environment from Ballistic Missile Defense System (BMDS) engagements and to support the full spectrum of impact debris analyses. In general, KIDD provides a link between lethality codes and post-intercept effects and observables codes.

(U) This volume serves as part of the documentation for the latest release, KIDD Version 7.0, which includes updates to the KIDD Kernel and Interactive KIDD (iKIDD).

A User’s Manual and Targets Manuals are also available. The former addresses installation, operation, inputs, and outputs while the latter describes the targets for which inputs are included.

For Official Use Only 2-1

SECTION 2. (U) KIDD OVERVIEW

2.1 (U) GENERAL INFORMATION

(U) The KIDD Kernel is designed to predict the initial debris environments resulting from hit-to-kill missile defense engagements. It is a model, rather than a simulation, so these debris environments are not propagated through time and space. The Kernel is designed to be implemented within other codes and simulations, particularly engagement simulations with propagation prediction capabilities. The Kernel is also distributed as a component within several other simulations including the MDA Post- Engagement Ground Effects Model (PEGEM) CLM.

(U) To facilitate integration into and with other codes, the Kernel itself is a subroutine. A single call to this subroutine will pass in the inputs defining a given intercept and the user options, execute the Kernel for this intercept, and pass back the predicted initial debris environment data. No external files are required or generated, though optional target and interceptor input files are available. The Kernel can then be called again for additional or different intercepts. Although the Kernel is designed for use within other codes, it can also be utilized in stand-alone mode and a graphical user interface (GUI), iKIDD, is available as an option to support this mode of operation.

(U) Inputs to the KIDD Kernel include information about the missile to be intercepted or target, the interceptor or KV, the engagement conditions, and the type of computations and outputs the user desires. Solid geometry models of the intercepted and intercepting missiles are not required, but libraries based on such models are available. In addition, default, generic inputs for each target type and the interceptor can be accessed by the user and complete sample input files for each target type are included in the package for use with iKIDD.

(U) The user typically supplies information about the type of threat warhead on the intercepted missile and information about the geometry and materials on both missiles. These data can be of low or high fidelity depending on how much information is known and the fidelity required for the debris results. The Kernel allows for wide latitude in terms of inputs and is highly compatible with capability-based assessments. The user has the option of accessing target and interceptor inputs from the available files.

(U) Engagement conditions include the velocity vectors of each missile and the lethality of the engagement. Lethality can be provided parametrically or obtained from a lethality simulation such as the Parametric Endo/Exoatmospheric Lethality Simulation (PEELS), another MDA CLM. Many of the debris models in KIDD are correlated to the input lethality level.

For Official Use Only 2-2

(U) KIDD is designed to be a fast-running tool and is capable of supporting real-time applications. It is a semi-empirical, medium-fidelity code based on data from hundreds of tests and for hundreds of thousands of debris fragments. These tests include flight, sled, gun, and arena tests and the data sets include recovered fragments and remote observables. Figure 2-1 shows an example of the latter and Figure 2-2 shows an example of the former.

Figure 2-1. (U) Debris Observables from Flight Test

(U) The KIDD Kernel has no look-up database as part of the code. However, the models and algorithms within the code are based largely on these data supplemented by a number of analytical models and the basic laws of conservation.

Figure 2-2. (U) Debris Recovered from Sled Test

Unclassified

For Official Use Only 2-3

(U) Fidelity and confidence are generally governed by the quality of data upon which the relevant models are based and range from high to low depending on the details of the missiles and their engagement, the type of debris characteristic, and the portion of the debris distribution of interest. The KIDD developer and Independent Verification & Validation (IV&V) Agent are responsible for generating KIDD Basis of Confidence reports for the MDA that document the VV&A status in more detail.

(FOUO) Targets may represent a separated RV, a separated booster stage, or a unitary configuration with the payload still attached to a booster stage. For the latter, the interceptor may initially engage either the RV, booster, or interface section of the missile.

For nuclear warheads, which include some high explosive (HE), and HE warheads, KIDD addresses the effects of HE detonation on the debris environments. Separated booster stages may also be detonated in self-destruct events. Targets may also represent non-lethal objects (NLOs), debris, chaff canisters, other KVs, resident space objects (RSOs), and decoys.

