A control device of an internal combustion engine using a neural network. When a value of an operating parameter of the engine is outside a preset range, the number of nodes of a hidden layer one layer before an output layer of the neural network is increased and training data obtained by actual measurement with respect to a newly acquired value of an operating parameter of the engine is used to learn a weight of the neural network so that a difference between the output value changing corresponding to the value of the operating parameter of the engine and training data corresponding to the value of the operating parameter of the engine becomes smaller.

In some examples, a system including one or more processors may receive sensor data from one or more sensors indicating one or more engine parameters of an engine including a combustion chamber. Based on the sensor data, the system may determine a homogeneity index indicative of a homogeneity of an air-fuel mixture within the combustion chamber. Furthermore, the system may determine an estimated amount of NOx in the exhaust gas based at least in part on the homogeneity index. In addition, based at least partially on the estimated amount of NOx in the exhaust gas, the system may send an instruction to control an engine component.

An emissions control system for a motor vehicle that includes an internal combustion engine includes a first selective catalytic reduction (SCR) device and a reductant injector, The system further includes a model-based controller that is configured to calculate a target amount of reductant to inject to maintain a predetermined ratio between an amount of NH3 and an amount of NOx at the outlet of the first SCR device, and to send a command for receipt by the reductant injector to inject the calculated amount of reductant. The model-based controller is further configured to send a command for receipt by an engine controller to influence NOx production by the engine by modifying an engine operating parameter, based on a calculated target amount of NOx at the inlet of the first SCR device.

A control system for fuel injection by predicting engine noise may include an engine noise predicting device configured to derive a predicted engine noise value in real time by a predicted engine noise coefficient which is pre-stored according to a currently measured combustion pressure value of an engine; and a combustion controller configured to determine a difference between the real-time predicted engine noise value derived by the engine noise predicting device and a target engine noise value for a current operation condition of the engine, and when the engine noise is determined as being degraded due to an abnormal combustion, configured to change the target engine noise value to control fuel injection according to the changed target engine noise value.

A control system for an engine is provided, which includes an engine body formed with a cylinder, a NO.sub.x catalyst, an oxidation catalyst, a PM filter, a fuel injector configured to perform a main injection and a post injection, and a controller configured to execute a DeNO.sub.x control in which the fuel injector is controlled to perform the main and post injections so that an air-fuel ratio of exhaust gas is brought close to the stoichiometric air-fuel ratio or becomes rich and fuel supplied into the cylinder by the post injection combusts therein, and a filter regenerating control in which the fuel injector is controlled to perform the main and post injections so that the air-fuel ratio becomes lean and the fuel supplied into the cylinder by the post injection causes no combustion therein, the controller executing the DeNO.sub.x control and the filter regenerating control consecutively in this order.

Methods and systems are provided for monitoring a NOx storage capacity of a passive NOx adsorption catalyst (PNA) included in an exhaust gas after-treatment system of an engine. In one example, a method may include, after an engine cold start and prior to an exhaust gas temperature reaching an upper threshold temperature, indicating degradation of the PNA based on an amount of NOx measured downstream of the PNA during a fuel cut event and while the exhaust gas temperature is between a lower threshold temperature and the upper threshold temperature. In this way, degradation of the NOx storage capacity may be inferred based on an amount of NOx released from the PNA and independent of a NOx storage measurement.

An exhaust gas control apparatus includes a fuel injection device, a NOx occlusion reduction catalyst, a fuel addition valve, an inflow gas adjustment device, and an electronic control unit. The electronic control unit executes a low flow rate reduction treatment for removing NO.sub.x occluded in the NOx occlusion reduction catalyst after fuel supply from the fuel injection device is stopped. The electronic control unit controls the inflow gas adjustment device such that a ratio of oxygen to the fuel added to the NOx occlusion reduction catalyst at a time when a temperature of the NOx occlusion reduction catalyst is below an activation temperature becomes higher than a ratio of oxygen to the fuel added to the NOx occlusion reduction catalyst at a time when the temperature of the NOx occlusion reduction catalyst is equal to or higher than the activation temperature during the low flow rate reduction treatment.

According to one or more embodiments described herein, an exhaust system for treating exhaust gas from an internal combustion engine in a motor vehicle includes a passive NOx absorber (PNA) device, and a model-based controller that controls an amount of NOx stored by the PNA device. Controlling of the amount of NOx stored includes computing a predicted NOx storage level of the PNA device using a prediction model of the PNA device, and in response to the predicted NOx storage level of the PNA device being greater than a predetermined cold-start threshold, raising a temperature of the exhaust gas by changing an operation of the internal combustion engine.

An exhaust purification system is provided with a NOx-occlusion-reduction-type catalyst 32 that occludes NOx in exhaust when the exhaust is in a lean state and reduces and purifies the occluded NOx when the exhaust is in a rich state, and a NOx purge rich control unit 140 for reducing and purifying NOx occluded in the NOx-occlusion-reduction-type catalyst 32 by putting the exhaust into the rich state by repetitively performing fuel injection control of at least one of post injection and exhaust pipe injection at a predetermined interval.

There is provided a technology pertaining to abnormality diagnosis of an NSR catalyst that enables the diagnosis that the NSR catalyst is abnormal to be made even when the degree of deterioration of the NSR catalyst is relatively small yet. An abnormality diagnosis apparatus is applied to an exhaust gas purification apparatus having an NSR catalyst and a fuel addition valve. The abnormality diagnosis apparatus includes a controller configured to perform a specific fuel addition process and diagnose the NSR catalyst. The controller starts the specific fuel addition process when the NSR catalyst is in a specific start condition, and diagnoses the NSR catalyst on the basis of the quantity of NOx flowing out of the NSR catalyst over a specific period in the period from when the specific fuel addition process is started to when the temperature of the NSR catalyst reaches the thermal desorption temperature.