One of the Doctor’s many New Year’s resolutions is to convince every heat treater to purchase and use simulation software for recipe creation and to aid in process control and troubleshooting. The time has come – the technology is proven and robust. Today, we will talk about simulators and controls for atmosphere carburizing. In our March 2018 column (since February’s focus is on nonferrous heat treatment), we will talk about low-pressure carburizing simulators. Then in a follow-up column, we will introduce you to nitriding/ferritic nitrocarburizing simulators. Let’s learn more.
Why Simulators?
We all know how important precise furnace atmosphere control is to successful heat treating. Meeting or exceeding customer specifications and producing both predictable and repeatable metallurgical and mechanical property results are the goal of every heat-treat operation. Nowhere is this more critical than in carburizing and other case-hardening processes where we deliberately change the composition of the furnace atmosphere at various points in the cycle. In addition, changes to material composition, part loading and/or process parameters (time, temperature, carbon potential) influence the final results.
In gas carburizing, the addition of carbon at the part surface followed by diffusion allows us to achieve both the desired case depth and hardness. Measurement and control of furnace atmospheres are often performed continuously using in-situ (oxygen probe) devices or via extractive (dew point or infrared) methods. A combination of techniques is often used.
Simulation and modeling software allow us to predict and determine in real time the carbon profile. Taking this one step further, one can use this type of software to control the process by utilizing the atmosphere inputs and varying the time to achieve the desired case depth and carbon profile at any given temperature. Prediction of hardness profiles is just an algorithm away.
Historical Overview
Years ago, the only effective way of trying to control the carburizing process was to establish a relationship between the volume of enriching gas to that of the carrier gas. This led to rules of thumb such as adding natural gas at 10% of the endothermic gas flow for carburizing or adding ammonia additions at 3-5% of the endothermic gas flow when carbonitriding. This was followed by techniques involving measurement of the carbon potential using either dew point or periodic shim-stock measurements.
These methods were common up until the early 1970s, when the first oxygen (carbon) probes were introduced. Once the zirconia carbon-sensor technology was proven robust, the technology swept the industry. Over the years, this measurement method has been refined, first by introducing a sensor correction factor and then using three-gas (CO, CO2 and CH4) infrared analysis in conjunction with the oxygen probe to input a more accurate carbon setpoint. It then evolved further with the introduction of continuous, nondispersive infrared analyzers that measure continuously rather than periodically, allowing for automatic operation.
Today, coil tests are being used to simplify shim-stock verification methods. Finally, true representation of atmosphere was achieved using metallurgical evaluation of parts in combination with the aforementioned methods. From all these results, proper correction to the carbon calculation can be determined, yielding more accuracy on the in-situ control parameter from the oxygen probe.
While modeling/simulation of the carburizing process began in earnest in the 1970s, it wasn’t until the advent of more powerful, smaller computers and their industrialization that programs emerged to allow complex calculations in real time. However, these tools only became of interest to the heat-treat industry over the past two decades as demands for higher quality and tighter control became paramount.
One of the lessons learned was that calculating the carbon potential for gas carburizing requires assumptions with regard to material chemistry, base atmosphere composition, temperature and changes to the atmosphere over time. The composition of the base atmosphere and the way in which the gas is measured were found to have significant effects on the accuracy of the calculated carbon potential.
Today, many control methods are time-based with changes to carbon and temperature setpoints. Due to the variable nature of the process, however, this method does not provide the necessary accuracy of control or repeatability of results. In addition, it is common to be conservative (“err on the safe side”) when providing a time for segments of the process. Simulators and their associated control systems (Fig. 1) optimize the carburizing process.
How to Optimize Results
The best way in which to utilize a predictive model in real-time control is to ensure that the variables accurately represent the conditions to which the parts are exposed; to build confidence by running data-logged values through the modeling software; and to compare those values to metallurgical results. Using this three-step process of building, verifying and controlling allows for accuracy in method and confidence in results. “What-if” analysis is another important aspect of these programs, allowing different inputs (e.g., time, temperature and boost-diffuse steps) to be explored.
Atmosphere Carburizing Simulator Example
One simulation program on the market is CarbCalcII (Super Systems Inc.). This control software package is capable of not only predicting the carbon profile developed during atmosphere carburizing, it can be used for real-time monitoring, real-time control, process replay modes and what-if simulation. It includes the capability to predict the carbon-transfer rate between the furnace atmosphere and steel during boost-diffuse carburizing or steady-state (constant carbon-potential) carburizing and during temperature drops and stabilization at hardening temperature. The ease of use, recipe development and ability to “tweak” the cycle are key advantages of the software.
Multiple screens and operating modes (Fig. 2) allow the user to input key process variables (including material chemistry and atmosphere composition), create multiple segment recipes, view the impact on results and view the carbon profile (actual versus desired). For example, the software comes with an alloy database that includes the most common steels used in heat treating. Within the standard materials is the ability to change an element’s chemistry (e.g., the ability to input the actual heat carbon content). It also allows the user to create custom chemistries to support new or foreign alloys that are becoming more common today.
In real-time and replay modes, the output allows users to display the carbon profile using data retrieved from the control instrumentation while process control runs in the background. Carbon profiles can be altered, and the program will modify the required time at temperature based on the actual carbon profile received from the control instrumentation. In addition, there is a load-tracking system to enable historical information on previously processed loads to be quickly retrieved and profiles re-created or modified.
Summary
Atmosphere carburizing simulators and similar predictive control tools help heat treaters do their jobs more efficiently, with better control, less operator intervention and at lower cost. In addition, they simplify record keeping and provide the type of documentation demanded by today’s manufacturing community. If you or your heat treater atmosphere carburize, make a promise to yourself and your company to insist that simulators are part of the process. Simply stated, it will improve quality and keep our industry competitive.
References
- Herring, Daniel H., Atmosphere Heat Treatment, Volume II, BNP Media, 2015
- Oakes, Jim, “Understanding Atmosphere in Carburizing Applications Using Simulation and Real-Time Carbon Diffusion,” Conference Proceedings, Thermal Process Modeling and Computer Simulation Conference, ASM International, 2014
- “Carbon Diffusion Model for Atmosphere Furnaces,” CarbCalc II Operations Manual, Super Systems, Inc., 2017