(U) Debris characteristics modeled in the KIDD Kernel include fragment length, width, and height; mass; material; thickness; minimum and maximum projected areas;

initial velocity vector, initial orientation and rotation rate quaternions; initial tumble rate;

average and maximum narrow-band radar cross section (RCS) for both principal polarization (PP) and orthogonal polarization (OP); standard deviation of the RCS; initial temperature; supersonic and subsonic ballistic coefficients; and supersonic and subsonic lift-to-drag ratios.

(FOUO) Fragment data for the intercepted and intercepting missile debris will be output for all fragments, from meters to microns, meeting the user’s criteria. Fragment materials currently modeled include aluminum, uranium, steel, silica phenolic, fiberglass, copper, foam, plastic, rubber, tungsten, beryllium, titanium, carbon, and high explosive.

Figure 2-3 illustrates correlation paths between the various models within KIDD. The size and mass model correlation involves an iterative algorithm.

(U) These debris characteristics are designed to be used, and have been used, to support a wide range of missile defense applications. There have been no attempts to incorporate defense-conservative modeling within the Kernel for any particular application. And when features are added or expanded to support new applications, the capabilities supporting previous applications are typically retained.

(U) The KIDD Kernel is generally provided to approved users as a Dynamically Linkable Library (DLL) for Microsoft Windows or a static object file library (.a) for Linux, Solaris, and OS X. Object and library files for other platforms and operating systems may be provided on a case-by-base basis. The Kernel is written in FORTRAN, but the DLL can be called from other programming languages as well. A structure-based Application Programming Interface (API) is included with utility functions in FORTRAN-90, C, and C++. A machine-independent random number generator is

For Official Use Only 2-4 included for complete portability across machines and OS. The KIDD User’s Manual defines all the inputs and outputs including units, which are all standard SI, and describes any associated restrictions or subtleties. Example source code for both the Structure API and the conventional Subroutine API are provided for stand-alone use or to help the user prepare an interface with another simulation. The iKIDD GUI is available for both Microsoft Windows and OS X.

Figure 2-3. (U) KIDD Kernel Models and Interaction

2.2 (U) DATA SOURCES

(U) KIDD development has mirrored the availability of debris data on which to base development and validation. These data have been obtained from a wide range of complementary tests.

2.2.1 (U) GUN TESTS

(U) Gun tests involve sub-scale, generally one-quarter to one-half scale, targets and projectiles. The lower fidelity projectiles are capable of achieving high speeds (e.g., 7 km/s). High mass recovery for the target, often exceeding 90 percent, is typical within the evacuated chamber and Unclassified

For Official Use Only 2-5 microdebris can be recovered or collected via witness panels. X-rays, laser photos, velocimeters, radars, radiometers, etc. can be utilized for the measurement of fragment speeds, sizes, temperatures, etc. during the test. The chamber walls limit fragment trajectories and provide a potential source for secondary damage, but this can be minimized.

2.2.2 (U) SLED TESTS

(U) Full-scale targets are typically impacted at sled test events, but impact speeds are limited to 2-3 km/s. Both interceptor and target debris can be recovered, with mass recovery typically about 50 percent for the target and less for the interceptor. Fragment recovery locations can be accurately measured with GPS.

Other forms of data collection include high-speed video, infrared, and radar data.

Atmospheric and ground effects must be accounted for in assessing debris data.

2.2.3 (U) FLIGHT TESTS

(U) Flight tests provide for full-scale testing in realistic environments. Debris trajectories through the exo/endo atmosphere, aerodynamic effects, fragment ground impact locations, and infrared/radar sensor data can be collected for both target and interceptor debris.

Low mass recovery, special interceptor trajectories, variable target fidelity, and potential secondary fragmentation due to aerodynamic loading must be accounted for in assessing debris data.

2.3 (U) THREAT TYPES SUPPORTED

(U) Models within the KIDD Kernel address the debris from intercepts of missiles with nuclear, chemical, biological, and HE payloads as well as flight test target missiles with no warhead. Nuclear payloads include a wide range of warhead designs. Chemical payloads include bulk, canister submunition, and chemical bomblet warheads. Biological payloads are limited to powder bomblet warheads. HE payloads are limited to unitary blast warheads and Runway Cratering Munition (RCM) submunition warheads. All of these threat types can be separated RVs or RVs still attached to spent booster stages and spent separated booster stages can also be modeled as targets. Threats also include NLOs, debris, chaff canisters, and decoys. Additional targets include other KVs and RSOs.

Unclassified

For Official Use Only 2-6

(U) Debris generated by the Kernel includes solid-state materials and does not include any dispersed liquids or powders unless contained within a solid-state material.

Examples of the latter include undamaged submunitions and bomblets. Within KIDD the submunition and bomblet debris, including undamaged, damaged, and fragmented submunitions and bomblets, are tracked and output separately from the other debris.

Debris from the intercepted target is tracked separately from debris from the intercepting missile and debris from the RV is tracked separately from debris from the attached booster. This separation allows the user to work with only the desired debris constituents.

2.4 (U) APPLICATIONS

(FOUO) Initial KIDD applications involved functional survivability and performance degradation of interceptors that had to fly through, and have their seeker look through, debris environments generated by previous intercepts and then potentially discriminate between ballistic missiles and debris fragments. The former requires modeling the small debris environments (i.e., microdebris) while the latter requires modeling the largest fragments in detail.

(FOUO) Subsequent applications included hit and kill assessment where KIDD is used to develop and test algorithms by which optical sensors and radars can remotely determine the outcome of an engagement through observables. This application requires correlation of the debris environments to the intercept lethality. Other in-flight KIDD applications involve interoperability between missile defense elements where upper tier elements should communicate debris generation to lower tier elements required to operate through and mitigate adverse effects of the resulting debris environments.

(U) In addition to in-flight applications, recent analyses have focused on ground effects, space debris effects, and flight test range safety applications. The probabilistic nature of these applications requires Monte Carlo capabilities and providing sufficient initial conditions to support subsequent simulation of the debris environments over a much longer period of time. The KIDD models were updated to improve the applicability and confidence associated with such applications.

(U) The KIDD Kernel can be used to generate binned and/or discrete fragment data. For the former the user defines bins reflecting fragment sizes and/or masses of interest and KIDD will provide aggregate results for the fragments in each bin including the number, integrated mass, velocity, and temperature of the fragments associated with each bin. This option is particularly suitable for analyzing small debris.

(U) For the discrete option, the Kernel will generate extensive data for each individual fragment based on the size, mass, or number thresholds set by the user. These data are intended to provide all the information needed for performing subsequent optical or radar signature, exo-atmospheric or endo-atmospheric propagation, or other debris simulations or analyses.

For Official Use Only 2-7

(U) For many applications, range safety for example, it is recommended to employ the Kernel in Monte Carlo fashion since a number of statistical variations are modeled within the code and the full range of variations may not be captured with a single Monte Carlo trial. The KIDD Kernel is configured so that multiple calls can be treated the same for repeatability or as separate trials for Monte Carlo assessments. The user is provided options by which the statistical variations can be captured with just a few runs. The user is also provided options to minimize the execution time, by eliminating non-essential calculations, where large numbers of runs or faster response times are required.

(U) The MDA Basis of Confidence document provides the user with the levels of confidence that the outputs are consistent with the developer requirements and documentation and would be representative of the real world. The user can use this information to determine the confidence associated with their particular KIDD application.

For Official Use Only 3-1

SECTION 3. (U) HISTORICAL OVERVIEW

3.1 (U) PRIOR VERSIONS

3.1.1 (U) VERSION 1

(U) Version 1.0 of the KIDD Kernel was released in 1992 allowing the KIDD breakup models to be used without requiring use of other parts of the KIDD suite. The initial application of the KIDD Kernel was integration with the U.S. Air Force Phillips Laboratory IMPACT and Debris Radiance (DEBRA) codes developed by the Aerospace Corporation. Such an integration was requested by the test bed community (e.g., National Test Bed) to provide a single debris model including Air Force efforts with regard to booster and satellite debris along with Army efforts addressing RV debris.

(U) Only lethal engagements of long-range nuclear RVs were addressed by the initial version of the KIDD Kernel and the models were based on limited radar data from two flight test intercepts. An expanded version released in 1993 (i.e., Version 1.3) incorporated debris data from numerous ground-based impact lethality tests and addressed both lethal and non-lethal engagements of long-range nuclear RVs. The ground-based tests allowed fragments to be recovered and analyzed in detail permitting higher-fidelity fragment models to be developed and implemented.

3.1.2 (U) VERSION 2

(U) Version 2.0, released in 1994, was based on additional data from ground-based lethality tests which allowed debris environments from a wider range of engagements to be addressed. The existing nuclear RV debris models were expanded to include lower-velocity intercepts of shorter-range nuclear RVs. New models were developed and implemented to address debris from intercepts of RVs with chemical submunition (CSM) warheads based on the lethality of the engagement. Improvements and expansions were made to the fragment shape and velocity distributions and new outputs defining fragment ballistic characteristics (e.g., ballistic coefficient) were included to support fragment endoatmospheric propagation analyses.

3.1.3 (U) VERSION 3

(U) Version 3.0, released in 1995, added a bulk chemical warhead capability to address additional engagements. A bimodal fragment size distribution model was implemented to facilitate this addition. This new approach was also used to improve the models associated with the other target types, particularly the missile (i.e., non-submunition) debris from intercepts of RVs containing CSM warheads, and to provide a more robust basis for future additions. A number of other improvements, which applied to all target types, were implemented in Version 3.0. These included a new fragment RCS

For Official Use Only 3-2 model, a new fragment material model, and improvements to the shape, mass, drag coefficient, and velocity models. A machine-independent random number generator was incorporated and several modifications were made to improve portability.

(U) Version 3.1, released in 1996, included further improvements to the bulk chemical warhead debris distribution model, fragment shape model, projected area model, ballistic coefficient model, fragment RCS model, and debris cloud center-of-mass velocity model. The bulk chemical warhead model was made applicable to a wider range of warhead mass fractions. Special fragment shape characteristics were modeled for the largest and heaviest fragments. Variations about the nominal, minimum, and maximum projected areas were addressed. Supersonic ballistic coefficient estimates were provided to supplement the subsonic data provided by previous versions. RCS data were provided over a wider range of radar frequencies and fragment sizes. Momentum transfer was correlated more closely with the closing velocity.

(U) Version 3.5, released in 1998, added the capability to address non-separating targets. The RV or different parts of the booster section of such targets could be hit and the resulting debris simulated. RV hits result in non-impacted booster section debris while booster hits will result in a single piece of RV debris. Since booster materials and designs are different from RVs, a more robust material model was also added. Although previous versions required material and thickness data for the target, these data, particularly the latter, did not play a large role in the calculations. The fragment mass model became a fragment surface area model which uses the user-defined thickness data to determine more-realistic fragment masses. Multiple thickness combinations could be associated with each material to further improve the fidelity. Since the input requirements increased significantly, an option was added to invoke a full set of generic, default inputs for any target type to help users unfamiliar with such data.

(U) Previous versions assumed a single debris cloud center-of-mass velocity associated with the target and KV debris based on momentum transfer between the two vehicles. This is a reasonable approach for highly lethal engagements of compact targets and perhaps for the small debris from other engagements, but does not apply to the larger debris resulting from highly non-lethal engagements or engagements of extended targets such as non-separated missiles. Although these cloud center-of-mass velocities were still computed and output, they were no longer used in the fragment state vector calculations or in the total energy or momentum calculations.

(U) Version 3.5 still applied a momentum transfer based velocity in the direction of the closing velocity vector, but the magnitude of this velocity was a function of the fragment size. Therefore, if large fragments were present, there was no longer a single cloud center-of-mass about which all fragments were dispersed. The dispersion velocity model added a term based on the KV-to-target mass ratio. A new option was added to allow the user to have the fragment state vectors adjusted to give explicit conservation of

For Official Use Only 3-3 momentum for the single trial and subset of fragments being simulated rather than for only a large number of trials.

(U) Further improvements in Version 3.5 addressed the CSM target debris models and bulk chemical target debris models as well as the nuclear target debris models. The latter reflected differences between lethal engagements with and without an HE detonation. The KV debris models were altered to better address non-lethal and booster engagements. Version 3.5.3, released in 1999, added a correlation of the fragment random mass to the width and height of the fragment, consistent with test data, and incorporated improvements to the average density calculation and the Monte Carlo logic.

3.1.4 (U) VERSION 4

(U) Version 4.0, released in 2000, added a unitary HE blast warhead target option.

Since no test data were available for this target type at that time, the models for this target were based largely on reviews of design documentation and tests done with other targets incorporating HE. For bulk chemical targets, the algorithms used to calculate the size of the largest fragment were improved based on data from more recent tests, which expanded the bulk chemical target database significantly. Parameters used to calculate the large and heavy fragment size and mass for long-range nuclear targets were also improved based on a large number of more recent tests from which extensive debris data were collected and analyzed. The shape model for the large and heavy fragments was also improved based on these test data.

(U) Material and thickness inputs were modified in Version 4.0 to allow the user more flexibility in selecting materials for each vehicle. Input materials not expected to be fragmented could be identified and the material fraction model would generate appropriate material fractions for the remaining fragmented materials. This allowed a wide range of calculations to be performed with only minor changes to the inputs. But it could increase the complexity of material inputs so the original input format was retained as an option.

(U) The momentum transfer model was updated in Version 4.0 based on more recent test data and analyses, but the basic form of the model was not changed.

Momentum transfer related speeds were based on the mass engaged rather than the entire mass of the target. This results in higher speeds when missiles with an attached booster are intercepted since momentum is only transferred to the RV or the booster. Note that intercepts in which both the RV and booster are engaged (i.e., interface hits) were not explicitly modeled in KIDD, but could be effectively addressed by making the engaged section of the booster part of the RV in the input data. The empirical momentum transfer model was explicitly correlated to the lethality and damage level associated with the engagement and a gamma distribution was applied to give some variation to the momentum transfer related speeds, consistent with applicable test data.

For Official Use Only 3-4

(U) The fragment RCS model was originally developed for metals, i.e., conductors, but Version 4.0 added an RCS adjustment for the non-metal materials included in KIDD. A new lift-to-drag ratio term was added to the outputs for the discrete fragments. While this initial model was simple, it was based on data from a number of endoatmospheric tests. The discrete fragment random mass, and random ballistic coefficient, models were improved to force integration over all the discrete fragments to equal an integration of the nominal masses for the same set of fragments.

(U) In Version 4.1, released in 2001, a total of six new materials were added. Two of the new materials, copper and tungsten, were metallic. The remaining four new materials (fiberglass, foam, plastic, and rubber) were non-metallic. The RCS adjustments introduced in Version 4.0 for non-metals were expanded to include the new non-metallic materials in Version 4.1. Improvements to the discrete fragment random material selection, random mass, shape, and temperature models were made as well.

(U) An additional input option was added in Version 4.1 to allow input material mass fractions identified as non-fragmented to be used in the heavy fragment mass calculation. This option was added because the heavy fragment will often be composed of most or all of the non-fragmented material. This option uses the same non-fragmented material inputs that were introduced in Version 4.0 for both lethal and non-lethal intercepts. Since the heavy fragment mass can have considerable influence on mass conservation iterations and most of the fragment distributions, users with high confidence in the non-fragmented material fractions for a vehicle can improve the accuracy of their results with this option. If the non-fragmented materials are not known, this option should not be used. The previous methodology was employed when the new option is not activated.

(U) The number of error and warning flags was increased from 9 to 31 in Version

4.1. This provided users a more detailed description of any errors or warnings encountered during a run, particularly potential errors or inconsistencies with inputs. If any of these flags were activated during code execution, the code may continue to completion, but results could be erroneous if a serious error occurred. Therefore, users are encouraged to review these flags. Since the Kernel is typically integrated into other simulations, only a few potentially-fatal errors will force a return and every attempt is made to run to completion otherwise.

(U) Several of the empirically-derived parameters were updated in Version 4.1 to reflect results of the most recent test data analyses and convergence criteria were also adjusted to improve accuracy with a minor increase in run time. Several equality conditionals using real numbers were modified to use inequalities in response to results of an independent assessment of the KIDD Kernel. Beginning with Version 4.1.1, released early in 2002, source code was no longer distributed. Object code and libraries were distributed to address configuration management requirements.

For Official Use Only 3-5

(U) KIDD Kernel Version 4.2.2, released in 2002, added several outputs needed by KIDD Bounds, a KIDD post-processor, and therefore required by PEGEM which uses this post-processor. A new option affecting how the random number generator was initialized for Monte Carlo trials was also incorporated for PEGEM. And the machine-independent random number generator was augmented by the optional use of the intrinsic random number generator since the latter is typically faster.

3.1.5 (U) VERSION 5

(U) Version 5.0 of the Kernel, released in 2003, added chemical liquid bomblet and biological powder bomblet warhead types based on extensive data collected and analyzed for both, but particularly the former. An exponential size distribution regime was modeled to address the unique characteristics of bomblet debris distributions. This version of the Kernel provided an alternative chemical warhead type and an initial capability for biological warheads. This was the first version to have formal IV&V.

(U) Improvements were also made in a number of the discrete fragment models, particularly the projected area models. In addition to a number of new outputs relating to the new target types, the version number was added as an output to facilitate developer support to users needing technical assistance.

(U) KIDD Kernel Version 5.2, released in 2004, included a number of enhancements related to debris from attached booster stages. Prior versions modeled debris from impacts of either the RV or the attached booster, but not impacts in the interface region between the two. Version 5.2 added support for interface impacts and generated appropriate debris from both the RV and the attached booster.

(U) In addition, booster joint breakup was explicitly modeled in Version 5.2. Prior versions assumed that all booster joints would break at impact. Version 5.2 allowed three options: (1) no joints break other than those associated with impacted sections; (2) all joints break consistent with the prior model; and (3) KIDD determines which joints break. The third option computed the booster loading based on KV interaction calculations, determined the axial and bending stress at each defined joint, and compared the resulting stress to the joint strength to predict joint failure. Joint strength was based on new joint bolt and/or aeroshell inputs. Joint failure was reported in new outputs defining the status of the joints and non-impacted sections. The effects of joint failure were reflected in existing bin and discrete fragment outputs. The new joint breakup model applied to RV, interface, and booster impacts.

(U) Other enhancements included replacing some single precision calculations with double precision algorithms to improve accuracy and avoid potential overflows and additional user input checking was implemented to avoid overflows. An externally callable function was included to facilitate version tracking by the user when upgrading

For Official Use Only 3-6 or when using different versions on the same platform. KIDD documentation significantly was improved, particularly with respect to the new user’s perspective.

(U) The iKIDD GUI was first included with KIDD Kernel Version 5.2. This represented the initial release of iKIDD, an optional stand-alone application available for both Windows and Mac OS X. Inputs in iKIDD were logically organized into related input panels with error and consistency checks. Input sets could be saved and reloaded and sample input sets were included. KIDD inputs were saved in standard NAMELIST format for use in other simulations. Outputs included a suite of common 2D and 3D debris plots and all KIDD statistical and discrete outputs could also be graphed in customizable plots. Plots could be exported in standard graphic formats. ASCII files of all KIDD outputs were also generated for user post-processing.

(U) KIDD Kernel Version 5.4, released in 2005, included the implementation of a non-runtime interface with PEELS. Coordination between the KIDD and PEELS developers resulted in the generation of Target Data Files for most of the PEELS target types that are also modeled in KIDD. These Target Data Files provided users the option of accessing KIDD target inputs based on verified and accredited PEELS inputs. For users of both KIDD and PEELS this ensured consistency in target inputs between the two. KIDD Target Data Files were made available for distribution to users at multiple classification levels.

(U) Other enhancements included support for faster execution times and the modeling of additional materials. For faster execution times, a KIDD Kernel option was implemented to pre-process many of the computations without knowledge of the final engagement conditions. This allowed extremely fast execution of the remaining computations when these conditions are known. This feature was designed to support hardware-in-the-loop simulations and testbeds with real-time requirements.

Improvements were also made to make calls to the KIDD Kernel more thread-safe for such applications.

(U) Two new materials, beryllium and titanium, were added to KIDD and several relatively minor errors, many of them identified by the IV&V Agent, were also corrected.

This was the first version to have IV&V performed before release to users.

(U) An update to the iKIDD GUI was released in conjunction with Version 5.4.

This version of iKIDD captured the full functionality of this version of the KIDD Kernel including the KIDD-PEELS non-runtime interface and use of KIDD Target Data Files. In addition, this version of iKIDD allowed outputs from multiple KIDD Kernel runs to be graphed on single plots, allowed more user control over the format and appearance of output plots, provided expanded user help functions, and supported range checking for all inputs.

For Official Use Only 3-7

3.1.6 (U) VERSION 6

(U) Version 6.0 was released in 2006 and included enhancements and corrections to both the KIDD Kernel and iKIDD.

(U) The KIDD Kernel in Version 6.0 added support for a new target type to address flight test targets that do not contain a specific threat-like warhead. Modifications in Version 6.0 also completed the efforts initiated with Version 5.4 to make the KIDD Kernel thread-safe for simultaneous multi-thread calls.

(U) The non-runtime interface with PEELS was updated to include additional information and minor modifications and additions were made to the Target Data Files.

Version 6.0 also included the initial implementation of the runtime interface with PEELS.

This interface builds upon the non-runtime interface and utilizes the new .KID generated by PEELS Version 10.3.1. This file includes PEELS inputs and outputs, for TRAJ option runs, which are then read by KIDD to ensure consistency and improve fidelity.

(U) Fragment velocity models were updated to reflect additional anchoring to flight test data and the RCS models were updated to better address non-impacted RV sections. Additional user input checking, default values, and warning flags were implemented. Corrections addressed issues with random masses for bomblet fragments, issues with shape and projected areas for the largest and heaviest fragments, and adjustments to the random mass and energy computations to account for the mass conservation convergence tolerance.

(U) The iKIDD in Version 6.0 fully supported the KIDD Kernel updates in Version 6.0 and included new tools to simplify preparing material component and size/mass bin inputs. It also allowed material components to be entered by mass or mass fractions and similar material components to be combined. The plotting capabilities were enhanced to support global custom plot libraries, allow additional plot customization, and maintain consistent colors and symbols in the default plots. Output data files were enhanced with increased precision.

(U) KIDD Kernel Version 6.2, released in 2008, included an upgraded flight test target without a specific threat-like warhead model, based on more-recent test and numerical data. The runtime interface format was updated to be compatible with PEELS Version 10.5, including the corresponding PEELS target model updates. An externally callable function was added to facilitate debug operations, create KIDD/iKIDD input files, and generate tab-delimited output files.

(U) Several minor model improvements were made in Version 6.2. Additional bomblet input error checking was added. Fragment binning algorithms were updated to eliminate potential negative numbers in very small bins. The RCS algorithm was updated so that the input radar wavelength will not affect the random number sequence. Cloud

For Official Use Only 3-8 expansion velocities for non-separating targets were updated to be more closely related to overlapped target mass.

(U) The default crossing angle calculation in Version 6.2 was updated to use the angle between the target and relative velocity vectors. And, while not validated, support for relative velocities down to zero was included. For Version 6.2 compiled libraries are provided for both 32-bit and 64-bit processors.

(U) In iKIDD 6.2, the PEELS runtime interface format was updated to be compatible with PEELS Version 10.5 file formats. Plotting capabilities were significantly improved to include more powerful plot styles and additional customization, improve plot zooming, have expanded custom and default plot libraries with improved plot library management capabilities, allow enhanced plotting from external data files, and identify target aliases when using the runtime interface.

(U) KIDD Kernel Version 6.5, released in 2011, added multiple wavelength RCS calculations, the capability to model debris from engagements of booster stages without an attached RV, debris subset options, and an optional structure API.

(U) The multiple wavelength RCS option allowed debris RCS characteristics to be generated for multiple radar wavelengths from a single KIDD run and eliminated the overall run time and random number sequence issues associated with using multiple KIDD runs to generate RCS characteristics for radars for different wavelengths. This has become increasingly important to support system-level BMDS events involving multiple elements.

(U) The KIDD-PEELS Run Time Interface was updated to allow compatibility with PEELS Version 12.0, including the corresponding PEELS target model updates.

These included three new targets and four updated targets.

(U) New debris subset options were added in Version 6.5. Previous KIDD versions allowed selection of all discrete fragments heavier than a requested mass, larger than a requested size, or a specified number of fragments. Small size thresholds are sometimes needed to investigate characteristics of relevant fragments (e.g., low ballistic coefficients), but can result in large numbers of fragments. Two new threshold options were added to allow numerically smaller subsets of fragments that extend down to a requested size threshold. One used a logarithmic step size, which results in more large fragments (i.e., fewer large fragments skipped and more small fragments skipped). The other used a linear step size, which results in equal size step skipping of fragments down to the threshold.

(U) All discrete fragment thresholds were expanded to an array in Version 6.5 to allow users to input different size, mass, and number thresholds for each vehicle. This capability allowed the number of fragments for each vehicle to be optimized for the

For Official Use Only 3-9 specific application and framework such that only the highest priority fragments are considered.

(U) An additional, optional API based on structure was made available in Version

6.5. The structure API argument names were longer, but more descriptive and structured, than the conventional subroutine API argument names. Both interfaces utilized the same KIDD Kernel to determine debris.

(U) The KIDD Structure API types and dimensions all KIDD interface arguments and constants. The KIDD Structure API also defines functions in FORTRAN-90, C, and C++ to load and save inputs to a KIDD NameList (KNL) file, a function to load target data from a Target Data File (TDF), functions to write outputs, functions to copy inputs between structures, and constants that define input settings and array element index order.

This capability significantly reduced the number of arguments in KIDD Kernel calls, removed the need for users to type and dimension numerous arguments, simplified the matching of target and interceptor inputs for system-level BMDS events, allowed users to ignore arguments irrelevant for their applications, supported pre-loading of inputs during initialization to eliminate file access during real-time operations, and simplified the process of updating interfaces to future versions of KIDD.

(U) KIDD Kernel Version 6.8, released in late 2011, added the capability to select pre-generated Interceptor Data Files (IDFs). This is similar to the Target Data File implementation. Available IDFs can be requested using the Special Access section of the KIDD Data Request Form (DRF). IDF content and release were coordinated with the corresponding BMDS element offices and new/updated IDFs will be included as they become available. The KIDD Structure API was expanded to include a function to load interceptor data from an IDF.

(U) KIDD Kernel Version 6.8.2, released in late 2012, added a Runway Cratering Munition (RCM) warhead type representing the first High Explosive Submunition (HES) warhead type in KIDD. The intrinsic random number generator option was removed and an improved machine-independent random number generator replaced the random number generator from previous KIDD versions. New material types for carbon and high explosive were added. The KIDD-PEELS Run Time Interface was updated to allow compatibility with PEELS Version 12.2, including the corresponding PEELS target model updates. Target Data Files were split into separate files consistent with Interceptor Data Files and Interceptor and Target CFF files were combined into the corresponding IDF and TDF.

(U) KIDD Kernel Version 6.8.3, released in early 2014, added a number of new outputs including orthogonal polarization RCS and initial orientation and rotation rate quaternions. To minimize additional memory requirements, several redundant or obsolete outputs were removed and the random RCS output was moved to an externally callable function. Minor improvements were made to the drag, lift-to-drag, RCS, and projected

For Official Use Only 3-10 area models and a new input was implemented to provide additional user control for targets with high explosives. The KIDD-PEELS Run Time Interface was updated to allow compatibility with PEELS Version 12.3. Several new Target Data Files and one new Interceptor Data File were incorporated.

(U) Support was added in iKIDD 6.8.3 to reflect all the changes to the KIDD

Kernel. An explicit Monte Carlo capability was added so that multiple trials can be run and plotted. The default plot options were expanded and a new option was added to allow saving of target and/or interceptor data in separate files for easier reloading in subsequent runs.

3.2 (U) CURRENT VERSION

Version 7.0 was released with this report and included enhancements to both the KIDD Kernel and iKIDD.

3.2.1 (U) KIDD Kernel

(FOUO) KIDD Kernel Version 7.0 added a number of new target types including NLOs, debris, chaff canisters, KVs, RSOs, and decoys. Self-destruct debris options were added for separated booster stages. Enhanced correlation between the largest and heaviest fragment, and therefore the entire fragment distribution, and the engagement hit point, strike angle, and closing speed were implemented based on test data. An option was added to have KIDD perform the initial (i.e., up to 1 msec) atmospheric effects on fragment speeds including reduced acceleration and subsequent deceleration. This results in more accurate “initial” speeds and avoids the need for users to perform the small time-step propagation computations required during this period. An external call was added to provide the material density and thermo-physical properties used within KIDD so that users can use these values in their post-processing. A minor modification was made in how multi-material fragments are reported and new error flags were added. Several new Target Data Files were incorporated including generic input data sets for each of the new target types.

3.2.2 (U) iKIDD

(U) Support was added in iKIDD to reflect all the changes to the KIDD Kernel and a number of minor iKIDD issues involving windows, visual glitches, menus, plot warnings, and reading older TDFs were addressed. The iKIDD codebase was upgraded to Xojo 2014.

3.3 (U) FUTURE VERSIONS

(U) Efforts and plans are in place for future versions of the KIDD Kernel and iKIDD as documented in the KIDD Long Term Development Plan and Acquisition

For Official Use Only 3-11

Program Baseline. Users are encouraged to submit change requests, problem reports, and requirements for future modifications and capabilities, and forms are provided on the distribution CD. Users are also encouraged to identify the use of KIDD in their associated briefings and reports to help promote continued MDA development.

(U) Current near and mid-term plans include anchoring to additional flight test data and first principle code results, expanded bin or cloud statistical outputs, more advanced RCS modeling, impact flash modeling, modeling debris from impacts of additional warhead types, and modeling debris from engagements by other types of interceptors.